Transitioning to a Multi-polar World (#30)

with Michael O’SULLIVAN, author of The Levelling: What’s Next After Globalization

Your host, Ben Robinson, is joined by Michael O’Sullivan, author of ‘The Levelling: What’s Next After Globalization’ and former CIO of the International Wealth Management Division at Credit Suisse. Michael currently serves on the World Economic Forum’s Global Future Council on the New Economy. In this episode, Ben and Michael discuss what is the role central banks will play in the transition period to a post-globalization, multi-polar world; what international organizations should be completely reshaped to meet the needs of this new world, what new institutions should be created, and more.

Michael Recommends


  1. One book: “Putin’s People: How the KGB Took Back Russia and Then Took On the West” by Catherine Belton
  2. One influencer: Chris Watling of Longview Economics
  3. Best recent article: Work done by Lisa Sanders for the The New York Times
  4. Favourite brand: “La Femme sans Tête”, bières artisanales de Paris
  5. Productivity hack: be ready to say ‘no’, politely.

Globalization is ephemeral, it’s in the ether, in the sky, in the way that people tend to look at it. There’s no ministry for globalization. So, it’s quite hard to get a grasp on it.

[00:01:23.24] Ben: Michael, thank you so much for coming on the Structural Shifts podcast. I’ve been really looking forward to this because it has been delayed a couple of times, but we finally got to do the podcast.

Michael: Yes, thanks! It’s a great pleasure. I’m delighted to do it.

[00:01:36.27] Ben: I wanted to kick off by just testing a little bit the premise of your book, ‘The Levelling’. Is globalization really over? Or are we just in a period where it’s sort of temporarily in retreat?

Michael: No, I think it’s dead. I think it’s over. I think there are many people, I suppose, for good reason, whose fortunes are tied to globalization, who don’t want it to be over, who deny its passing. Maybe it’s been dwindling as a force — in the last two years, we have been storing up many of the side effects or the perceived side effects of globalization. And that, I think, it’s been dealt a fatal blow by COVID. And I should say, I’m in favor of globalization. I mean, it’s done so much good. Billions of people have risen out of poverty, it’s transformed cities like London, Dubai, etc, it has given us so many technologies. So, I’m in favor of globalization but my reading, as it were, of the situation is that it’s dead and going. We’re moving on to something else now.

[00:02:50.15] Ben: And how objectively can we show that it’s over? Because, again, might it not just be changing form because, if we look at financial flows or trade flows, sure, we can show that it’s in retreat, but what about things like the flow of data? Is it not just becoming more digital?

Michael: I think there’s maybe three things just to bear in mind. One is that globalization is ephemeral, it’s in the ether, in the sky, in the way that people tend to look at it. There’s no ministry for globalization. So, it’s quite hard to get a grasp on it. We do have the benefit of history that we’ve had a wave of globalization from 1870 to about 1912, which looks very like what we’ve seen in the last 30 years or so. And that wave of globalization came to a juddering halt with economic crises, nationalism, etc. So, all of the warning signs are there. And then, thirdly, if I debate globalization with people, what I try and do is bring that debate down to indicators — the movement of people, the flow of ideas, trade — and all of those have been coming to an end, they have been cut off in different ways. So one example I give is that globalization began with the fall of communism, the opening up of Eastern Europe — not just economically, but democratically as well — and now we have events like the shutting of democracy in Hong Kong, which is the bookend to what happened with the fall of communism. You mentioned digital globalization. I think that’s quite interesting because tech and digitization have played a really strong role in globalization but the channels of digitization are being funneled in different ways. So one example I give you is Google in the early 2000s had about, I think, a third of the search market in China. Now it has close to zero. So, what we’re actually seeing is we have more digital activity, but it is becoming more regional. You look at TikTok as another example — potentially global company now being shuttered in terms of how it can be used not just in the US, but also in China and other parts of the world.

One of the problems we have, there really is no Minister or Prime Minister for globalization. It is an interconnected, interdependent activity, and many of those interdependencies are breaking down. In our world, the role of the nation state is still very, very important —Michael O’SULLIVAN

[00:05:18.01] Ben: Why do you think is over? Because as you yourself said, it’s had massive, massive positive effects. It’s a positive that raises economic activity and wealth for everybody. So why do you think it’s overall coming to an end?

Michael: I think that the prime concern I would have is that the economic engine of globalization has slowed. So, in many emerging countries, the rate of growth is slowing. In the developed world, productivity has been slowing — notably so in countries like the UK, where it’s at multi-decade lows. The financial side effects of globalization, the negative ones have been rising, so the world is becoming more and more indebted, which will slow future growth. And then, there’s a range of economic problems that people associate with globalization, such as inequality. To my view, it has really, nothing to do with globalization, but rather the way individual countries have harnessed it. So if you look at the most globalized economies in the world — Ireland, Netherlands, etc. — the income inequality in those countries is actually reasonably well managed, because they use tax to distribute the benefits of globalization. If you look at the US, where inequality is really egregious, they have not used their tax system to spread the benefits of globalization, and that creates discontent with globalization,

[00:06:52.07] Ben: Had globalization been better managed, then it wouldn’t be coming to an end?

Michael: I think that’s largely true. And I think one of the problems we have, there really is no Minister or Prime Minister for globalization. It is an interconnected, interdependent activity, and many of those interdependencies are breaking down. In our world, the role of the nation state is still very, very important. I think COVID is a great example of this: different countries have managed and digested globalization in very different ways with different consequences.

The post-globalized world order will be a multipolar one. What I mean by that is you’d have at least three big regions: China, Europe, the US — for the reason of their size will be dominant but also who will do things increasingly differently — Michael O’SULLIVAN

[00:07:35.16] Ben: You talk quite a lot about Brexit in the book. I suppose we should see Brexit as being part of something bigger, i.e. the end of globalization. But can we really draw that conclusion? I mean, could Brexit not just have been a political mistake, a referendum that should never have happened, a unique set of circumstances? You know, the refugee crisis, etc. Can we not just see Brexit in those terms? I mean, does it have to be read as part of this end of globalization?

Michael: That’s a very good question. To give context, we’ve had two big waves of globalization. The first of the early 20th century, which was led by Britain from London. The second was an American wave of globalization. So, both waves have been led by Anglo-Saxon countries. And the two main Anglo-Saxon countries — the US and the UK — are now in political crises. That much is clear. And those political crises are also crises of globalization. In the case of the UK, I think there is an argument that the calling of the referendum, or the way it was constructed, was accident-prone and could have been done better. I think, however, the way I tend to look at globalization, it’s like a big block of ice that’s begun to fragment and the first really big crack in the world order was Brexit — something that many people thought would be inconceivable happened — and that respected was the first shock, and I think the first really big event that’s taken us into the post-globalization age. And why I’m convinced that it’s linked with globalization is that many of the underlying problems in the end of globalization argument — low productivity, inequality, a country dominated by one city, one elite — you find all of those in the UK. And I think what I’m also drawn to is this whole idea of the rise and fall of nations and the fact that nations go through cycles and globalization is doing the same. The UK is now in a cycle where it’s in, I think and I hope, it’s sort of a bottoming out phase. And in the next few years, we’ll begin to see things improve and be reconstructed. So there is that logic to it, I think.

We would probably feel the end of globalization more severely if it weren’t for central banks, but the cure is arguably storing up worse down the line — Michael O’SULLIVAN

[00:10:09.09] Ben: How does a country like the UK fare in a world that’s deglobalizing? Because, you know, it’s detached itself from a very large trading bloc and so, now it’s seeking, I guess, new alliances, new trading partners in a world that has become less interested in trading. So, how does that play out, do you think?

Michael: Yeah, it has a lot of risks. And one of the risks, I think, is that the discourse in Westminster and in London is not really focused on what’s happening to the world order. It’s really very inward-focused — and this is a risk that things shift outside the UK, and it doesn’t adapt. So I think that there are several things. The most important, I think, is that the post-globalized world order will be a multipolar one. What I mean by that is you’d have at least three big regions: China, Europe, the US — for the reason of their size will be dominant but also who will do things increasingly differently. So, if you look at the way the EU is imposing itself on the tech world in terms of regulation, that’s a good example. So the UK needs to ask, “Okay, first of all, do we want to be outside this very powerful block?” — and that decision is already made. So, then it becomes a relative decision: “How do we position ourselves between these big blocks?” And in particular, two of them, the US and Europe are very, very close political and trade partners. And my hope would have been that they pay acts to arbitrage these regions. It’s a sort of third independent, but not neutral party, whose upholding of the rule of law makes it more where people feel very secure to do business. So, in that plate, the declaration in Parliament that the UK would break the rule of law is extremely worrying. And I think the international consequences of this have not been thought out at all.

Coming out of the global financial crisis, central banks were like doctors who gave financial morphine to the economic patient — Michael O’SULLIVAN

[00:12:17.07] Ben: What role do you think central banks are playing? I mean, do you think their role is helpful or not? In the sense that are they smoothing the transition period by cushioning economies from some of the worst economic effects of the transition? Or do you think they’re prolonging the transition and exacerbating some of the ills of the current age, such as wealth inequality?

Michael: Yes, this is a very good question — quite a complex area. We would probably feel the end of globalization more severely if it weren’t for central banks, but the cure is arguably storing up worse down the line. So the way I tend to phrase that is that coming out of the global financial crisis, central banks were like doctors who gave financial morphine to the economic patient. You know, a doctor, if you’re injured, he or she may give you morphine for a couple of days to take the pain away, but they won’t do it every day for 10 years. And that’s what we’ve had. And to that end, markets, investors, many other players have become dulled and stultified to the reality of economics and what’s happening. So, we’ve had no inflation, except in asset prices, valuation for government bonds for many equities, the technology equities are at all-time highs. What we have is we have bubbles in financial markets, which potentially rob future generations of the returns they will need for their pensions. And I think, at the same time, what the cover of central banks has done is to make it much less urgent for politicians to address underlying economic problems. So, if you look at Europe, there’s actually been relatively little reform on things like capital markets union, banking reform — all these things that were super urgent nine years ago, where we had promises from finance ministers that they would be addressed — pretty much nothing has been done.

[00:14:27.09] Ben: Yes, slowing down reform. And also, I guess, it’s also removing some of the market censoring that happens to politicians and to political actions, right? So, you know, you might argue — I’m sorry, I don’t want to get into counterfactuals, but you might argue that in the run-up to Brexit, there would have been harsher market movements, because the central bank was intervening.

Michael: That’s absolutely true. And I think it raises many legal questions, too, as to what end should central banks go in terms of trying to intervene in markets and economies? And you now have a situation where central banks are trying to mandate themselves or justify themselves on a whole range of criteria. The European Central Bank is now adopting the mantle of the green economy; in the US, the Congress has proposed that the Fed do everything it can to reduce racial inequality, which is a just cause, but it’s much, much better left to politicians and lawmakers than central banks. I don’t know how they would go about doing that. What we want in democracy is we want parliaments and governments to address these problems, not central banks to, if you like, swamp the whole political economy in terms of what they’re trying to do.

[00:15:57.16] Ben: If we break those things down, then, so first of all, how do we sort of de-politicize central banks? Because as you say, it’s absurd to think that the body in charge of monetary policy could affect racial inequality. So, how do we de-politicize central banks?

Michael: So, the world went through — from the ’70s onwards, many central banks were politicized in that their governors or presidents had, in some way, ties to the governments who appointed them. And then we went through a phase where central banks were trying to crush inflation, which was politically very unpopular — the best example was Paul Volcker in the US. And we’ve gone through an era of independent central banks who are run by technocrats, civil servants, independent from politics. And that is beginning to change, I think. There is a sense that, maybe in the States, that the heads in the central banks are somehow siding with the mandates of individual governments. It’s been the case certainly in Japan. And I think that there’s several things we can do. I think the appointment process for central bankers is important, and also the people who populate the committees of the central banks; stressing their mandates, and narrowing their mandates is also very important. And maybe, I think, in future what we need — we probably need another crisis to get over this — is a curbing of extraordinary powers, like quantitative easing. That, when it was introduced, was something that was considered off the charts and now it’s normal. So, there needs to be a debate about the extraordinary powers that central banks have.

I would prefer a total reboot. I think some institutions have remained relevant, like the OECD, because they’ve had a good sense to attach themselves to a bigger framework in the shape of the G20. Others, I think should be completely reshaped — Michael O’SULLIVAN

[00:17:52.21] Ben: Yes, that was gonna be my second question, which is, how do we roll back the balance sheets? How do we de-leverage the central banks? Because I think in the book — you know, the book is already out for a year, so they’re even higher, particularly COVID — but you said they’re as high as they’ve ever been since the Napoleonic Wars. So how do we de-leverage central banks and get back to some sort of sensible level of leverage?

Michael: So the background picture to this is that world debt to GDP levels have been rising, they’re passing out the previous high of the Second World War, on course, as you said, to hit the highs of the period around the Napoleonic Wars. And that’s quite extraordinary if you think of the events associated with that. One of the reasons debt is rising is because interest rates are so low, governments, companies find it very, very easy to borrow and, in particular, government debt has been hoovered up by the central banks. Now, there’s a number of ways of walking back from this. One is the enlightened approach where policymakers decide, “Look, there’s just too much leverage in the world system. We need to pare it back.” And that’s done in a collaborative way across central banks. I don’t think that’s going to happen — that would be too ideal.

[00:19:20.20] Ben: This is what you refer to as the ‘New West Failure’, right?

Michael: It is. We need a reordering of finance in the world, along the lines of the Westphalia Treaty of 1648. That’s a very grand example. But, you know, the size of central banks, the size of debt in the world, is the biggest in centuries. So, we do need a grand setting for this. I certainly think that we are storing up the ammunition for the next financial crisis which will be a crisis of debt and ‘debtness’ and it will probably come about by, you know, additional QE from central banks not working or not functioning, a loss of confidence, and then, for the first time in some time, people begin to look to get the money they’ve lent back, either from governments or countries, projects, companies — and it’s just not there. You get a deep recession. Maybe central banks will try more QE and that won’t work because often, in a credit crunch, monetary power can — in a deep credit crunch, monetary policy can lose its power. So that’s sort of a negative, nasty scenario, which I don’t like to paint, but in the absence of enlightened policy, that’s a potential route for the future.

[00:20:51.26] Ben: And I can’t resist asking you this — I know we’re gonna go slightly off-topic here — because you used to be a CIO, so it’s too tempting not to ask this. But how does one position a portfolio for a world where there’s a massive governmental debt crisis on the horizon?

Michael: Yeah. So, I think you do a number of things. The first thing, I suppose is to figure out what really is a safe asset. So what government debt will you feel really safe holding? I mean, that’s why German bonds, for example, continue to trade-off at a negative yield, because it’s one of the true safe assets — maybe Dutch debt as well. So, if I had a portfolio and there were private equity in it, for example, I would switch that into things like distressed debt funds, long/short credit funds. I think what you can begin to do, as well, is to look at tail-risk strategies — that’s quite a technical term, and what it means is that these are strategies that would form a small part of the portfolio, but if an extreme, negative event happened, they would pay off quite handsomely, they would tend to be kind of derivative-based or derivative-type strategies. And then also, I think, you know, in other parts of a portfolio, you want to aim to hold companies that have got a lot of cash and less debt. Ironically, some of the big tech companies fall in that category, even though they have somewhat stretched valuations.

[00:22:33.13] Ben: So, just coming back to the book and the end of globalization, how do we get back to sustained organic growth? Because some of the things that you talk about in the book have become, I guess, slightly passé or old fashioned, which is you advocate for things like education and rescaling and things that don’t seem to be at top of the political agenda right now. So, is that what we need? Do you think we need to get back to education and re-boosting productivity? Because that’s the only route back to sustained organic growth.

We can continue with globalization as it is, which I think is unlikely. We can have a lapse into chaos and disorder, like the 1910s and ’20s. Or I think what’s much more likely is we’ll have a multipolar world, which I think the way to look at it is that it’s not so much a world dominated by big regions, as to the fact that these big regions do things increasingly distinctly or differently — Michael O’SULLIVAN

Michael: I think someone said — maybe it was Paul Krugman — for developed countries, in the long run, productivity is everything. And I agree with that because it’s the main — unless you’ve got rapidly changing demographics — it’s the main and perhaps the only way to boost economic growth. And I think the first step is that there needs to be a debate and a movement amongst governments to focus on productivity and focus on what I would call ‘trend growth’. So, what’s your sort of trend level of growth? And the best way to improve that is through productivity. As you said, many of the things that drive productivity — you know, rescaling education, careful investment in technology in country strategy — they have been forgotten. If you look at what’s happened to education and educational attainment rates in the US, that’s quite alarming. And many of the other factors around that, things like human development, are regressing. So you’re right. I mean, they’re worse than passé. They’re just being degraded. And there’s really no other way for developed countries to grow in the long term. You can do trade wars, you can try and convince countries to reshore investment, but that’s all one-off kind of short-term stuff. The real driver of growth is through productivity. I think the problem many politicians have with this is that we are in short, political cycles. So, by the time you’ve invested in the factors that drive productivity, you only see the benefits of that maybe 5–10 years down the line.

[00:24:58.03] Ben: I wanted to ask you about the US elections in November because you make the point in the book, you sort of draw the parallel with the 1980s and you say, you know, people were fixated with the end of communism, and they felt as bought what was coming next, which is this period of rapid globalization and internationalization of trade and everything. And to some extent, you say the same thing here. Now, we’re obsessed with a little bit the political circus, and we’re missing this bigger shift. But to what extent is Donald Trump and this US election, a circus versus being really important in determining how we transition to this new world, whether it’s a smooth transition, an elongated transition, etc.?

If there’s any doubt that globalization is over, four more years of Trump will entirely smash that — Michael O’Sullivan

Michael: He is very important. And I think what is important is that, you know, Donald Trump as an individual has a lot to be responsible for. But he has come at a time when that’s maybe apt in that it’s the end of an era and he is the human wrecking ball breaking down the old order. He is not going to be the person to build up the new order. And he has also been enabled by many people principally on the Republican side for whom it’s convenient to have him as President. I think it’s quite clear that, you know, four years of Donald Trump have broken many things. He’s broken America’s diplomacy with Europe, with parts of the Middle East, but not all; obviously, broken diplomatically with China, and sown general confusion. He’s also, I think, devalued many of the institutions in the US — the State Department, some of the financial institutions as well. And then, in that background, four more years of Donald Trump would make permanent all of these ruptures and it would cause the rest of the world to maybe ignore America. Europe in particular will tend to go its own way. I think countries like China will be more and more convinced that the US is in turmoil, and it is weakened, and Russia will probably have the same view. And they will act accordingly. And I think that will well and truly break globalization. If there’s any doubt that globalization is over, four more years of Trump will entirely smash that. Joe Biden, for his part, I think will not be a transformative president, but a restorative president in that he will restore the status quo, he will repair the State Department. His team, I think, is very strong on foreign policy, maybe less so on the economy. And he will, I think in particular, restore relations with Europe and try and rebuild the partnership with Europe, to the detriment, I think, of Russia, and maybe China as well.

[00:28:10.06] Ben: But I guess the point that many people would make is, as you said, he’s a wrecking ball. He’s gutting institutions domestically, globally. But, in a way, is that necessary? Do we need to break these institutions in order to remake them? You know, would a Biden presidency as you said, restore the institutions or, you know, other institutions still fit for purpose? So, I suppose the question is, do we need four more years of Donald Trump to hasten this transition to the new world? Or would it be a disaster and lead to a much, much more uncontrolled transition to the new world?

The level of denial amongst governments and companies that things can go on, as they were, I think, is still very, very high, and that denial needs to be broken — Michael O’SULLIVAN

Michael: I hesitate to recommend four more years of Donald Trump, but a lot of what you’re saying, which I accept, which is that we’re in the midst of a paradigm shift, which is a very overused phrase — and you only get paradigm shifts for a number of decades, maybe centuries, and you only get the rebuilding when a lot of the old order is broken down, there is chaos, and then people come forward with new ideas and new initiatives. And the reason I think we need more breaking is the level of denial amongst governments and companies that things can go on, as they were, I think, is still very, very high, and that denial needs to be broken. So, if you take as an example the debate over the World Trade Organization, which is potentially a defunct and irrelevant organization in the context of a multipolar world, the debate now is to whether we just have a new leader and someone who’s not from Europe, or the States and everything would be fine. And you have similar debates about the World Health Organization, the World Bank, and not about, you know, should these institutions be radically changed or displaced? And what institutions of the future do we need to have?

[00:30:19.06] Ben: And do you have a view on some of that or not? So, you know, I don’t want to run through the list of all the international organizations like the IMF, and so on. But, are some of them still capable of doing the job we need them to do or do you think it’s a total reboot?

Michael: I would prefer a total reboot. I think some have remained relevant, like the OECD, because they’ve had a good sense to attach themselves to a bigger framework in the shape of the G20. They’re sort of the think tank of the G20 now. Others, I think should be completely reshaped. The World Bank, I think should be relocated physically to Africa, which is the one part of the world that really needs the help of the World Bank. And then, I think there needs to be a debate on what the institutions of the 21st century need to be. Do we need, for example, an institution on climate change that has got the power to fine governments and companies in a forceful way for climate damage? Do we need an institution to govern cyberspace and cyber warfare and cybercrime? We don’t have that. We don’t have an international police force for the internet. So, there’s all these things that are beginning to crop up as future problems, and have not yet been framed either in philosophy and law or by institutions.

[00:31:48.20] Ben: Moving to what comes after globalization. So you’ve used the term ‘paradigm shift’, and you’ve also used the term ‘multipolar world’. Is that what globalization gives way to? Does globalization become a multipolar world?

Michael: I think so. I mean, there are many scenarios. We can continue with globalization as it is, which I think is unlikely. We can have a lapse into chaos and disorder, like the 1910s and ’20s. Or I think what’s much more likely is you’ll have a multipolar world, which I think the way to look at it is that it’s not so much a world dominated by big regions, as to the fact that these big regions do things increasingly distinctly or differently. So, again, coming back to the internet — I mean, the US has got the big internet giants, and they are stockmarket monsters. China has cut off its internet, but it has a thriving e-commerce sector. And then Europe doesn’t have any of these big internet companies, but it is becoming the regulator of the internet, looking to protect people from the ills of the Internet.

We’re in a world where lots of issues are emerging and lots of well-placed frustrations are being vented. So, it is definitely a period of turmoil, where democracy and rights and liberties are being contested — Michael O’SULLIVAN

[00:33:12.23] Ben: For how long can it seek to regulate the internet without those platforms? Because, to use an analogy from a previous podcast, it’s sort of trying to control the seas without a Navy or an infrastructure, right?

Michael: Yeah. It can because of its size and the power of its economy, and the fact that Europe in particular is very, very sophisticated in terms of policy, and regulation. So, if Britain were to try and come up with its own set of rules for the internet or how British people consume the global Internet, it may well not be able to do so. But Europe is obviously much bigger. I think the way we’re going now is we’re going towards a values-based, multipolar world. So what I mean by that is that when it comes to economics and politics and climate change, each of the big regions has got very different values and approaches- and those values will inform how they build out their economy. So, Europe wants to protect its citizens and their data. It’s also very strong on the environment, and many of the new policy plans we hear about from Europe, are focused on the green economy. Much less so in the States where there’s just a very different balance between society and the economy. And then China has its own very distinct set of values, which I think we don’t spend enough time trying to understand, in the West. And I think it has its own risks and its own complexities which are maybe not readily apparent in newspaper headlines. And if you look at the Chinese Communist Party, which is a very big machine, inside it you have lots of different groups and rivalries that effectively mirror what you have in western politics just that they’re all under the same big disciplined umbrella. But I think in China, the secret is not just the vision, but also the implementation. They can implement policy for such a big country in a very, very speedy kind of way.

[00:35:38.00] Ben: The gravitational poles are US, China, Europe. Will countries that aren’t in those countries or in those regions have to choose between them? So, you know, will Africa have to align itself with the USA? Or will India — and it looks increasingly like India is already aligning itself with the US — is that what it will be? It’ll be a question of choosing between one of these poles for everybody else?

Michael: Yeah. So, I think there’s maybe a few things here. So, I think countries who fall between the poles — so, Japan in Asia, Australia, they are between America and China; the UK, Russia, in Europe — they’re on different sides of Europe, obviously — they will potentially find life more difficult because they’re not as big as the three poles and they’re sort of mid-sized powers and they have to come to terms with that and reshape their own identity. I mean, Russia has its own crisis in that it’s a military power, but not an economic or financial power. I think other parts of the world are in the stream. India, if you add it on to the area of the Emirates, is potentially in time, another pole, but it has a lot to travel in terms of its development in order to get there. And then I think for countries like, you know, Nigeria, Bangladesh, who are populous and growing, I mean, they have lots of choices. They can sort of say, “Well, let’s co-opt ourselves to China and the Chinese model, or do we still follow what the Americans have done? Or what the Swiss or the Irish have done?” Or they can kind of say, “Well, look, we just do it our own way.“ So, I actually think that these countries — there’s maybe 10 of them — you know, big, populous countries in Asia and Africa, who have not yet really globalized or developed, where they go in the future and how they do it, and what templates they use.

[00:37:35.14] Ben: And you see them as potentially in a strong, almost bargaining position or arbitrage position, do you think, between those poles?

Michael: Not yet. I think what they need to do — and this is where some of the new institutions of the world order would come in — is that these countries perhaps need to collaborate better in terms of building up their own cooperative institutions so that they collectively have more power vis-a-vis the US or China, as it were.

[00:38:05.16] Ben: You’re a big fan of small countries, right? Because you just used the example of Switzerland and Ireland — so, small countries that are very globalized in the sense that they’re magnets for certain kinds of information, trade flows. Are you still as bullish about those small countries in this sort of de-globalizing world?

Michael: I am, obviously, as an Irishman who’s lived in Switzerland, quite biased. And how I got involved in the whole globalization debate is I wrote a book years ago on Ireland and globalization — at the time, Ireland was the most globalized country in the world. I’m still bullish about them because I think this has been shown by the Coronavirus crisis, is these countries in general, they have a resilience, and they have a robustness. So, by virtue of being small and being open to world trade, they’re also very much aware of what’s going on in the world and what they need to do to correct against some of the imbalances that they will suffer. And the countries I have in mind are the likes of Sweden, Switzerland, Norway, Ireland, Singapore, etc. They’re all very different in terms of their culture and politics, but they have similar problems, but also similar ways of dealing with them. They tend to be the best countries who have this mix of strong rule of law, very good institutions, trust in experts, investment in education, etc. So, that model for me is still one that’s really relevant today.

[00:39:52.01] Ben: And how do you align that or reconcile that with also being a fan of supranational bodies like the… I mean, would it not be a good thing for power then to be devolved down increasingly to sort of nation states or even cities within nation states. Because if small is better, small is more agile, if small is more resilient, then is now an argument for fragmentation?

Michael: I mean, you can call it ‘fragmentation’, you can call it ‘devolution’ as well. I think what you may see is you may see countries like France become a bit more devolved to its regions in terms of taking some of the power that’s concentrated in Paris and devolving that to some of the regions. That doesn’t mean that France is going to break apart or that Europe is going to break apart — quite the opposite, because all of these countries are still happy to have the umbrella of the EU and to enjoy its laws on data protection, etc. I think, also, that globally, we do need to look at governance. So, climate change is my favorite example here, where many countries signed up to the Paris Accord, but it doesn’t really have teeth and it’s not, in my view, really contributed anything in terms of lessening climate damage. What you find is many of the big cities in the world, however, are much more progressive, and much more green than their individual governments. So, my suggestion would be that you have a sort of Paris Accord between big cities that it has teeth. They can, for example, tax their hinterlands, they’re better in control of pollution in their economic hinterland, they tend to be more advanced in terms of infrastructure and green policy, than individual countries. So that’s maybe one way to look at governance, I think.

we are beginning to see a fracturing of political systems where dominant parties are being pushed aside — Michael O’Sullivan

[00:41:56.07] Ben: Well, practically for other issues, because you still need a sort of supranational accord, that the cities sign up to. Would you need the same thing when it came to cybersecurity, or what else is coming down the road — you know, genomics, or digital currencies? How would you simultaneously have global accords in a de-globalized world with city states taking more responsibility?

Michael: Yeah, so I think cities are probably apt for climate change. I think something like cyber, you know, you need a cyber accord probably between at least five countries — and most of those will be on the UN Security Council. They are the cyber powers of the world plus a bunch of others because they are the ones who are either initiating or defending many of the cyber attacks and cyberwars. You know, this is an area of activity where there are no rules. So there’s no template or rule that says, “If Russia hacks me, I’m allowed to fire back a missile, because hacking me or hacking my hospitals is a declaration of war. It’s an aggressive act.” So, I think you need increasingly to match the institutions to the locus of the problem, the locus of the issue.

[00:43:24.06] Ben: Who are The Levellers? And why do they matter? And one of the related terms is ‘The Agreements of the People’ — what are the agreements of the people?

Michael: Okay, so the Leveller is a somewhat obscure story, but probably one of the most important in British history. And I’m guessing many people haven’t heard of them.

Ben: I used to live in Putney, and I’d never heard of them.

Michael: So my starting point is that we’re in a world where lots of issues are emerging and lots of well-placed frustrations are being vented — you know, from gender inequality, you’ve got black lives matter, and you’ve got then democracy-based struggles in Belarus — and you can go on for quite a long time. So, it is definitely a period of turmoil, where democracy and rights and liberties are being contested. In some respects, I’m not a fan of a lot of this political debate on Twitter, and I wish a lot of it were more constructive and that they were more constructive channels. I look at a lot of these protests movements, and you ask yourself, you know, can we go beyond protesting? And how would you take a movement forward and make it concrete and begin to embed it in changes and in laws?

Michael: I recall reading a couple of books some time ago on the Putney Debates, which happened in the middle of the 17th century in Britain. So the king was captured by Cromwell’s New Model Army and with the king being captured people had an inkling of what a parliamentary democracy might look like and they began to debate this down in Putney — it was called ‘the Putney Debate’. And it was primarily a debate within the army, the soldiers, and the officers. And one very important cohort was a group called ‘The Levellers’, who effectively were kind of the social democrats of the day. And they and their leaders came up with a template called ‘The Agreement of the People’. Quite a short template, but it’s really, in tangible form, what the people wanted from government and from parliament. And I think that’s missing today. And that needs to be reconceived.

Michael: For lots of somewhat bizarre and interesting reasons, these agreements of the people were written down, but then not fully transcribed or rediscovered for another 200 years. So, it’s an example for a while lost to history, but it was the first popular expression of what a constitutional democracy would look like, at a time in Europe when Europe was just beginning to throw out some huge innovations in politics and nation states. So, a very important time. And I actually think that if people today could look at these agreements of the people and use them as a rough template for what they want — because they were very practical. You know, they talked about people being treated equally by the debt courts, they advocated limited political terms to reduce corruption — so, foreshadowing many of the things we have today. And I think the use of a template would be taking many dissatisfactions with politics and many movements and making them concrete and also constructive, which I think is largely missing.

[00:46:59.27] Ben: I really like your idea of getting off Twitter, because, first of all, you know, Twitter tends not to lead to anything that’s particularly concrete. But also Twitter tends to lead to massive bifurcation which is, it’s very difficult to hold the center ground on Twitter, because it’s the extremes that gain traction on a platform like Twitter. But I suppose the question is, how is this so different from electing representatives with well-defined mandates in the first place? Are you just saying that we, as constituents, would put forward our ideas, and then almost the political parties would be formed to then put those into statute and into place? Would that be the difference then? So, rather than the political class coming up with the manifestos that we then vote on, it becomes much more bottom-up?

Michael: I guess it should ideally be more bottom-up. I think there’s a sense also that many political parties are somewhat jaded in terms of what they represent. And maybe one question, which I think we haven’t seen yet, is whether we get new political parties, new political entrepreneurs coming through. And I think there is dissatisfaction — you know, what we are seeing as well, we are beginning to see a fracturing of political systems where dominant parties are being pushed aside. You’ve seen that in France, the two dominant parties pushed aside. It’s interesting, we haven’t seen it in the UK or the US yet; the two-party systems remain dominant. But I think there needs to be channels constructed for taking, you know, what I think are well-founded grievances, and getting those into law. For example, there’s a lot of work now being done in social media, such as petitions for parliament in the UK — a lot of work being done there, in social media to get people’s petitions into parliament. So there are changes beginning to come through, but certainly not as fast as I thought would happen.

[00:49:10.03] Ben: I like that idea, too, because I think there’s sort of contrast between Twitter — again, to use that analogy — and the way we do politics it has now become enormous. And the politics hasn’t kind of responded yet. Because, you know, I think what Twitter is doing is giving us the sense that we know as much as the political class. I think we really saw that during COVID, which is everybody thought they knew as much about how to cope with this crisis, as the political class did, because it gives that impression of the narrowing of, or this sort of information asymmetry disappearing, which is an illusion. But I think what would be good would be to sort of reboot politics to be more like social media, in the sense that, as you say, we could treat it like a platform, which is, you know, we could contribute all of our ideas, and then you could build new political parties on top of a platform, which would then be much more networked and responsive to changing ideas. Is that what you had in mind, then, that it will become more networked and responsive?

Michael: In a way it is. And again, one lesson I remark on in the book with The Levellers is that they were idealistic. And parties are a result of that. I mean, they were very good at things like pamphleteering. So they were the social media geniuses of their time, but they were totally outmaneuvered by the incumbents. So I think there is a cynicism required as to how the political system works, which needs to be matched with idealism and a desire to change things. You know, I use the example of Emmanuel Macron, who is seen as being a revolutionary political figure. But I think what he figured out was that the best way to do a so-called ‘revolution’ is to take the system from the inside, not the outside. Now, he had the help of many parts of the system and the institutions in doing so. But it seems to me that that’s sort of a speedier way to changing politics than trying to do it from the outside.

[00:51:18.15] Ben: I can’t resist asking the question, which is, how satisfied are you with Emmanuel Macron? Somebody living in France? How much do you think he’s really changed politics? Because he seems to be a bit like Obama or a bit like Donald Trump. He used digital means to campaign in a completely different way. But the governing has been almost very traditional. Do you accept that?

Michael: It has been traditional in that he — I mean, he has replaced one elite with another younger one, for sure. And pretty much all the people around him reflect the fact that France is still very much elitist in terms of politics, in that they all have the same formation, the men and women have the same views, education. I mean, it’s more stark than, say, the UK is. So he and the people around him are still very much a product of the French system, the French elite. They’ve just, I think, made a lot more fresh. I mean, I think there are areas where he stands out. I don’t think in any way he’s corrupt. I think he’s absolutely sincere in what he wants to do and he’s very, very driven, in that sense, in terms of implementing his vision. I think what he has changed, in my view, compared to the two previous presidents — Sarkozy, and Hollande — I think, with both of those, there was a sense that they were kind of filling a gap and that they might not be around for the next four or five years after. Whereas Macron, I think people have a sense that perhaps he will be here for another term, and that we will have a Macron era that he has the time to implement changes. And he, I think, before [00:53:08.12] he had already implemented quite impressive labor market changes. France needs more of those. I think also, one area that’s clearly open to him is Europe and European politics. And certainly, the energy for European politics and political initiatives in Europe is very much in Paris. It’s not in Germany. We will soon have the post-Merkel era, politics will be a lot more fractured, which will leave France as the dominant country driving policy in Europe.

[00:53:46.04] Ben: If Brexit was the first sign of the end of globalization, is TikTok, the first sign of the post-globalized world?

Michael: Actually, it’s an important sign because it shows how a service that should, I suppose in many respects, be harmless and that should be global, and that people in many countries can use it, but has been used for political ends, can become carved up in the manner of this multipolar world. So, you know, in the US and China, there are clear barriers around the use and the ownership of different parts of TikTok, which begin to, at the same time, sketch out the map of this multipolar world that’s coming.

[00:54:47.12] Ben: Last question. COVID: do you think that this is, again, laying bare the fact that we have reached the end of globalization? Or do you think it’s a reason to be optimistic because it’s a crisis that all countries face and really should galvanize us to work together?

Michael: It’s certainly been a test — a dramatic test. And I think what individual healthcare companies, universities have worked together. The absence of collaboration between countries and regions, for me, has been the litmus test — a litmus test that shows that we are at the end of globalization and heading into a more singular, maybe more selfish, multipolar world. In previous crises, you’ve seen countries and governments collaborate — global financial crisis being an example. And we just had an absence of that this time. They’ve squabbled over vaccines and masks, etc. So that’s the lesson that needs to be borne in mind.

[00:55:57.08] Ben: You would just argue, then, that COVID — again, exemplifying what the new world would look like — at the same time is probably bringing it forward faster?

Michael: I think so. I think it accelerated this whole thesis, for sure.

Ben: Michael, thank you very much indeed for coming on the podcast.

Michael: It was good!

Strategy in the Post-fixed Costs Economy

October 2020

Evaluating strategic options in a world where businesses have never been easier to start, but never been harder to scale


We’ve talked often about the diminishing importance of supply-side economies of scale. In its simplest expression, digitization flips the industrial age equation. What was scarce in the industrial age was supply; what is scarce in the digital age is demand (attention).

In the industrial age, to scale supply meant mass production to spread the fixed cost of large capital investments over large volumes. And the industrial age was an age of mass produced, relatively standardized goods. This applied to goods and services provided by the private sector, but also to state-provided services, such as education and public services.

Since the advent of the internet, this is changing. We first noticed the shift in industries where both supply and distribution could be digitized (e.g. media) because supply became abundant faster and this highlighted our limited attention sooner.  But it’s becoming increasingly apparent that all industries are being disrupted as software has eaten the world. More and more physical goods have software components to them, making supply more digitized. Where supply cannot be digitized, distribution nearly always can. And where supply-side economies of scale remain important, they can be borrowed.


Renting scale and the end of fixed costs

AWS was not originally intended to be a platform on which third-parties would run their businesses. But, thanks to Amazon’s business model, it was possible to open up that infrastructure service to others. In doing so, Amazon created a massive and highly profitable business which contributes 65% of group operating profits.

But, as big as the impact has been on Amazon, the broader societal impact has been truly dramatic.

Before AWS, companies had to make large upfront investments in computing hardware – a significant barrier to entry. Even buying hardware was a major source of risk: buying too much could bankrupt a company while buying too little could cause a major bottleneck to growth. And so, AWS removed both risk and cost for new businesses. As a direct consequence, it also contributed to the creation and success of hundreds of thousands of new businesses.

This boost to global GDP over and above the value captured by Amazon itself is difficult to calculate, but it’s certainly very significant. It’s probably not an overstatement to suggest, as Charlie Songhurst does, that AWS has been the single biggest factor in the rise of angel investing.

But AWS is not the only internet era platform. From Shopify to Stripe, examples abound of platforms that share their scale economies to remove the cost and complexity of doing ecommerce – allowing companies to form and start trading faster, with reduced risk and at lower volumes than would have ever been possible.

And not all internet-era platforms are providing digital services. As Rita McGrath, Columbia Business School Professor, discussed on a recent Structural Shifts podcast, by helping establish prices and create trust, digitization is making more and more non-digital assets tradeable on marketplaces. As she put it:

“What we’re seeing with the advent of the digital economy is that more and more transactions can be conducted in markets that used to require a firm.”P

Uber was a pioneer in this regard, but we now see this “uber of x” phenomenon everywhere – even in the enterprise market. At aperture, we are very much an embodiment of this, providing strategy and go-to-market as-a-service that enables companies to avoid the fixed costs and risk of underutilization and underperformance in these functions.

In effect, it’s becoming easier to rent all services, physical and digital. All become liquid and on-demand. Capex gives way to opex or, as Younes Rharbaoui says: we have entered the post-fixed costs economy.

“Signs of a post-fixed costs economy are all around us: companies switching to full remote, increased reliance on independent workers & freelancers, on-demand software where cost matches usage, are all creating lean financial structures for growth.”

And, of course, like many other secular trends, the impact of the pandemic has been to accelerate it. If COVID-19 drew a binary distinction between online and offline services, lifting the former and sinking the latter, then it was disproportionately brutal in its treatment of those offline businesses with high fixed costs – oil companies, airlines, hotel chains and so on. From now on, all fixed cost investments will be more heavily scrutinized and, where they exist, variable costs alternatives will be more actively considered. Even Warren Buffett, a regular character in aperture blogs, is starting to consider the wisdom of some high fixed-cost business models.

The fact is, if fixed costs were already becoming passé, they definitely will be in the post-pandemic world – ushering in a faster transition to a new, internet-era economic structure.

Platforms, aggregators and the long tail

For the best explanation of the difference between platforms and aggregators, we recommend this classic essay from Ben Thompson. In it, he uses the Bill Gates platform definition, namely that: “A platform is when the economic value of everybody that uses it exceeds the value of the company that creates it. Then it’s a platform.”

On this definition, the business we have already mentioned – AWS, Stripe, Shopify – are all platforms. They make the large investments in fixed costs – datacenters, fulfilment centers, payment networks, etc. – that mean their clients don’t have to. They make it cheaper and simpler to do business.

It’s this commoditization of supply and associated end of fixed costs that’s now starting to give rise to a long tail of providers. Thanks to lower and variable input costs, it’s possible to make money at lower volumes than in the past, which in turn means a higher number of providers can co-exist.

Take newspaper publishing, for example. The massive costs of producing and distributing physical newspapers gave rise to significant economies of scale and produced an oligarchical market structure. Compare that with today, when a platform like Substack allows independent writers easily and cheaply to publish, distribute, and monetize paid newsletters. These writers can make a living with only a small audience, allowing potentially tens of thousands of them to co-exist – and giving rise to a broader phenomenon, the “Passion Economy”, where more of us can pursue our craft or our talent and make a living from it.

However, the difference between the long tail as it’s conceived now and the original theory, is that supply is abundant, not demand. The constraint on all digital-era businesses is demand and the gatekeepers of demand – the most profitable actors in the digital ecosystem – are aggregators.

In a world of abundant supply, aggregators help match buyers and sellers. They are in a position to do so because they provide interactive content that rises above the noise to command our attention. When in possession of our attention, they can monetize it by charging advertisers to reach us. This has become the biggest cost for many online companies, accounting for 40% or 50% of the investments they make in growing their business.

As Clayton Christensen predicted in the Law of Conservation of Attractive Profits, as one part of the value chain commoditizes, the value is captured elsewhere. As platforms helped generate an economic surplus, aggregators increasingly captured that value – especially Google and Facebook.

While it has become cheaper to start a business, a sharp increase in customer acquisition costs more than offset these savings.

More precarious

But it’s not just high customer acquisition costs that prevent long tail companies from rising to a size where they exploit scale effects. There are other factors at play.

First, the falling costs of starting a business is a double-edge sword. If one online retailer can set up on Shopify, so can any other. Platform companies are lowering the barriers to entry for everyone, making it harder to defend a business than in the past.

Second, if a business model has network effects and a company can grow large enough to exploit them, this market leader becomes more powerful than an industrial-age leader. This is because, unlike supply-side economies of scale, demand-side economies of scale are subject to increasing returns to scale; the more they exist, the stronger they become. And so, where the industrial age gave way to oligopolies with clearly defined industry boundaries, the internet age gives way to winner-takes-most aggregators, large-scale platforms, and a long tail of suppliers operating across the economy in general.

Third, there’s also one important supply-side economy of scale which makes size important even in the absence of network effects – and reinforces them where they exist. This is the ad score. Basically, the bigger a company gets, the cheaper its relative cost of customer acquisition becomes because it pays a lower cost per lead thanks to a higher ad score (the algorithm that advertising platforms like Facebook and Google use to calculate the likelihood of a customer clicking on an ad).

Effectively, the post-fixed cost digital economy is one where is it simultaneously cheaper than ever to start a business, harder than ever to defend and scale it, and where the returns to scale have never been more important.

If it was hard to cross the chasm from startup to large, scaled business in the industrial era, in the digital age it is harder still.

So where to next?

Let’s look at strategic options for new entrants, from where to play to how to scale.

Where to play

Marc Gruber, a professor at EPFL and former podcast guest, wrote a book on “Where to Play”. At the risk of grossly oversimplifying the narrative, it argues that companies spend too little time thinking about which market opportunity to pursue – assuming a good product eventually finds a market – and instead provides a framework to select the right market before commencing activities.

Like Marc, we believe it makes sense to invest the time upfront to consider carefully where to play, even more so in the post-fixed cost economy. Marc’s book provides the methodologies for this choice, so we limit ourselves here to explaining the rationale.

In a digital world, where returns to scale are bigger, incumbents will be harder to displace. Therefore, it follows that any startup should focus either on creating a new market or, more likely, on market blind spots: the niches where consumers are underserved or overserved.

Underserved and overserved markets

In the digital age, underserved markets are likely bigger opportunities than in the past because geographical limitations are removed. A micro market in one country might be a big market when addressing all countries collectively.

B2B marketplaces are a classic example of this phenomenon.

While there are B2C marketplaces for seemingly everything, many entrepreneurs overlook B2B marketplaces because they seem less scalable. They think that it will be difficult to build a big business if your buyers and sellers are very specialized; that there won’t be a generalized pull effect. Buyers of fish are unlikely to be drawn to a specialist marketplace for, say, cement and building materials. But when you are dealing with a global B2B vertical, a two-sided network is more than sufficient to build a massive business: the wholesale fish market, for example, is worth USD150bn!

Another reason B2B marketplaces are sometimes overlooked is because it can be a long game. Many of the businesses we work with are patiently helping incumbents digitize their offering as a precursor to enabling one-to-many transactions. But their ultimate goal is to become a platform for enabling a many-to-many marketplace.

Trade Ledger Business Finance Lending Platform

There are also plenty of opportunities to target overserved customers. As Gary Pisano discussed on another episode of Structural Shifts, companies often push so far with innovation with an existing product or within an existing business model that they overshoot customer demands and leave themselves open to disruption from new entrants providing more user-friendly products or offering more convenience. He gives the example of subscription-based razor blade services like Dollar Shave Club disrupting the overengineered and expensive Gillette razors (“I can only shave so close before it’s scary!”). But this concept of overshooting is especially prevalent in B2B software where, in order to meet the enterprise buyer’s demands, traditional companies have overshot the demands of the end user creating the opportunity for disruptive innovation. This idea is brilliantly expressed in the following excerpt from an a16z article:

“Since effective top-down sales require a highly choreographed (and costly!) dance between pre-sales and the customer, product teams are incented to add more nobs to the product so these teams can sell more value and extract more dollars. Vendors get crossed off the list in vendor discovery if their product doesn’t check all the right boxes for the enterprise buyer, even if many of the boxes don’t actually deliver any value. This often creates a vicious cycle where more complex products give rise to longer sales cycles for more dollars, which then incentivizes even more complex products. For any user of legacy enterprise software, it doesn’t take long to realize that designing a seamless user experience is by no means a top priority for the vendor.”

Direct to consumer

Targeting underserved and overserved customers is what Clayton Christensen refers to as “disruptive innovation” and, as he tells us in the Innovator’s Dilemma, disruptive innovation is about simpler and cheaper products, but it’s also about marketing:

“disruptive technology should be framed as a marketing challenge, not a technological one”

This is even truer today than when Christensen wrote it because digitization opens up new routes to customer. As a result, product, monetization and customer acquisition have to align seamlessly around these new distribution opportunities.

The big trend is direct-to-consumer.

In the retail space, this mostly refers to the phenomenon of avoiding any intermediation – retailers or wholesalers or even any physical retail footprint – to sell directly to the consumer.

In the enterprise space, direct-to-consumer is different. Historically, it was not worth going directly to end users because they didn’t have much influence – or budget. Therefore, it was necessary to go through procurement teams and the choreographed dance of RFI, RFPs and workshops mentioned above.

But what is changing now is that technology products are not just sold directly, but are consumed directly. This makes software-as-a-service a much more disruptive phenomenon than people think: it’s more than a cheaper deployment method, it is a way to circumvent the central buying function and reach the end user.

In this context, it’s clear SaaS companies should have products marketed to end users, simple enough for them to consume without heavy configuration, and priced so they won’t appear on the central procurement team’s radar (e.g. freemium models to test and deliver value ahead of the paywall).

Slack is the example many cite. It markets directly to end users, who can try it for free. Once it has taken hold in an enterprise, it spreads virally thanks to strong network effects (even across enterprises). All the while Slack reaps the benefits by having a pricing structure that reflects usage.

Some argue that it’s harder to make this bottoms-up, direct-to-consumer approach work in areas like fintech, where regulation is important and IT security teams have more muscle. But we see it happening everywhere.

One fintech example that we came across recently, in the context of our upcoming wealth management report, is Hydrogen.

Hydrogen offers a classic bottoms-up approach: a user can provision for themselves a free sandbox environment from the Hydrogen website, pre-integrated with the services they’ll need (like Plaid). The customer only begins paying once they cross certain thresholds, such as API calls. In addition, that also take a jobs-to-be-done approach to solving end user problems by offering discrete services as no-code plug-ins to existing applications, meaning an end user can add a service like tax optimization in minutes.

Avoiding the aggregator tax

Going direct to underserved or overserved consumers is the new playbook for disruptive innovation, but it doesn’t mean companies can avoid the aggregator tax. In fact, costs just get reallocated: in retail, CAC is the new rent, while in enterprise, CAC is becoming the new senior sales rep.

Nonetheless, while unavoidable, businesses can minimize the aggregator tax.

One way is to invest in brand.

A lot of startups seem to believe investing in brand is a luxury. We don’t agree. Marketing is like a pump: first, you have to fill it, if you want to draw it down. Sure, you can generate some leads from well-targeted paid campaigns, but it’s not a sustainable endeavor and you’ll end up paying more to aggregators over time. Instead, you want to run paid campaigns into a customer demographic that’s heard of your company and thinks positively about it, which will improve your ad score and lower your ad cost.

You should also invest in data. Going direct to consumers means you know more about your customers than your industrial age predecessors ever did. It’s critical to capture this information in order to:

– build proprietary routes to customers

– learn more about your users to better target and to aid self-discovery through recommendations

– learn how consumers use the solution so you can constantly improve the utility of the product, everything from ease of purchase to completeness of solving user’s problems.

You must leverage the power of networked buyers.  At the most basic level, your product should be differentiated enough that customers will want to advertise it on your behalf. That customer advocacy might come in the form of a positive review or post, but as former podcast guest Julian Lehr highlights, it may also come in the form of signaling.

In addition to advertising, the networked customer can be an acquisition channel.

Wherever possible, you should try to build into your solution the viral features that make the product better when it’s used together with others (e.g. messaging), that lead to customers acquiring other customers. And, even when it’s not possible to build these viral features into the product, it’s possible to build them around the product. A Nike running shoe has no inherently viral features, but the community it has built around its products, the millions who share fitness information, definitely does. And it’s not just consumer brands that can create communities. Consider Salesforce, for instance, which has a community of over 2 million organizing events and sharing content. This attracts others, but also binds together the users in a way which makes it hard to leave the community (by choosing a different product).

Don’t just rent commodities, rent luxuries

If you’ve read this far, it probably won’t come as a surprise that we advocate for renting commodities. Don’t buy your own servers, for example, or write your own accounting software. This will keep operating costs low and variable.

However, we also advocate for renting specialist skills. It allows more cost flexibility and avoids underutilization, but it also reflects a structural shift.

The best people increasingly don’t want to work for a single company. They like the variety and the speed of learning that comes from working across multiple companies and projects. And, platforms are emerging that go further than just matching companies with freelancers – platforms that put together, manage and take responsibility for (the output of) interdisciplinary teams. Effectively, they give businesses greater flexibility and quality at scale and specialists the security and freedom to keep learning.

Achieving internet escape velocity

Brett Bivens, a venture capitalist at TechNexus, came on the Structural Shifts podcast earlier this year to talk about his theory of “Internet Escape Velocity”.

Essentially, internet escape velocity is what happens when a company successfully executes the strategy playbook described above. That is:

– it identifies an underserved or overserved niche

– it leverages internet distribution to reach those customers directly

– it unleashes a growth loop by combining the reinforcing properties of product, distribution and monetization, and

– it uses data and marketing to avoid the aggregator tax

At that point, it hits internet escape velocity, becoming capable of crossing the chasm of precarious long tail supplier to become an aggregator itself, using the pull of its loyal customer base to pull in more suppliers or to launch new own-label services. Brett uses the example of Spotify, which he describes as follows:

“over time, as they expand and gain leverage, podcasts are a higher margin business, social products are a higher margin business, marketplace products are a higher margin business for them. And so, by owning the consumer demand via the lower margin streaming business, they have the opportunity to expand into those areas.”

When a company hits internet escape velocity, it also has the option to invest in fixed assets. As we showed in a paper we wrote with Dave Galbraith in 2016, internet-era companies tend to become asset heavier over time as they seek to entrench their position and deliver better customer fulfillment. But the critical point is that they invest after they achieve a sustainable route to customers. Assets grow like the roots of a tree, downwards from distribution rather than upwards from production.

If, however, a company that hits internet escape velocity doesn’t want to invest in fixed assets, the good news is that it’s never been easier to rent supply in the post fixed-cost economy.

When Software has Eaten the World (#29)

Structural Shifts with Belén Romana, Spain’s former head of Treasury

There is a lot of anguish over what’s happening online these days from the rise of hate groups to media manipulation, the propaganda to interference with elections — are the positives of our digital world even worth it? Well, today, your host, Ben Robinson, digs into this question with Belén Romana García — Spain’s former head of Treasury, and an economist who has worked in both the public and private sectors. Belén is also a board member for several public companies and foundations. She says that people are primarily driven by three things: power, money, and knowledge — and she is especially driven by knowledge and curiosity and a desire to understand the world and its possible future. Today, she and Ben discuss, should our elected officials have to learn how to code to better understand the world that we’re living in? Should we scrap GDP as a metric since it’s not accurately reflecting our service economy? Does democracy mean equal voting? And how does the information and infrastructure of our online world affect our freedom or a sense of freedom in real life? And more. 

Belén recommends:


  1. One book: A world without work, Patrick Susskind
  2. One influencer: Azeem Azhar
  3. Best recent article: The geopolitics of information, by Eric Rosenbach and Katherine Mansted (Belfer Center for Science and International Affairs)
  4. Favourite brand: Spotify
  5. Productivity hack: Never waste a chance to learn something new. Listen to podcasts while driving, flying, cooking or having a walk.

Information in itself is a huge good that we have in abundance. Of course, when you have any scientific advances or new ways of creating value or knowledge, that does not come without its flaws, and without its problems. It takes time for societies to understand the real implications — good and bad, by the way — of any advance. It’s sort of trial and error. We are understanding the implications, the advantages, the disadvantages, and it will take some time until we do understand the whole thing.

[00:01:32.14] Ben: Thank you very much for joining us! So, the key thing that we wanted to pick up on today is one that you talk about a lot, which is the notion that software is everywhere, and it’s this idea that, as software’s become more powerful, it’s proliferated, and it’s become much more pervasive in our lives, our communication, our politics, our industry — and I thought maybe a good jumping-off place might be the quote from Peter Thiel, the one where he said, “We wanted flying cars. Instead, we got 140 characters.” So, has this world of pervasive software delivered on its potential? i.e. Do you think that the world is now better for having so much software in it?

Belén: Definitely! I think that we are better off. We can communicate better, we can find more data, store that data, analyze that data, deploy that data. So, I think that information in itself is a huge good that we have in abundance. So that’s a much better world. Of course, when you have any scientific advances or new ways of creating value or knowledge, that does not come without its flaws, and without its problems. It takes time for societies to understand the real implications — good and bad, by the way — of any advance. So, I think that we are better off. It’s sort of a trial and error thing. We are understanding the implications, the advantages, the disadvantages, and it will take some time until we do understand the whole thing. And so, it will take some time until we get the world right. And we will try again and again and again. And finally, at some point, we will have a reasonably good set of rules. So, I think that we are better off but we should be quite modest in terms of we have to understand that this is a journey that started, as societies, and it will take us some time to really get it right.

As human beings, we usually get to understand, control, and react. So, the fact that some advances have dangers, that’s always the case; […] And reality leads us to think that once we get to understand, then things are much better after that advancement than before. The problem, of course, it’s always the transition period.

[00:03:44.01] Ben: If we think about some of the negative applications of digitalization, we might think about, you know, its scope for manipulation, for example, or we might think about, you know, some of the scary things that people say about where AI is headed, right? You know, that we’ll be controlled by software, rather than it kind of being used as a tool by us to improve, almost. But you believe that, on balance, overall, it’s been a force for the good?

Belén: I think so. Of course, it has many dangers. But that happens with anything that you can think of. When the car started, there were so many dangers around that and there were no roads. So, it took some time — decades — to set the rules and understand what is good, what is bad, what should be done, what shouldn’t be, who should be controlling that? How should the authority control that? How can we drive around the world? So there were a number of things that happened over decades. But finally, the car gives us many things — of course, good and bad — and if you understand the real implications in groups, for example, climate change, then you get to have a much better deployment of that advance. You know, as human beings, we usually get to understand, control, and react. So, the fact that some advances have dangers, that’s always the case; that happens with medicine, that happens with pharmaceuticals, that happens with anything you could think of, basically. And reality leads us to think that once we get to understand, then things are much better after that advancement than before. The problem, of course, it’s always the transition period. So, since this is a trial and error, and it will take time for us to get it right, there will be many things that will happen that won’t be good, and people that will be harmed. That’s absolutely the case. And that’s a very difficult point because, as I said, when it’s something new, we don’t fully understand the final implications. So, this trial and error… And you see that everywhere. So, we didn’t know that Facebook could be potentially dangerous for political institutions, the suffering, that outcome to understand, “Yeah, this may happen.” And now we are starting to think, should we do something? What can we do? What are the limits? So, does it mean that social networks should be banned? I don’t think so. They should be here for good. It’s like with cars. I mean, the fact that a car is a dangerous tool doesn’t mean that you shouldn’t control it, you shouldn’t regulate it.

We are on a very, very early stage and it’s very difficult for rulers to set the rules. One of the things that I worry about around this — setting the rules — is the fact that Parliaments understand laws, but they don’t understand codes. It turns out that codes are also ruling our lives. So, if we don’t get them to understand code, to understand that language, it will be very difficult for them to set the right rules. And I think that we’re far from that.

[00:06:38.21] Ben: And if we just continue for a second with that analogy of the car, where do you think we are in this transition? You know, it took a long time before cars had seatbelts, it took a long time before cars had emission standards. How far into this digital transition do you think we are? And how good a job do you think the rule setters are doing?

Belén: We are on a very, very early stage and it’s very difficult for rulers to set the rules. One of the things that I worry about around this — setting the rules — is the fact that Parliaments understand laws, but they don’t understand codes. It turns out that codes are also ruling our lives. So, if we don’t get them to understand code, to understand that language, it will be very difficult for them to set the right rules. And I think that we’re far from that. If you had a council of wise men, old men that couldn’t read, and they had to set the rules for the printing press, they would just say, “Okay, whatever is printed that I cannot read should be fair, balanced, whatever.” But they cannot enforce it, because they cannot read the book. I think we are in a similar situation where individuals setting the rules do not speak the language, cannot read the code and hence, all they’re doing so far is giving this open recommendation of should be fair and avoid bias and these things. But then, they are driven by the outcome, and they cannot prevent it from happening, because they don’t understand. So I think that we are at a pretty early stage. We need rulers that understand code.

because of the cyber GDP that we are not measuring, we are not taxing, we are not tracking. So, we don’t know whether we are getting richer or not, we don’t know whether we have a fair tax system or not. Basically, there’s a part of the GDP that is not there.

[00:08:30.29] Ben: I think the EU Cookie Policy is a very good example of a policy that’s set by people that don’t understand the nature of the digital world. What about economists? So, you’re an economist. How good a job do you think economists are doing at understanding the new world and changing their measurement tools? To make it more concrete, we basically live in a world where we don’t think we have any inflation, we think productivity growth sucks. But is that really the case? Because you know, there might not be inflation in the price of streaming music, but there feels like there’s a lot of inflation in other areas — healthcare and such. So, do you think we’re using the right measurement tools for the digital world?

Belén: We, economists are really struggling because we are all educated in an industrial world. So, if you take GDP — GDP nominal versus real is inflation, GDP is a very industrial concept in many ways. I mean, it started as one physical thing times the price, and that is GDP. And the evolution of both the number of physical things and the prices did see it to the GDP real and nominal. We started struggling with services. So, for decades, for example, financial services were not part of the GDP because no one thought it could be measured in terms of any value added. So, with services, which is something in the middle between digital and industrial, we struggle to understand productivity to measure; you don’t have a physical thing — you know, whatever the assets times the price. It’s services. And you know what that means, in a very industrial world. So, we tried to build this bridge with services — did a very good job, but reasonably good, I think. But, if you jump to digital, you don’t have this vertical approach to sector, which has been industrial. And there’s this huge debate now in the economists’ world, around inflation and productivity as you were rightly saying. There are now two camps: some people say we’re heading into a world of no inflation forever, and some people are thinking that we’re heading into a high inflation world because finally, monetary phenomena. But, as I said, I struggle with that, and it’s difficult to get the cyber economy into this, the cyber factor into the GDP.

Belén: Now, I’ll give you one example that I always think of: the value for Google of having Germany, where is it? Is it part of the German GDP? It doesn’t look like that because that value is not the revenue coming from the German part. But more than that it’s network — so, it’s part of the network that increases the value of the whole network. That’s not part of the German GDP. Where is it? My point is that probably it’s part of the Google market cap and that’s why, because of this cyber GDP, we are not measuring, we are not taxing, we are not tracking. So, we don’t know whether we are getting richer or not, we don’t know whether we have a fair tax system or not. Basically, there’s a part of the GDP that is not there. And you see that again, and again, and again. You gave the music industry as an example. That’s a perfect example of something like that. So, we used to have a music industry that could be measured in an industrial way — the number of LPs or the number of stores, or the number of concerts, and then you have tickets times price. Then all of a sudden comes streaming. Basically, in terms of GDP, the music industry has disappeared, because you don’t have much employment, you don’t have tickets, you don’t have LPs. So, you don’t have physical things that you can measure. However, we have never been able to access so much music in our lives. Not necessarily for free. You pay for that — Spotify, you pay for it. But you access a huge store of music that you can choose to confine whatever. That’s not part of the GDP.

The competition policy is basically based on the knowledge that if the prices are low, competition works because the prices are the final signal of a monopoly or oligopoly. It turns out that’s not the case because we’re not paying with money.

[00:13:06.14] Ben: Yeah, it seems to me that the problem that you talk about translates into consumer surplus, right? Which is, you know, by definition, not captured because it’s surplus. And I often wonder if we were able to better understand consumer surplus, and somehow feed that back into GDP, that might be a way of capturing some of the benefits. But I just wonder, in general, if GDP is just, you know, we should scrap it, if it’s just obsolete. And we should start again. Because you know what I mean? It’s like you’re saying, you know, it’s a bit like inflation — you have RPI minus x, RPI minus y. And it’s like, how many things can you augment a broken metric for, in the end, before you have to just start again.

Belén: But the fact that we are using not a perfect measurement does not mean that we should use none. So, for me, the key thing is understanding. Again, understanding that we’re missing part of the economic evolution, wealth — whatever you want to call that — and that we need to develop other means. We will still have a physical world, we will still have services. But we are missing digital. So I think that one of the key things for, first, universities and probably also statistical authorities, is developing that understanding of how to measure. This is a new thing. It does exist, it does create value. So, this is again, a very old debate: value versus price. And that impacts everything. So I’ll give you another example. It’s the competition policy. The competition policy is basically based on the knowledge that if the prices are low, competition works because the prices are the final signal of a monopoly or oligopoly. It turns out that’s not the case because we’re not paying with money. We’re paying with data, and data has no value, no price. So, they’re for free. So, all of a sudden, you have a competition policy based on a very industrial concept that needs to adapt to something different, which is we’re paying with our souls, so to speak — our data. So, I think that the key thing is for, firstly, universities to start thinking about this — and I think that some have already started — thinking of this concept of cyber GDP. And value versus price happens when you have value, but no price or a price that does not fit much with the value as in the case with data. So, I wouldn’t say, okay, we don’t use GDP anymore. But I think that we should, at once, develop other complementary ways of measuring and understanding the digital world.

when you have a huge concentration of power, you have a problem. And as a state, the concentration of power is always dangerous. So, one of the things that I really wonder is why people — that happens especially with younger generations — do not care about giving away their information to companies but they care about giving information to the state.

[00:16:03.20] Ben: I just wanted to return to that idea of regulating the digital economy, because, as you say, you know, in the past, the litmus test was, you know, our prices going up to the end consumer, therefore, there’s sign that the company has market power, and can manipulate pricing. How do we regulate networked businesses? And how worried are you about the increasing size and influence of some of the largest platforms?

Belén: Over the years, when you have a huge concentration of power, you have a problem. And as a state, the concentration of power is always dangerous. So, one of the things that I really wonder is why people — that happens especially with younger generations — do not care about giving away their information to companies but they care about giving information to the state. And if you live in a democratic state, you do have rules, you have a transparent system where you know your rights, and you have tools to defend your rights. That’s not the case with the large platforms. And there are millions of that; you have Facebook measuring video views and charging for it. So, it’s the same as the judge and the defendant is the same. Or you have no ability to prove that whatever video you’ve posted has been more or less than another one. So, you know, I’m starting to read articles on the corporation as a courthouse, because within the Amazon world, it is much more efficient to solve conflicts within Amazon, than using the courts. That’s, I think, hugely problematic, because, as a citizen, even though you know the rules, those are not the standard procedures. I mean, if you live in a democracy, and you don’t like the Prime Minister, you can vote against him or her, and at some point, the guy leaves. You cannot vote against platforms. So, you don’t have any access to understanding how it works, what your rights are, and how you can defend your rights, what kind of tools you have. Whereas democratic societies have all those things very clear. So for me, as a citizen, I would rather give my information to a democratic state than to someone I cannot access. I don’t know where I am.

The problem with data is that the value comes from the aggregation. So, on your own, you cannot get value from your data because your data or my data have no value on themselves. They have to aggregate and aggregation means something that goes farther than the individual.

Belén: So, these platforms have huge power in economic, financial, political, social, and they have no rules. So, I think that we do need to regulate those platforms. And they basically act as monopolists in different fields. So you have a set of monopolists. And that’s, by the way, nothing new. We had that in the late 19th century, where, in the US there were, again, a number of huge monopolists. And the state, at some point, reacted and they said, “Okay, hold on. We need to do something about the mobile or we need to do something about many different industries — the oil industry and then telecoms” — because the power was too much. I think we are exactly where we were at that point. And, in order to avoid that, the competition policy was born. Now, I think we need to think of another competition policy. But I don’t know whether that’s enough. All the states have a regulation for networked industries, but networked industries are something much more national. This is international, isn’t it? And we don’t have common rules. We have common rules on how do we rule the seas. And that’s similar. So, you know, there’s an international law around these. We don’t have an international law around digital and that means sometimes we don’t even have a national law about digital. So I think that we need to develop that.

[00:20:28.23] Ben: Yeah, I suppose the only good precedent there, you know, whether we think is a good piece of legislation or not, something like GDPR, even though it’s a law that’s imposed within one sovereign area, which is the EU, it does tend to have ramifications outside of that sovereign area. Because if you want to do business in the EU, you have to treat customer data in a certain way. And what tends to happen is those policies then tend to become globally applied. But I just wonder, in general, I think you said earlier on that we have politicians that understand the law, but they don’t understand code — I think is what you said — and that was never more evident than in the times that the big platforms are brought to Congress, for their annual grilling, right? And I’m just wondering, you know, if it’s difficult to impose regulation cross border, and if it’s difficult for our current generation of politicians to impose the right kind of regulation at all, is the answer maybe to devolve more responsibility down to us as individuals, and in some way try to give us more transparency so that we make better-informed decisions about the platforms we use and what we share with them and so on? So, i.e. you know, put more responsibility in our hands.

we are clearly moving into a world of fragmented internet

Belén: For a number of reasons. One is they are monopolists. So that’s the first point. The second point is the nature of data. The problem with data is that the value comes from the aggregation. So, on your own, you cannot get value from your data because your data or my data have no value on themselves. They have to aggregate and aggregation means something that goes farther than the individual. So, the combination of the two things makes it very difficult for individuals to really be responsible. So, you would be asking them for something they cannot do because they don’t have the means. I’ll give you one example. For me, GDPR is the first step. But I’d really like the right to be forgotten because that’s something that, you gave away your information, or someone gave away your information, in a situation when we didn’t really know the consequences of that. And then, you suffer those consequences. You know, the right to be forgotten was a concept that was born in Spain — it was a Spanish case that led to this thinking of ‘this is unfair’ because someone did something three decades ago, and the guy is suffering from it again, and again, and again. And the problem is, I’ll give you one example. If you have a public appointment in Spain, your number gets published. And, at the same time, it looks like that should be private. And, you know, years ago, I had an attack, and the police told me, “Everyone knows where you live, and your number.” And I said, “Yeah, but that came from the fact that I was a public official.” So, all of a sudden, you have something that was ruled thinking of a different world, and now all of a sudden becomes a threat. And you cannot do anything because that is not part of the right to be forgotten. That’s nothing wrong, but it’s part of my privacy. And even then, it’s there. So, as an individual, I cannot do much. I need infrastructure or authorities, courts, that help me protect my rights — understand and protect my rights. So I don’t think that individuals are the way out.

[00:24:18.08] Ben: So regulations are still a thorny issue. And the other issue I wanted to talk about was your idea of policing the seas, right? Because what are the seas? I think we’re getting lost slightly in the analogy, but like, you know, what’s the landmass? What’s the sea? And if Europe is a landmass and we don’t have any really large platforms, where do we stand? I mean, for example, for how long can Europe impose regulations like GDPR when it doesn’t have platform companies of its own? Because it’s a bit like, you know, we’re imposing legislation on companies that aren’t even in our jurisdiction.

This dream of an open Internet for every country I think it’s over

Belén: But that’s why I like the analogy of the seas because, in the international law, you do regulate companies that are not part of your jurisdiction. And I think that the internet has been almost a global ocean. I think it’s not anymore. The exception being, of course, China. Right from the beginning, the Chinese thought, “Oh, if this is an ocean, I want to control my ocean, I want to set the rules.” And I think that we are clearly moving into a world of fragmented internet, where we, again, have not an ocean, but different oceans or seas if you want. And you said that the Russians announced that at some point, they wanted to close their internet and have a sort of a narrow channel into their internet so that they can control both — their internet and the channeling. The Chinese, of course, control that. I think that the US is thinking also, we need to think of how we control and have a walled garden — that we know what’s going on, who’s doing what, and we can react to that. So this dream of an open Internet for every country, whatever, I think it’s over. And, of course, that worries me in terms of Europe. So Europe has advantages and disadvantages. The disadvantage is that, of course, we lack any sort of platform. That’s a huge disadvantage. The advantage is that we still are the largest market in the world. We are interesting. The problem is when you’re not interesting at all — then you’re done. But when you can add value, then you have some tools to regulate that value that you are creating. So I think that Europe has that right and that ability. And also, the tradition. I think that we do have the tradition. We may over-regulate sometimes, but we know how to regulate and we think of the individual as someone that has to be protected. That comes from a long, long tradition of European thinking, philosophy, political science, sociology, anything you can think of. So, from that point of view, Europe has a chance.

in the 21st century, those countries or regions or companies, for that matter, that will rule the world will be the ones that will be able to do the three things: produce, store, and analyze data, move around those data — the platforms — and then set the right rules.

Belén: The thing is, for me, that if we use another analogy, and we think of the late 19th century, and how the industrial power was built — and that means power or states or companies — there were three things: the ability to produce things, the ability to move those things — so you needed a physical infrastructure — and then the rules of the game. And if you look at the industrial world, the industrial powers what they did is, Okay, I produce, I build infrastructure — the train being the first one; if you look at how the trains were designed in the 19th century, you knew who was powerful in terms of countries, and who was losing the battle. And only looking at that map is clear. And then, the rules of the game, which is free trade. That, coincidentally only applied to industrial goods, not to agricultural goods. Okay, 100 years have gone, so then digital power — who can produce, store, and analyze data? Very few countries or companies. The infrastructure is the platforms — we lack the infrastructure. And then the rules. So far, we haven’t got any rules. No rules at all, not even this free trade rule. We didn’t have a rule because we didn’t think we needed it. So, in the 21st century, those countries or regions or companies, for that matter, that will rule the world will be the ones that will be able to do the three things: produce, store, and analyze data, move around those data — the platforms — and then set the right rules.

[00:29:14.16] Ben: So how worried are you about Europe? Because at the moment we’re trying to set the rules to some extent — you know, GDPR, for example, PSD2 — but we don’t produce or move the goods, right? Or we don’t have the infrastructure, so we won’t be able to set the rules very long in that case. So, where does that leave us? And do you think the game is over in terms of creating the infrastructure?

The European Union, as a project, is built on conflict, it has always evolved out of problems. It’s not when everything works, and it’s peaceful. Whenever there’s a problem and a crisis, right from the beginning, the inception, the European Union evolves and tries to build on that.

Belén: I think we’re really far behind but it doesn’t mean that we have to say, “Okay, we give up.” I think we should fight. And fighting means a number of things. One is if we could understand how to rule this right — how to read the code — we would have a huge advantage. And that’s something that we could develop. Another thing that for me is hopeful is the fact that we’re starting to listen to the European Commission talking about digitization, which is quite a new thing, in a much more thoughtful way. And I’ll give you one example. So the European authorities came to the conclusion I think — that’s my understanding, I have to say — they came to the conclusion that “Okay, we have lost the war on personal individual data, but there’s a huge wave of data coming, which is data coming from systems, physical things. We are an industrial power, why don’t we build on that?” I think that is great because, by the way, physical things you get to personal data. So even though we lack the platform for personal data, which is the case, we could build platforms around non-human data, if you want. It has stemmed from good systems. I thought this is good, because instead of wasting resources fighting something that’s going to be very difficult, because you’re on late, let’s try to build something that is not built anywhere else, and then we won’t be late.

Belén: So, for me, that’s a very useful way of understanding that. And I think that with COVID-19, that the European Union is using it in a positive way, so to speak. The European Union, as a project, is built on conflict, it has always evolved out of problems. It’s not when everything works, and it’s peaceful. Whenever there’s a problem and a crisis, right from the beginning, the inception, the European Union evolves and tries to build on that. I think that now is starting to happen that. I mean, Europe has a soft power, and Europe has someone that understands that there are two huge giants rising, which is the US and China and thinking that they don’t like each other as much as they did in the past — and Europe is in the middle; that could be a potential advantage if we are smart enough. So, for me, the worst part of it would be if we thought of it as when you have a castle on your right, another castle in your left, and you are on the plain, then whoever comes, you’re done. That is not the way to understand that, I think, because then we’re done. If we build on our chances, I think — and we do have chances — I think that’s quite clear. You know, with the cloud, we’re late, but still, Germany and France are thinking or are starting to create a cloud. Will it be like that? Will it be as competitive? We’ll see. But I think that’s the right move. We should have done that move long ago. But at least we’re starting. So, in relative terms, we are behind the US and China, but we are ahead of the rest of the world. So, I’m hopeful. And I think that Europe is built on this notion of the individual has his rights, which is differential, I think.

[00:33:34.17] Ben: It’s quite interesting what’s happening with Tik Tok because I think, you know, is probably a good way of thinking about this, which is you’ve got this Chinese castle, the US castle. Do you think what we’re now seeing is, you know, those two citadels are trying to now establish and define their spheres of influence? And so, you know, Tik Tok was a case of… That’s almost like an invading army and we’re not going to allow that into our Citadel. And then everything that’s happening with Tik Tok in India seems to be a bit like, you know, India’s kind of aligning around the US castle. And then, I think your current comparison of Europe as kind of no man’s land is quite accurate because Europe hasn’t yet, I don’t think, really decided which castle is going to align itself with. And, as you said, maybe, for a short period of time, that gives us a position of influence because we can arbitrage between those two castles. Is that the way you think about the world, which is that that’s where we are now? The idea of this, you know, Pax Americana kind of world that was global is over, and increasingly nations are gonna have to choose between which of these two castles they align themselves to — or do you think they’ll be more castles like the Russian castle, the African castle?

when we all have access to information, that does not give us knowledge

Belén: A long list of things, as you mentioned. One is India. I don’t think that India is aligning with the US. I think that India is thinking, “I’m large enough and I’m advanced enough, and I have the knowledge and human capital needed to build my own castle. So, what I think that they are doing is, “Okay, let’s create our own apps.” So then it’s much more the understanding of, this is exactly a backdoor that you can use to look around and I want you to have that. It’s not only the influence, but it’s also the information that you get. One of the things with digital, it looks like the network effects come with a huge amount of data and that huge amount of data, if we’re talking about personal data comes from large populations. India, which is a very large country has the means to create its own network effects. They don’t need to have anyone else. As it happens with China. They are large of their own, they don’t need to have anyone else to create network effects. So, I think that they have understood that and they want to build on that. And from that point of view, neither the US nor Europe, on our own, I don’t think we are large enough in terms of that sort of pool of population. Probably, we need to think of something that combines both, otherwise is very difficult. If you think of if the US thought, “Okay, I’m closing my castle, it’s only us”, that’s not enough. That’s clearly not enough. They need to think of other pools of population. And probably we will see, at least, I think that large countries will try to build their own castle. Brazil is another example of a country that thinks that they are large enough, that they have the means, they have the knowledge to do something on their own. And so, we will have that as it happens with it. If you think of the industrial world, of course, the first superpower was the UK, but then Germany reacted pretty soon and then the US. So, you end up having a short number of industrial powers, but a number of industrial powers. So I think that Google sees something similar. I think we will have a number of powers. And I hope that Europe will use that as a chance.

[00:37:36.17] Ben: I think what’s interesting about India is with Geo. They’ve almost sort of separated infrastructure from production, which is to say, “You know, since we own the infrastructure, we’ll allow foreign companies or American platforms to operate on the infrastructure, but we own the infrastructure.” So, they have a certain level of sovereignty, that Europe, for example, doesn’t have, because we need to have the platforms or the infrastructure, right? So, I wonder if that might be the model. You referenced Brazil. I wonder if that might be the model that is followed by others. At least, then, you have a stake in the digital world, whereas if you own neither the infrastructure nor the platform you have, nothing.

Belén: I agree. I think that at least you need to control the highway. And then, you get to decide which car can go through that highway and whether you have to charge or not and the whole thing. If you lack everything — if you lack the highway and the car, you’re done. So, as it was the case in the industrial world, you needed to have infrastructure for the trains that you built, controlled. Otherwise, if we didn’t have the money to build our own infrastructure in the 19th century, the British companies that built it, they did it for their own, not for the Spanish people. It was more to extract value out of mines, basically. So, we had exactly the wrong infrastructure that we did not own, not control, not design. You’re done. You cannot succeed in the industrial world, with those weaknesses.

[00:39:16.15] Ben: I want to speak of something that you talk about, which is that democracy does not equal voting. What do you mean when you say democracy does not equal voting and how is that sort of relevant to this digital shift?

Belén: People tend to think that if I vote, that’s fine, then I can defend my rights. But that’s not true. For example, you have certain rules to control propaganda or financing in a democracy or what kind of information you can give or how long — are you always in campaign or can you be bombarding people again and again and again, do you give them one day they have to think? Whatever. So a number of things that complement the voting. But I think that we have completely forgotten that that infrastructure of rules is key — and if you don’t have that, then voting doesn’t… Any dictator can organize voting. And, quite coincidentally, is always to his advantage, it’s always 95%. It’s not the voting. It is the whole thing that comes before the voting, that grants that that voting will be a legitimate exercise in terms of democratic access. So, that, for me is key. You can opt-out Facebook — that’s voting; you vote with your feet, which does not change much, perhaps, because you’re out, but that’s not much.

when information is for free, we tend to vote out of emotion

[00:41:02.27] Ben: So you’re arguing almost we’re succumbing to this illusion that because we get to vote more often and there are more referenda that we’re somehow more empowered and we have more control over our affairs, whereas you’re saying the opposite is true? Because if I understand what you’re saying, rightly, you’re saying a lot of the underpinnings of democracy as we think about it are being washed away or eroded by digitization. So, can we just delve slightly into that? So if we think about some of the things that are critical to a democracy, such as having accurate information, is that principally on information that you think that democracy is being eroded? Or do you see other areas where the waves of digitization are washing?

COVID-19 was the rise of nation states. They were reborn, and all of a sudden, people turned to them to be protected, in many ways — limiting the movement of citizens, offering healthcare, buying vaccines. So, the nation state has, again, become a key player in the economic and social world, which means that people, when in danger or in a difficult time, and in various situations, they turn to those that are closer to them and have the size to protect them

Belén: I’ll give you one example. I can access a lot of information. Does it mean that I can self-diagnose my illness? That’s completely wrong. I have information but not knowledge. So, the difference between information and knowledge now is quite clear. It has never been the case over centuries in human history. So, we need to understand this difference. And, you know, there was this very interesting exercise. IBM has this Project Debater and they created a machine that can debate with a person and they had this public exercise that you can watch on YouTube, of a guy debating with a machine.

Belén: And that exercise was arranged by Intelligence Squared US. And they have this mechanic whereby the public first says that they are against or for whatever proposition and then after the debate, they vote again. The party that has moved more wheels, if you want, or more opinions, wins. They did exactly the same. And their proposition was, “Should states finance, pre-schooling education?” And the machine was told to say ‘yes’, and why and the individual, ‘no’ and why. And the exercise was, for me, quite telling, because, of course, the machine came with hundreds of examples, of data, if you want. And the individual less, but some. But he, even though he voted he was against that proposition and most of the public was for the proposition at the beginning, he could move more opinions, because he could understand emotion. All of a sudden, when we all have access to information, that does not give us knowledge. That means that we can look at things that we would like. And, of course, you can always find data that justifies your prejudice. That’s always the case. You can look at part of the information, and that leads you to say, “Yes, I’m right!” Or the other part, which is, “I’m wrong.”

Belén: Finally, when information is for free, we tend to vote out of emotion. And when I saw that exercise that I found so interesting, I thought, “Now I understand Brexit.” So finally, is it, “Do I trust you or not? Do I think that you have an intention or not?” So, one of the things that I find quite interesting is that there’s this crisis in democracy, at least democratic states, that people are losing trust in institutions, being those private or public. And then comes COVID-19. And for me, it was quite a surprise because COVID-19 was the rise of nation states. They were reborn, and all of a sudden, people turned to them to be protected, in many ways — limiting the movement of citizens, offering healthcare, buying vaccines. So, the nation state has, again, become a key player in the economic and social world, which means that people, when in danger or in a difficult time, and in various situations, they turn to those that are closer to them and have the size to protect them. If it’s tiny, if it’s only a city, they don’t have the size to give me what I need. I need something larger. I think that’s good for democracy, of course. Any economic crisis raises the danger of populism and all these things. So, I’m not naive from that point of view. But the way I see it, I think that democracies with this huge crisis, have a chance to rebuild themselves. And, for example, the fact that the European Commission is now trying to get to buy enough vaccines, I think that increases their legitimacy. The problem was, in the beginning, the European Commission did not react at all. So it’s not only they lacked legitimacy, they lacked action. The fact that you need someone that can protect you, I think that gives a chance to democracies to react and rebuild themselves. But of course, they need to develop an understanding of the code, to control the rules. So, all these things that we have been mentioning, and of course, build your own infrastructures, as well.

[00:46:55.29] Ben: You often talk about how the conception of democracy is bound up with industrial age concepts, that I suppose the other question is, you know, if democracy depends on nation states, and nation states are with a long term view in trouble, and if democracy depends on the institutions of the industrial age, then is democracy in trouble from two sides?

Belén: The key distribution of wealth comes through wages to work, which is quite an evolution in terms of the history of the humankind. So if you work more, you get the chance to be wealthier. And that means that we are basically all the same because we can become the same over time. And so, why can you vote and not me? And you being first a landowner, then a man or a white man. So, if I am a woman, and black, and I am poor, I can become what you are. So, why can’t I vote? Now, with the digital world, it’s, again, if we find the way to assign prices to value, that can bring us again, to a situation where I can earn my living in a reasonable way, and hence, this whole structure can survive. If that’s not the case, if we cannot understand how value is created and distributed, and therefore, how fair our economies are — we don’t know how fair they are, because we don’t have the right measurement, again, so we will measure part of the fairness or unfairness; but the rest does not exist. So, if democracy is a key to grant fairness, then we need to have those tools. Some people will be helped by training. Some people won’t. So you need to think of how can you protect and help people that won’t be able to be retrained. You know, with COVID-19, I found quite interesting the fact that live sports have suffered so much, and then eSports are thriving. So, all of a sudden, you have a whole sector of eSports that is growing, and that entails employment, in many ways.

[00:49:25.12] Ben: Is that a positive? Because I absolutely 100% share your view. I’m an optimist, I believe that the world that awaits us will be more positive than the world that we leave behind. But I’m worried, like you are, about the transition, because this transition is now happening so much faster on the back of COVID-19. Does that somehow reduce the scope for wars or civil unrest or whatever nastiness that could normally come to because it’s happening now just, you know, at an accelerated rate?

Belén: In terms of the COVID-19 I’m worried about the short term economic effects it brings. That’s absolutely true. And it’s also true that is somehow accelerating these trends. But the positive side of it is that they’re becoming visible. If a problem exists, and it’s visible, you can tackle it. The problem that we have is that these trends have been there for 10 years, 20 years, and they weren’t visible at all. So, there was no public debate, public worry, nothing at all. Remote working has been there for a number of years and some companies were very good at using remote work, and some specific groups of people wanted to only work remotely. But it was sort of a lateral and receivable, but it was clear that it was a trend that at some point would affect many others. Why don’t we have an industrial organization with a service economy? That’s exactly the case. Now, all of a sudden, we have found that, indeed, we didn’t have an industrial organization. Well, now it’s visible. Now everyone is thinking of, “How can we do this? Should it be five days a week? How can we really apply this technology that we already had, but we didn’t use?” So, from that point of view, COVID is making many of these trends visible. I think that is, a tree that falls in the forest — is it falling? No one hears or sees it. It’s exactly the same with problems. And we only look at problems when they are big enough, as societies. Now they have become big enough. So, I think that that will also accelerate the transition in terms of their reaction.

[00:51:47.14] Ben: I almost feel like the furlough scheme, or you know, whatever it’s called in the country. Is, in some way, almost like an experiment in the universal income in a way, which is, how do we compensate people whose jobs are not going to exist in a digital world?

Belén: Yeah, yeah, absolutely. And instead of having the finish with this experiment, now we are all making it. And there will be things that won’t work, but we will find which ones do. So I think that’s another example of things that I think are accelerated. You will get to see that. If you use that analogy of the COVID and digital, some populations are suffering more than others. And so, they get isolated, so you need to protect them more, because they are isolated, because they need to be isolated.

we have the chance to understand better the world that we are already living in. I think that societies have stopped and are thinking — which is really, really helpful because it’s not just the elite thinking

[00:52:38.13] Ben: A lot of people use the war analogy for what we’re going through, which is, the enemy is the disease, it’s not another country. And that’s interesting in a couple of ways, right? One, because the government’s taking an unprecedented or warlike intervention into the economy to boost aggregate demand, and so on, which is interesting. But then, what a lot of people say is that won’t be the rebuilding exercises necessary after a war, which then sustains aggregate demand. But in a way, don’t you think that might also force governments to figure out how to redefine the tax base?

Belén: Not only the tax base but also the economic structure. If you look at the European Union Fund, it talks about two things. One is green, and the other one is digital, which is we need to rebuild the societies thinking of those two things, which is quite a novelty, I have to say. So, we have an enormous amount of money to rebuild the economic structure. And that, I think, is a unique opportunity. So, we don’t have physical infrastructures that have been damaged, as is the case in a war, but we have indices that have been damaged by the digital world, and they have not been able to react. And so, I think that’s a huge chance. It’s sort of a Marshall Plan effort in a different way. So, it’s not bridges and buildings, but it is platforms and data. I think that there’s no clear answer to that. It will depend on the society and how they feel that they need it and whether they have someone that will finance that.

[00:54:35.16] Ben: But that Marshall Plan, if I’m not wrong, that Marshall Plan is something that Europe is thinking about, but it’s not something… You know, whether we consider the UK in Europe or not anymore, I think in four months we’ll find out but the UK doesn’t have an equivalent of the green new deal or the digital Marshall Plan, and the US may or may not, depending on elections in November. So, do you think many countries will take this once-in-a-generation opportunity to rebuild?

Belén: Again, I think about that, and it depends on the country. But in terms of the nation states, we are seeing, again, states are not only regulating but owning and managing a large part of the economy. So, that’s more sort of the ’70-like situation where you had the state doing two things: doing things and regulating those things. And I think that model that looked completely over is back, and it has advantages and disadvantages. So, back to the network effect, much of the technological revolution in the US came from the collaboration between private and public institutions. We never had that in Europe. Now, I think we’re going to have that. And it’s a precondition. But you need to have both sides of the world collaborating. I think that’s key. When you lack both, you lack size or the ability to combine private and public forces, that country will struggle, I think. It’s not only the quality of the politicians, which is key, but also the tools that they have. Probably they won’t have the same tools.

[00:56:29.29] Ben: I want to finish on, if we can, on a really positive note. I think you started by saying that digitization has been a force for the good, right? We may not see it in our GDP statistics or our productivity statistics but we feel it. We feel it in particular as consumers. But then we got on some slightly more negative topics. So we talked about how democracy is struggling in the face of digitization, we talked about how digitization is kind of dissolving the global world into nation states, we talked a bit about how the transition is going to be tough. Maybe we started to get into more optimistic ground with how the pandemic may find a route to faster transition. But I just wanted, if you would, to finish by giving us your most optimistic projection of the future.

Belén: It’s difficult for me because, by nature, I’m not an optimistic person. I think that the problems that you don’t identify, you don’t solve. So, I tend to focus more on the problems than on thinking that there won’t be problems. So that depends on how you define an optimist. But, in my view, I think that we have the chance to understand better the world that we are already living in. I think that societies have stopped and are thinking, which is really, really helpful because it’s not elite thinking. I think the whole society is thinking. It happens — this crisis — at a time when we have already suffered some of the negative impacts of the digital world. So, we have either Facebook political things and things like that. So I think that awareness is higher and I think that’s very good at all levels of society. So, for me, an optimistic or a positive — much more than optimistic, a positive outlook into the future, would be democracies get stronger, because they understand the rules of the game. We get to measure track, and if needed, to control the value and impact of digital and hence, we really create abundance and give it away to us. I think that Europe has the size, the human capital, and now the understanding. So, we have the tools, so let’s use them.

[00:59:19.08] Ben: Belén, thank you very much for your time!

Belén: Thank you! Thank you so much!

What we learned from doing the Structural Shifts Podcast

What the world looks like when seen through “a great series of conversations with people who are building the future”


In the break between seasons — season 2 starts again on September 24th — we thought we’d take stock and reflect on some of the things we’ve learnt from making our podcast.

When we started Structural Shifts (initially without a name, that came from episode 14), it was just an excuse to reach out to and chat with people whose life and work we found interesting. What surprised us was two-fold: firstly, that people liked it and recommended / introduced other people we should interview, making the whole endeavour sustainable; and, secondly, that many of the evergreen topics we set out to explore bubbled up to the top of the people’s agenda and consciousness.

The fundamental transformations that used to quietly shape our world — over years or decades — suddenly became topics for mainstream conversations. The acceleration everyone’s talking about — we felt it too and it has influenced how we make and develop the podcast.

It’s safe to say that doing the podcast has taught us valuable lessons.

“I found this podcast borderline post-modern, and refreshingly frank. A profound look into what’s next. Love every bit! The guests, the music, the format.” — Simone Cicero, Co-Creator of Platform Design Toolkit

A crisis can be clarifying

The worldwide lockdown had wide-ranging effects, some of them even on the positive side. For example, it enabled us to interview thinkers we greatly admire, but who are geographically remote — like Rita McGrath and John Hagel — as well as to grow our audience.

In a world already hungry for meaning, the pandemic triggered a pressing need for strategic thinking. First, it made people pause and reflect on what truly matters — for their lives, work, for the planet.

Then, because institutional and private reactions to the pandemic left many disillusioned, they became determined to gain a stronger understanding of big topics — fintechinternet business models, geopoliticsthe climatethe future of work.

We had profound, unhurried conversations with people who are thinking and doing things differently. Their thoughtful observations, distilled from decades of practice and reflection, challenged our received wisdom on a range of topics — from innovation to marketing — as well as encouraged us to entertain contrarian viewpoints.

Instead of a just-do-it mentality, the pandemic reinforced the timeless value of reflection and flexibility, reflexes that all our podcast guests share. If you keep an eye out for it, you’ll notice that in every episode we publish.

Good questions are catalysts for change

Good podcasts depend on two key ingredients: interesting guests and good questions.

Our listeners increasingly took care of introducing us to great thinkers, some of whom — like Brett Bivens or Julian Lehr — we caught on the rise to becoming big stars. And we concentrated on trying to get the best out of the conversations.

In the past six months, we’ve spent a lot more time on research. As our audience grew, so did our sense of responsibility to get the best out of every conversation. Many weekends and late nights were spent reading the books our guests had written, which made us well-prepared — and hopefully improved our the return on our listeners’ time.

Some of the book authors we had invited at Structural Shifts podcast

Our goal was also, for ourselves and our listeners, to delve into diverse topics such as the ethics of technological change or building a safety net for the self-employed. A risk because many podcasts listeners like to keep digging into a given topic like investing, we hoped to create the context for the cross-pollination of ideas, frameworks, and viewpoints that can serve both professional and personal pursuits.

As the inner workings and implications of the networked age leapt into view for the entire planet, we developed an even keener focus on asking questions that help us have better, more stimulating conversations. Questions are essential to decode, deconstruct, and rebuild our vision of the world as it is — and as it might become.

The case for techno-optimism is one of our favorite examples of such a conversation, providing signposts to use when engaging in mainstream conversations around key topics in tech and their society-wide impact.

“A great series of conversations with people who are building the future. Each one is like having a dinner conversation with a smart friend who has come back from a voyage. I listen when driving or jogging — the miles just melt away and I arrive with a refreshed mind.”

It’s easier to connect when you share purpose and focus

Another thing we noticed while doing the podcast, especially in the past 6 months, is that people who share the same principles tend to resonate (or “click”) more easily when having conversations remotely.

It was surprisingly easy to delve into complex topics with them because everyone was eager to dive in. Maybe you’ve also noticed how small talk takes less and less time in online meetings as we have more of them.

This desire to have important conversations, to support clarity and good decision-making translated into our guests sharing personal perspectives more openly.

What’s more, it was easier to connect with new guests who dedicated even more time than before to share their expertise and experiences. We’re grateful for every minute they spent with us!

Capturing attention in a roaring world is a big challenge

As Herbert Simon predicted, a wealth of information gives way to a poverty of attention.

Our response to this has never been to compete on giving information, but to focus on carefully curated insights. A great fan of craftsmanship, we meant for the conversation — except for maybe the couple we did on previewing the post-pandemic world — to be timeless; as relevant now or in two years’ time as they were the day they were recorded

We also found that the lockdown period — or more specifically the extra time that many people gained through not travelling and commuting — opened up more demand for the long-form product we offer.

“Always insightful and informative. It is a relaxed conversation with people who have had interesting experiences and something to say. Ben Robinson, brings out the best in each guest.”

The Structural Shifts podcast remains one of our favorite projects, in which our enthusiasm for the topic and our guests’ generosity combine to help you see farther — and more clearly.

Helping ourselves and our network to move from scalable efficiency to scalable learning and, in do doing, to prosper in our networked age is why we do the podcast.

We hope it helps you achieve the same.

“I really love this podcast series. There’s not much content like this coming out from Europe. Should serve as an example to others” — Bozidhar Hristov

Our thanks to all of our guests and listeners and to Sarah Mikutel, our podcast editor. In series 2, we’ll be back with more mind-expanding conversations, covering the token economy, the future of finance, the end of globalization, the startup community way, the new precariat and much more…

Changing the Business Model of the Internet (#12)

Structural Shifts with Andy YEN, Founder and CEO of ProtonMail

This podcast was recorded at Fongit — Switzerland’s premier innovation incubator. Ben Robinson is joined into the conversation by Andy Yen — Founder and CEO of ProtonMail — and we discuss the inherent trade-offs that the current Internet present to the users, the importance of privacy and security but also how to compete against Google head-on via business model innovation.


Ben Robinson (Ben): Welcome to the Aperture Podcast. For this episode, we’re at Fongit and we’re with Andy Yen, founder and CEO of ProtonMail. Andy, welcome to the podcast.

Andy Yen (Andy): Yes, thank you. Thanks for having me.

Ben: Andy, I want to start with the idea of trade-offs, which is sort of inherent to the Internet because I mean, it’s become such a sort of integral part of our lives. I mean, it’s completely embedded in our lives. Nobody could imagine living today without the Internet, but the Internet was kind of started on the model of everything being free and that’s come with all sorts of negative externalities. So, I’ll start by asking, have we paid too big a price for the Internet?

Andy: I think it’s the opposite. We haven’t paid a high enough price for the internet. We live in the world where we’re used to getting things for free, but as any economist will tell you, there’s no such thing as a free lunch. Nothing is free. Everything has a cost, and in fact, if you look at the internet today, if you use it for free, if something is free, then you’re actually the product that is being sold and repackaged and actually you don’t bid it out to the highest bidder. So because of that, I think we’re not paying enough for the Internet and we’re paying in ways that we probably don’t understand and we probably shouldn’t be paying, namely today we mostly pay for our Internet services with our privacy in our personal data and that’s a price I think most people don’t realize and understand.

Ben: No, I think you’re right. I think most people would just think these services are free, full stop. And I think also with this idea of surveillance capitalism and the surveillance state and so on. What’s so bizarre, about this state of affairs, is that whenever people thought in the past, in 1984 about dystopias they thought that it would be really intrusive. So the state would be monitoring you in an active way, but actually what we’re doing is we’re surrendering our own privacy. Whether it’s a conscious choice or an unconscious choice, it’s a choice that we’ve made. So, do you think we can ever turn back the clock on that?

Andy: Well, I would argue that people haven’t actually made a conscious choice, because let’s take an average person. When you agree to a Terms and Conditions, you don’t actually read it. You don’t know what you’re consenting to. In fact, we have no idea. If you imagine a 13-year-old child on social media, cause that’s the age in which you’re allowed to go on social media now without parental consent.

Ben: So there is actually now some notion of parental…?

Andy: Well there’s always been a notion of certain minimum age. It’s not really adhered to and respected. It’s not really verified.

Ben: I didn’t even know that was the case.

Andy: Yes, by law, it’s supposed to be 13 in most countries. If you take a teenager going on Facebook or social media for the first time, I would actually say they don’t really understand what it is they’ve consented to. They don’t realize the implications of giving up their data forever. They don’t realize that at the time that they’re posting a photo on Snapchat or on Instagram that this data could come back and haunt them for decades better. This is not what you think about in your first social media experience. So we haven’t really surrendered. We haven’t really consented. In some ways we’ve been sort of tricked into doing this and I think that’s an important distinction, because consent must be informed and without information, you cannot willfully consent to something.

Ben: And what role should the government play here? Because I didn’t even realize there’s a minimum age. But clearly somebody, somewhere has set a minimum age because they understand that anything like drinking or having a gun or whatever… there are some… you want a person to have a certain level of education and to understand the implications of what they’re doing. So there have clearly been some small efforts here to try to safeguard the consumer or the individual, and we’ve seen Europe issue GDPR, again sort of top-down pieces of legislation to try to safeguard the individual. We have these cookies, laws and things. Is it possible in the same way as you know, I asked you if we could turn back the clock and change customer behavior? I don’t think we can, but can we protect customer behavior through legislation?

Andy: Well, the customer behaviour point we can revisit later, cause I’m not sure I entirely agree that we cannot change customer behaviour. Let’s talk about a little bit about government regulation. To understand regulation really, we need to look at history and kind of understand history. Governments have been in the business of regulating businesses since essentially the beginning of time. If you look at, for example, finance. That’s another industry that is quite widespread, known for some pretty bad abuses, things like that. The first Stock Exchanges appeared maybe in the 17th century, but the first legislation really only came after the Great Depression. It was the Securities Act of 1933. And then onward, even that wasn’t enough. We had a financial crisis in 2008 and then more regulation came afterwards. So what that really shows is regulation and governments in particular, it takes time.

Ben: And they’re also reactive as well.

Andy: Yes, reactive as well. So you cannot say, let’s rely on government regulation to keep us from going off the edge of the cliff because historically, that has never been the case. They regulate after you got to the edge of the cliff, and if you want to take action before then, I think it needs to come from the private sector. It cannot be government driven. Government will not drive the change in this area, because they historically never have done that.

Ben: And what about the consumer? Because you see a different kind of model in China, for example, where there’s an expectation from the consumer that they do need to pay for a lot more services in China, so some of the things that in the UK, US and so on are free, like this podcast. In China, there’s an expectation you pay for podcasts. So do you think the role of the private sector working together with the consumer is to try to actually create a price for using the Internet?

Andy: Well, I actually think China is not a very good example here, and I’ll tell you why.

Ben: Yes. I think it also has very strong trade-offs.

Andy: Yes. In China, there’s surveillance capitalism and surveillance in general. You are not fooling or tricking anybody. Everybody in China know they live in a surveillance state. They are aware the government is spying on everything that they do. They have the ability to read all their personal information. This is kind of a known fact. They live with it. And in fact, I would say that they have a much higher awareness of the risks of surveillance than Western society, because in China you write the wrong thing online, you end up in jail. Here we don’t have that sort of pressure and that’s not part of our everyday reality. So in fact, they’re not being tricked because they know exactly what’s going on. Whereas in the Western world, in fact, most people have no clue. It’s not part of their daily reality. So that’s why you cannot compare with China because the level of awareness is completely on a different level.

Ben: And so just to revisit that idea of it won’t be the government that solves this, or at least proactively or in advance of whatever the next scandal is, so you’re saying it’s more the role of private companies and possibly I would argue private companies in alliance with the consumer. So how do you think that will play out?

Andy: Well, ultimately when you look at an industry and how it evolves, it always comes down to consumer choice. Consumers will make a decision and their decision translates to how they pay for things, their wallet. You pay with your wallet. You vote with your wallet in many ways and that decides the future of any industry. If consumers decide that electric cars are the future, then that’s the future. If customers decide that they want to have more privacy in the future, then that’s where things are going to go. My view and the reason that we’re doing ProtonMail and working so hard in this space is we actually view privacy as an innate human need. It’s something that everybody has a need for. Whether you realize it or not is another question, but you do have that need. You have your passwords on your accounts. You have curtains on your windows. You have locks on your doors because we have an innate need for privacy and that need for privacy so far hasn’t translated to the Internet yet, but it will. It’s just a matter of time.

Ben: And do you think it can happen without education because maybe we can go back and discuss customer behaviour. But in order to change my behaviour, I have to realize there’s something wrong with that behaviour or at least the risks associated with that behaviour. So you can have a situation where suddenly the private sector allows us choice. For everything, we have two versions — the free version and the paid version and one gives us privacy. One sacrifices that privacy and we make that choice. How do we make the consumer aware and make that informed choice? Because at the moment as you said, they’re not making a conscious choice because they’re not aware of those inherent trade-offs.

Andy: Well, there’s a couple of things there. First of all, to make a choice, there has to be a choice, and until ProtonMail came along, there really wasn’t a choice. You wanted email; you’ve got it from a service that was going to invade your privacy and mine your data. That was it. So what industry and private sector can do is a) we can provide that choice. That’s the first step. Once you have the ability to make a choice, then it’s just a matter of time in some sense, because Google’s business model, Facebook’s business model, this is not something that can be hidden forever. People will eventually figure it out, and the reason their awareness today is higher today than it was, let’s say 10 years ago is because the technology is newer. Back then, people didn’t really understand it. Now they do. And in 20 years, of course, more people will understand.

Now certain things like education. Certain things like data breaches and scandals, these things will accelerate that trend, but knowledge tends to spread and the information will get out and that’s why it’s really just a matter of time. So what we are controlling an industry is providing the choice and then seeing what we can do to accelerate that shift.

Ben: And then do you see in your own business, in adoption and so on a positive correlation with the things that you were just saying, like scandals, you know. So after Cambridge Analytica, did you suddenly see a spike in demand and interest for ProtonMail?

Andy: Yes. So each time there’s a scandal, there are always spikes but you don’t build a business or change an industry based off of spikes. What you need is you need a strong underlying trend and that trend is not going to be based off a scandal. That trend will be based off of user awareness and user knowledge and user sophistication, and that really comes naturally. The first time you go on Facebook, your first thought is not how do they make money, but after seeing some ads that are suspiciously like what you’ve searched for and hearing some stories from people who have used the service and know more and maybe friends who are well educated, you slowly begin to know. So that knowledge permeates out into the society, but it takes time. It’s not overnight.

Ben: You yourself used the term change an industry, and the industry you’re trying to change is a very big industry. And so implicit in what you’re doing is that you’ve got just massive ambition, and also you are really early with this. In preparation for this, we watched your TED Talk and you were saying things back in 2014 that people have only really started saying recently about these sorts of negative externalities of a free Internet. And so a) how did you get into this? And b) how do you arrive at wanting to sort of take on Gmail and Google and the biggest corporations in the world? I suppose most people would have thought this is a great business, but it’s just so big and scary that I’m going to do something smaller. So how did you arrive at this and how do you manage with a business that has such ambitious mission?

Andy: Well, I think what you mentioned earlier, and I think back in that TED Talk, we are trying to in fact change the business model of the Internet. Because the only way that our mission succeeds is if you build a new business model, and the business model cannot be around advertising. so that’s of course very ambitious cause it’s trying to disrupt an industry that today is maybe over two, $200 billion a year. So that in fact is the mission, and I don’t think you’d go into this with that in mind. You don’t go into a job and say, okay, I want to disrupt a $200 billion industry. Some people maybe go in that way, but the way that I went into it and the way that our team went into it was we saw a need. It was a need that we think is important. Not just…because there’s a business here, but really because of the social impacts. Democracy relies on freedom of speech. Freedom of speech relies on having privacy. It’s all connected. So we solved the social angle that was very, very important. And that’s what made us decide that we needed to pursue this. And I think a lot of that comes from the history of the company. It was founded at CERN. It’s created by scientists in fact. The World Wide Web was also created at CERN, so we had a connection to that.

Ben: You had a kind of symmetry there.

Andy: Yeah, and we saw how the Web had transformed. It was invented to foster communication. It was invented also to build a better world, and we think the Web today is not really reaching the full potential of what it was created to do. It’s become in many places actually a tool of oppression. And for us, the key thing is we need to try to reverse that trend somehow by providing choices and options and tools for people, and that is really what drove us into the business. After coming in, I guess we’re not stupid, we realize that it’s quite challenging, of course, as you say, but we think this is something that’s very important to do, success or fail. Someone needs to do it. And I think that’s the reason why most people go into Science, because you do it because somebody needs to do it. And we thought that calling to go out there and do it.

Ben: So I’m guessing you’re a physicist?

Andy: Yes, that’s correct.

Ben: Yes. So you’re a physicist, and that’s how you ended up at CERN, and how you go from CERN to creating software to change the Business Model of the Internet?

Andy: I think the joke is that it’s all Math at the end of the day.

Ben: Correct.

Andy: So, in fact it is all math. It is all logical reasoning. So the great thing about physics is basics. You can kind of go into essentially any field out there. What physics does, it doesn’t teach you any specific skill so to speak, but it teaches you how to think. So in fact, quantum physics to new technology was not a huge jump.

Ben: And tell us how it works, and by extension, the business model for it. How you make it pay?

Andy: The business model is a freemium. So that means you can use it for free. But if you want more advanced features and services, then you have to pay for it. So, in essence it’s quite similar to Dropbox and it’s either you pay for the service or the service is going to sell you to pay for itself. That’s the model that we picked for the service. It works by using something known as end-to-end encryption, and what end-to-end encryption does is it encrypts data on your device before it reaches our server, and the beauty of that means that all the data that we store on you is encrypted in a way that we actually can’t access. So, this guarantees you two things.

One is privacy because if we can’t get your data, we can’t sell it to third parties. It’s as simple as that, but also it gives you security kind of for free, and that’s because the data that we store is encrypted. If we were to get hacked or breached, a hacker cannot steal from us something that we don’t have. So, really security and privacy go hand in hand. There are actually two sides of the same coin. And that’s why I think it’s really important because it’s not just for privacy. It’s also for security and cyber-crime today is one of the biggest challenges that we’re going to face in the 21st century. So, it sort of solves both of those problems at the same time.

Ben: Yeah. And what I like about it, to revisit that point from earlier is, doesn’t really require any behavioural change. Instead we carry on doing what we always do, which is emailing many people at once, big attachments, all the things that are a little bit risky, but this time they’re encrypted. So if you like, you’re not asking people to change behaviours that have grown up with the internet, you’re just protecting that individual from exposing themselves to the risk of hacking.

Andy: Yes, yes, exactly. And I think that’s really, really important because if I go out on the street and ask anybody, would you like more privacy and security? No one’s ever told me “No” to that question. It’s always, who do I give up? What do I sacrifice? What’s the trade off? And our mission, if you want to sum it up in the sentence, is to make that trade off zero. So we know we want to reduce all the technical hurdles, reduce all the user interface hurdles and make it as simple and as easy as the services you already use that don’t have encryption. So the encryption needs to be fully automatic, fully transparent, invisible to the user. And that is really, our goal, and that’s actually in many ways actually our main innovation.

Ben: And so if you send me an email, which you have, from ProtonMail, it comes into my email server displayed in a way that’s like every other email?

Andy: Yes, yes, that’s correct. Now in that situation, obviously my inbox is encrypted. Everything on my side is secure, but if you’re not using ProtonMail, you are using Gmail, then of course Google will get a copy of your email. And that’s why in Proton there’s kind of a strong network effect, because if you want to protect the entire network, you want to get your friends and family on to use the service. And that’s something that has really been driving the growth because there’s obviously still value in protecting your own data, your own inbox. But if you want to protect your whole community of people on contacts, then you also get them on the same system. And we see more and more of that as we continue to grow and scale.

Ben: So really pleased that you raised that because I was because I was going to ask you about Customer Acquisition Cost and the speed of adoption and so on. So, there are two things, based on what you said that really would suggest to me that this would have very low customer acquisition costs. One is the freemium model. And so the first question is what percentage of users eventually become paying users? And then the second model is to what extent are the network effects really driving user adoption?

Andy: Well, we know that the largest segment of growth, so the single biggest reason that people use ProtonMail today is actually word of mouth referrals. So network effect is in fact the biggest driver of growth. So that’s that then I think freemium is also important, but for us, freemium is really going back to the mission of the company. Our mission is to make privacy accessible to anybody in the world that wants and needs it. And you can only do that by offering the service for free. I’ll give you example. We have a VPN service as well, and the country that has the most users of Proton VPN is actually Iran.

Ben: Okay.

Andy: They’re under sanctions. So in fact, they don’t have credit cards. They’re not tied into the banking system. They have no way to pay us, and if the service wasn’t free, then those hundreds of thousands of users in Iran would get cut off. So that’s why for our mission to succeed, in fact, we must make it free to make accessible and we subsidize that by people who do pay us.

Ben: So there’s actually… I don’t know if you’re familiar with this term… shared value transactions. But it’s this idea that the heaviest users or in this case, the most demanding users maybe pay for all the features that are then offered to everybody, and you only need a small number of those people to become paying users for the whole model to work. So what’s the conversion rate on freemium into premium in your case?

Andy: Well, it depends a lot on country. And the way a service like Proton works because we’re so global is in fact you essentially have the First World subsidizing the Third World.

Ben: So it’s like probably what the model for climate change.

Andy: Exactly. I think that’s actually a good thing because in fact, there’s sort of like a hidden transfer of wealth there, but that is in fact needed to create a more equitable and just world. So I think that’s actually a good feature. And conversion rates of course are much higher in Western Europe and the US. They’re much lower in places where people cannot afford to pay, but we’re okay with that because overall, we have a model where it’s sustainable, it’s profitable. So that means it can continue to grow and it’s a cycle that fuels itself and that’s what we need to have a business that can survive and be stable.

Ben: What is the trigger point to go from freemium to premium?

Andy: It’s very hard to say. It depends on the person.

Ben: Sorry, to ask it in a different way… it’s a set of features that I sign up to that then makes it premium, right?

Andy: Yes, that’s exactly right. So paid accounts have certain features which are not available…

Ben: Such as?

Andy: Well, you can have additional storage, additional features like auto-responder on emails. So it’s a subset of features and storage that you need you don’t get on the free account, and that’s what is driving conversion. But I think a lot of people pay for the service really to, I think support the ideals of the company and what it stands for. And they pay because they know that privacy matters. They care about it, and if they don’t pay, they simply won’t have the privacy anymore. Someone needs to pay something if you want to have this option. So it’s really about keeping a private option available to the world and I think that’s very important.

Ben: And then do you have many businesses that adopt ProtonMail for all of their employees?

Andy: Yes. So today there’s probably between 20,000 and 30,000 businesses who are using the service. They obviously all pay, and for business the value proposition is very clear. You can go to a cloud but still keep control of your data. That’s important to a lot of banks and financial industry customers. But then secondly, is your communications within the company are secure and private by default because every employee will be automatically security-encrypted.

Ben: Because when you laid out the binary choice earlier between either having privacy in your email communications or not having privacy in your email communications, and you said you had this sort of free options like Gmail or the paid options like yours. There’s a third actual option, which is you use your corporate email account, but in which case the corporation you work for can see everything that you’re seeing and doing. Right? So I’d argue that there are actually three choices for individuals. Would you agree with that? But in this case, I guess the corporation allows you free access to email that’s secure and where they can’t spy on you.

Andy: So actually, I would argue against that… because what is email fundamentally? Email, I would argue is actually the most meaningful passport of the 21st century. In the 20th century, your identity was tied with your passport. That was who you were. That was how you verified to the world who you were. That was how you accessed all the services that the world could offer you. Today, everything’s moving online, and every online service that you sign up for is actually linked to your email. It’s your online passport, and the idea of using your corporate email account doesn’t actually work because that would be like say, your employer is your identity. It’s not true. So that’s why for that reason, you know, that’s not really an option for most people.

Ben: It’s so true because that’s happened sort of insidiously, which is so many services just say sign up using Gmail, sign up using Facebook. And again, unconsciously you don’t think about the ramifications of that. But you’re right, it’s become way more linked to our identity than…I mean, you saying that it’s made me realize, wow, that’s true, and I hadn’t even thought about that myself.

Andy: Yes. So Gmail isn’t just, let’s say controlling your data. They’re also controlling your identity in some ways. And you don’t realize that when you set up for it. But that’s what is happening. And that’s why we feel it’s very important to have an independent identity, which is built on the idea of privacy and security at its core and respecting your rights online.

Ben: Some people make an argument … I obviously know where you stand on this, but some people say the genie’s out of the bottle. We can’t go back and reclaim our privacy. So almost more logical would be to make everything transparent and if everybody could see everything, then there’ll be less abuses of our rights. Where do you stand on that? How do you challenge the logic and the soundness of that argument?

Andy: Well, I would ask that person that tells you that for their email password and see they give it to you.

Ben: Good point. And so I can imagine that many people for many reasons — because they want to challenge what you’re trying to achieve in your mission, I imagine you’re the subject of attempted hacks a lot. How do you counter that?

Andy: Well, the best way to protect data is to not have it in the first place. So the way that we counter that is simply by encrypting as much as we can so that we don’t have anything that can be stolen from us.

Ben: But if they solve — I don’t know exactly how it works — but if they can figure out the encryption keys, then they can get access to the encrypted data.

Andy: To this day, not possible to break the encryption that we use. This of course maybe won’t be the case two decades from now, but we encrypt data in a way that even the strongest supercomputers in the world theoretically should not be able to crack the data. So it is for all intents and purposes inaccessible to a hacker.

Ben: And so that’s one risk. I think so you’ve convincingly argued that you were ahead of that risk. And I suppose in the same way that the sophistication of hackers will grow, the sophistication of the technology you use will also improve. So it’s a race, but for now you’re ahead. The other big risk I guess is coming under pressure from governments to just hand them the data… Because a bit like Apple says you know our devices are encrypted and safe and so on, and then the US government says we want access to it. Have you come under that kind of pressure? How do you deal with that?

Andy: Both the answers of course is Yes. Like any major tech company, we do come under that type of pressure. We’re based in Switzerland. Switzerland has very strong privacy laws and a strong history of protecting privacy. So that is obviously a very strong advantage. But the technology also works. If you do the technology properly, in fact, there was no way for us to decrypt the data, so it’s mathematically not a possibility.

Ben: So you couldn’t, even if you wanted to, decrypt the data you have on your servers?

Andy: Yes. That’s the point of intent encryption, is that it’s encrypted before it comes to us and we don’t have a way to decrypt it.

Ben: So your argument is simply, no matter how much pressure you apply, we simply can’t respond to your desire.

Andy: Exactly. But there’s also a third aspect to this, which is that the opinion of governments has shifted quite a bit from 2014. In 2014 it was, okay, encryption is a major threat to national security. Now the conversation is we need more cyber-security because they’ve seen the hacks and the scandals and the issues of the past five years. So in fact, governments that were maybe pushing and trying to pressure us in 2014, some of them are now our customers.

Ben: Really? I was going to ask about users. We’ll come back to it.

Andy: Yeah. So this kind of shows a shift in how the attitude has changed because people realize that in fact we’re not just providing the privacy but also security, and security is the key to securing the 21st century digital economy.

Ben: Yeah, and again, coming back to this point, I just think it’s Canute-like to sort of say, okay, let’s roll back what we’ve put in place and instead it’s much better to say what we’ve got in place that’s secure. And that’s what I really like about everything that you’re saying. The mission — obviously everybody buys into the mission — but also, it’s a mission and a business that doesn’t require… it doesn’t introduce loads of friction into our lives basically, because we’ve got used to friction-less experience using the Internet.

Andy: And the other way to look at this is Does encryption come with certain risk? Of course, it does. It’s not black and white. Yeah. Is it possible that terrorists could use something like ProtonMail? Of course, but we also know that they also use Facebook. They also use Twitter. They also use buses and planes. You cannot possibly ban everything a terrorist could possibly use. The way you address this question, the way you look at it is you say, what is the overall social benefit? And when it comes to trains and cars and planes, of course they can be abused, but there’s a very strong benefit there as well. And it’s the same for encryption in security technology. We needed to secure our data and secure our future because the future is data and the future is online. And the social benefit from that outweighs the risks that may come as a result of this technology.

Ben: So we’ve started to touch on a couple of times, but now, if you don’t mind, delve into a bit like who uses it? So I imagine, millions of individuals like me use ProtonMail, but I’ve read on your website, you were talking about the activists that use it. You’ve said governments sometimes use it. I guess terrorists could use it as you said. So who uses it and what do they use it for? First of all, give us a sort of breakdown of the demographics of the kind of people that use it, and then is there anything about their usage that makes you uncomfortable as a CEO?

Andy: As a web service and a email provider, the users are basically anybody that would use Yahoo, Gmail or Microsoft. It’s the same spectrum of people.

Ben: Is it though, because I guess particularly, I don’t know…I mean, we’ll come back to numbers of users if you could share that with us, but I imagine particularly beginning, the early adopters were people who were hot on this idea of privacy in a way that others weren’t.

Andy: I think in 2014 that was true, but today with more than 20 million users, it’s really kind of come into the mainstream and everybody will have a reason to use a service like this. People that want privacy and security. It isn’t just the paranoid guys or the government people or the journalists or the activists. That’s a need that actually even your mother might have. In fact, I would say, of course, we may not have as many mothers using ProtonMail, compared to let’s say, you know, Gmail. But that demographic is represented. So we capture actually the full spectrum globally, but it’s really people that have a higher awareness. What links them together is not what the field, they work in their backgrounds or where they live, but relieve their level of awareness of how the internet works and their level of concern for privacy and security.

Ben: And that 20 million number, that’s staggering. The first time I ever heard you speak was at an event here at Fongit and I was blown away by what you’re saying, the ambition of what you’re doing. But even then, which is I think just a few months ago, the number wasn’t anywhere near 20 million. So that would suggest to me that you really are seeing these network effects kick in. It’s like exponential type growth.

Andy: Yes. The growth is pretty rapid right now. It’s really because of awareness. People are just more and more aware today than they were in the past. If you go in 2014…, 2010, Mark Zuckerberg basically said that privacy is no longer a social norm. He comes out this year and he says privacy is Facebook’s main focus now.

Ben: So do you believe him?

Andy: Of course not, but the fact that guys like him have shifted tone so dramatically in the past decade really shows where things are going, and that’s why I think if the awareness is up and we want to be ready, we want to have the service that people can rely on, and we want to really be part of the change that we want to see in the world.

Ben: And do you see more users or greater adoption in countries where their privacy is potentially at greater risk? And do you see — and I suppose you don’t even know that much about the demographics — but do you see, just to go back to this topic of activists cause you yourself were talking about this in one of your blogs, is it people and places where their privacy is more at risk?

Andy: I think adoption is certainly higher in those communities. Among journalists and activists, it’s very, very popular. A lot of people are using it. But if you look at our VPN service in Iran, that’s a huge user base there as well. So of course, if you have a need and if it’s something that your life depends on, you’re going to use it right, and you’ll use it may be potentially more and preferentially more compared to somebody else who isn’t at risk. So there is of course a correlation.

Ben: Where do you stand on freedom of speech? Because activists are using your platform to defend their freedom of speech and to defend their freedom of protest, and somebody like Mark Zuckerberg, let’s bring him up again. I suppose he would make any argument that would absolve him from taking any responsibility for protecting freedom of speech. Where do you stand and what are the limits of freedom of speech if we can’t see what people are saying?

Andy: I think freedom of speech is obviously something that’s very important. It’s the cornerstone of modern Western democracy right now. Can freedom of speech be abused? Sometimes. There’s hate speech, there’s hate crimes, there’s far-right and far-left people who can abuse services like this. Email though, which we do is inherently a private conversation, and what you say in private, in fact is protected. Now if you use technology like Twitter or Facebook or even ProtonMail to let’s say, espouse far-right hate speech, then that is actually against the law. That’s against the law in Switzerland and we as a company would adhere by Swiss law. And I think on topics like this, it’s important for tech companies to really follow the law. We cannot take on the role of judge, jury, executioner, adjudicating and passing judgment on these types of issues. We must adhere and respect the legal framework.

Ben: And so just again to get back to the idea of the legality of some of these things and therefore government’s role in regulating the Internet, do you remember the law that was passed in the Clinton era that essentially absolved lots of these internet companies from taking responsibility for any of this stuff, from having any responsibility for safeguarding the truth on their platforms? Do you think it’s time to change that law then? Because if you’re saying that ultimately your responsibility is to adhere to legal frameworks, do those legal frameworks therefore need to be upgraded for the internet age?

Andy: Well, ProtonMail for example as a mail service because mail is private communications, is a complete separate category from other social media where it’s publicly view-able. So this discussion for example, wouldn’t really apply to a service like ProtonMail, which is private communications. Now the law — I do know the law you are talking about, what this basically says is platform providers like Facebook and Twitter cannot be held liable for the content that is posted by their users, because the user is actually the author of that and my personal view is, in fact, I think that law is in fact correct because you cannot…it’s probably in many ways almost impossible to force platform providers to really police everything that is said and posted on their platforms. I think it’s actually impossible. Where I think there does need to be some more strengthening is that if something posted is false, defamatory, or malicious, there needs to be ways to probably legally compel the take down of such information after it’s proven to be false, defamatory or malicious.

Ben: Do you think they have a responsibility themselves for that kind of fact checking?

Andy: I would say there’s a moral responsibility to do that, if not a legal responsibility

Ben: I’m not sure morality and ethics is high on the list of some of the people that run these platforms.

Andy: I would say they’re more concerned about their profit margins and their bottom and top lines, and spending time checking content and responding to legal requests is probably not very cost efficient. So yes, most of the platform providers probably hide behind this law and say it’s not our obligation, but I do think they do have an obligation to remove data that is just false and wrong.

Ben: Do you think it was a sort of necessary step that we needed some sacrifice of our own privacy in the beginning to get the Internet moving? Would you argue it was a necessary precondition to widespread adoption for us to be giving up some of our privacy and now has come the time to solve, now it’s adopted in our lives to then just tighten up?

Andy: Yeah. Well the Internet was never really free to begin with. You pay for your internet service; you pay for internet protection. In fact, from the beginning of time the Internet was actually not free. If you look at things like Hotmail and Yahoo, in the early days those services were not fully free. The free Internet and free Internet services is a relatively modern invention in history of Internet. It really came in the mid-2000s, so it was paid first. It became free and now is shifting back towards being paid again.

Ben: So you argue it needs to just go full circle?

Andy: Yeah, yeah.

Ben: Where do you stand on net neutrality?

Andy: I think it’s important to protect it. I think we need it. Without net neutrality, you essentially give big companies certain advantages that smaller people cannot compete with.

Ben: Yeah. So you would argue it was a mistake to make that change?

Andy: Yes. I think net neutrality is important for the future of the Internet. That’s my personal view.

Ben: So are you campaigning for it to be reintroduced?

Andy: Well, I’ll give you an example. In the UK there’s a new law that allows ISPs to censure various websites if they believe it’s adult material, and our VPN service — Proton VPN — was mistakenly categorized as adult by some ISPs. And the ISPs are very hard to get in touch with and not very responsive, and in fact, the only way to solve this issue is to make an official complaint under EU net neutrality rules. So that’s an example of a small company being able to compete in a market because of net neutrality. So in fact I think it is very, very important and without that, you don’t have innovation. If we didn’t have this legal recourse to do this, then we’re at a disadvantage because a big company like Google will never get blocked because they are too big.

Ben: Want to talk a bit about the scaling journey you’ve been on. So you’ve got 20 million customers, which is …again, I’m staggered by that number and I guess you’re adding tens of thousands a week or whatever the number is. What have been the challenges along the road to scaling this business to that extent, and what trade-offs have you had to make because clearly you want this to be the most secure, but that costs lots of money. Clearly you want this to be as economical as possible for the users, but that’s a difficult thing to achieve and to produce a really high quality. So what trade-offs have you made and what have you learnt in scaling a business to this level?

Andy: It’s no different than any other business actually. There are three levers that you can control. There’s cost, there’s time which is related to cost to some extent, and there’s quality. If you want high quality and high security, invest more time and as a consequence of investing more time, you are investing more money, and all businesses to some extent are automating these three levers and depending on the company you go to, they are could be on very, very different sides of the spectrum. For us, security and privacy are our core business. Quality is very important. So we’ve optimized the business in such a way that we care about quality first and costs and time are sort of secondary. So we have a ‘don’t cut any corners’ type of philosophy. The tradeoff of this is we are probably higher cost than our competitors. We probably develop products more slowly than our competitors, but I would argue that we also provide up a product that is in the end, higher quality, more secure and more private than our competitors. You must make trade-offs all the time, and in any sort of project, any sort of business, you need to decide where in that space you want to position yourself.

Ben: And who do you see as the competitors? Cause presumably you would put people like Gmail in a completely different category. They are not a direct competitor.

Andy: I think today in fact the competitor is Gmail, cause if you look at where most of our new users are coming from, they’re coming from legacy email services. They’re coming from Google, Microsoft, Yahoo, even AOL sometimes, so they are in fact our biggest competitor. And if you want to think about the future, that’s the target that we need to aspire to go after.

Ben: But is there another secure email service offered on a freemium basis? Is there anybody else who has exactly the same business model as you out there?

Andy: Oh yeah, of course. After we entered the space and opened up the space, a lot of people came in as well. This is sort of normal for any other space, and I think that’s actually a positive thing. It’s good to have more competition, more activity and more attention on your space because that drives innovation, that drives growth and competition’s always good in my mind. Us being sort of the first movers into this space and being by a significant margin the biggest player, that’s something that is quite durable because there is a network effect here..

Ben: Any business with network effects tend to be ‘winner takes most’ markets. That’s right.

Andy: And I would say communication apps in general also follow that sort of trend. So that is inherent advantage for us being in this space. But at the same time, we’re also not the ones who just kind of sit on our hands and say, we’re happy. We want to push innovation in the space. We want to continue maintaining our level of quality and we want to keep doing new things and doing things better and better. So that remains our key focus because we must do that if you want to take down the real competition, which is Google.

Ben: Yeah. I think what you’re saying is interesting because when you have got a business that is underpinned by network effects, the temptation is just to sit back and let them kick in, because it becomes such a strong moat. But what you’re saying is you’re not resting on the network effects moat. You’re also pushing the innovation moat as well.

Andy: Yes. Because network effect moat protects us in the encrypted space, but that’s maybe the 2% of the market that we’ve captured so far overall. The opportunity is the other 98% of Google users who are not on the system yet, and to go after them, you must innovate. You must be better. You can’t just be more secure and more private. You also need to be maybe easier to use, faster. You have to look nicer. It needs to be more stable. You’ have to compete on all these other things that people care about as well.

Ben: Yeah, they do. And would you argue today that your service is it’s the most widely used? Would you also argue it the most secure?

Andy: I would say so, yes. What I also want to say is there’s no such thing as a hundred percent security. By definition it doesn’t exist. Any system can be compromised and can be hacked. All we can do on our side is adhere to best practices, don’t cut any corners, do things as carefully as we can and engineer the cryptography in as strong a way as we can. So these were things that we try to do. It’s a reason that we use cryptography that’s open source and open standards which have been embedded by the community. So we do our best to check every single box, but I think it’s irresponsible and not honest to claim that things are 100% secure because they by definition cannot be 100% secure.

Ben: So I would imagine that everybody that’s listening to this now wants a ProtonMail account. I want one. How do we get one? And if we’re a business, how do we port all of our existing email addresses onto the Proton network?

Andy: Well, if you are a consumer, you just go on the website, get an account. That’s it. If you’re a business, we have an importing tool and we also have a customer support team who is there to help you with the migration. And we’ve migrated probably hundreds of companies just in the past month. It’s something we do all the time and we’re also improving these tools so they’re getting better and better every single month. Our whole ride is to make it easy.

Ben: Just to get back again cause I think we talked about this scale of the ambition for the company, which is great, but what was the sort of Eureka moment where you were like, okay, that I’m going to stop trying to discover whatever you guys are doing at CERN and if you want to tell us what people do at CERN, that might be also be interesting, but how do you go from that to something like okay, I’m going to launch a secure email system service?

Andy: Well, I think as scientists we have natural curiosity. So there really wasn’t an Eureka moment in the sense that you thought about the problem and said okay, let’s do that. It was more like; this is something that I, myself would like to have. I like to have email that is secure and private, and if you’re a scientist, I guess the benefit of that is you can actually build your dreams.

So we just actually went out and built it ourselves cause we had the know-how to do it. And after we did that, we simply released to the public. We said, okay, we built this. Now you can come and use it. And when we released it, we discovered that a lot of people had the same desires and same wishes as us, when it came to security and privacy, and this was something that we didn’t expect. We didn’t anticipate, we didn’t project or foresee. It was just let’s do it for the sake of the science. Let’s do it for the sake of having technology, and then, as they say, the rest is history.

Ben: The last thing I want to talk about, so when I saw you speaking at the Investor Day here at Fongit, I mean I was struck by the ambition, keep talking about this. But the other thing that I found really sort of heartening was that a business with this potential scale is a) in Europe, because we don’t have many platform-type companies in Europe and b) in Switzerland. And so I think you’ve talked a bit about why you’re in Switzerland, because you worked at CERN centers in Switzerland. There’s an inherent advantage because Swiss privacy laws are very robust. But are there other advantages to being in Switzerland? Have you at any point thought okay, this is such a big business now. I think we need to take this to California?

Andy: Yeah. Of course, we get asked this question quite a bit. Investors in particular ask this question so why Switzerland and why Europe?

Ben: I’m very pleased that you’ve chosen Switzerland and Europe, so, I’m not challenging you…

Andy: I think it’s an important question. I can start from a couple levels. What is the most important thing for a tech company to survive and thrive? I would say the most important raw material, if you will, of our industry is raw talent, and if you consider Europe, not Switzerland but Europe as a whole, this is a population here of around half a billion people. That’s a bigger population than the US and it’s a highly educated population.

So if you can leverage the potential of all of Europe in one company, then in fact you have access to a bigger talent pool at a lower cost and you have more raw materials here in fact than in the US, so that is from a high level why it makes sense. And I think the fact that we want to do it in Europe and leverage its potential also explains Switzerland in some ways because Switzerland is maybe the most cosmopolitan country in continental Europe. It’s the place where you can bring Germans, Poles, Spanish and French and Romanians wherever into one place, and they can all feel at home and feel welcome here, and it’s only by leveraging all of Europe that we will succeed.

I think Switzerland also has various, in addition to the strong laws, it also has a lot of support structures for startups like us. There is Fongit where we are based, which is offering a lot of support in the office space, organization, management in helping us to operate the business. There’s State of Geneva, which is very open providing whether it’s funding, advice, whether it’s tax breaks or other sort of support. These support structures make it a lot easier for us to run the business and focus on growth instead of worrying about other things. So I think it’s this combination of factors that made it so that we were able to succeed in Switzerland and continue to grow within Europe.

Ben: I don’t know if you were there, but Neil Rimer from Index Ventures, he gave a presentation on where he thought Switzerland was at in terms of attracting and retaining great digital-age businesses. And even if you weren’t there, the slide was circulated on social. You may have seen it, but he basically had pluses and minuses and I think there were more minuses than pluses, and you’ve talked about many of the pluses that he talked about. We haven’t addressed the negatives. So he was talking about the high cost of living, the difficulty sometimes getting visas for people. Have you encountered those kinds of problems? I guess the high cost of living, for sure.

Andy: So actually, on the visa and permit side, having a close relationship with the State of Geneva has essentially solved that issue for us. So we’ve never had any issues there. I would say cost is in fact a problem. The cost of living is very high. Geneva is better than Zurich in the sense that you’re close to the French border, so you can sort of reduce the cost of it that way. This is why I recommend actually Geneva over Zurich, which is not a popular opinion in Switzerland, but that’s my opinion. But I think you can’t build a business like this just in Switzerland. You also need to leverage other offices, other locations in Europe, and that’s what we’ve done. We have other offices in other locations throughout Europe which allow us to bring the overall cost basis down even further.

Ben: Where are your other offices?

Andy: So we have offices in Prague. We have offices in Vilnius in Lithuania, and also in Macedonia. So these locations help to kind of lower the average costs in Europe. Then I think the other thing is when it comes to costs, yes, Switzerland is expensive, but you have to look at the alternatives. If you’re not building a tech company here, then the only logical place you would actually go is San Francisco. If you compare Swiss costs with California costs, in fact, we’re quite a bit cheaper here.

Ben: Really?

Andy: Yes.

Ben: I am slightly surprised cause so housing, wages, these things are now more expensive in California?

Andy: Yes.

Ben: Okay.

Andy: Yes. So it’s completely shifted in the past five years. I think in 2014 they are about the same. Today we’re maybe around 30 or 40% lower in Geneva compared to San Francisco.

Ben: Wow!

Andy: Yeah.

Ben: And I also liked the fact that you’re challenging the received wisdom of Geneva over Zurich. That’s cool. Earlier on today as I was introduced to some of the members of your team and the first thing that you observe is it’s very multicultural. The second thing is these guys are very young. Very smart and very young. The third thing you observe is there aren’t actually that many of them. So would you also argue that to create a platform business today, you don’t need hundreds of thousands of people that you might need to create a really large industrial age business? So that also perhaps gives an advantage to places like Switzerland where you can get very high-quality talent and you don’t need hundreds.

Andy: I think to scale you do need a lot of people. There’s no way to get around the need for people. But it is true that if you hire very smart people, very talented people, and that you leverage the newest technologies, it’s possible to do a lot more with the same amount of people that would have been possible even one or two years ago. So, in fact it’s true. You don’t need that many people to run a pretty massive business, but you still need people. It’s still your main raw material and the fight for talent is still going to be the main thing. And the other reason I say Geneva instead of Zurich is also kind of mindset and competition. The competition is all for talent and in Zurich you’re fighting against some very big established players and…

Ben: Including your nemesis…

Andy: Including Google, yes, but not only that. It’s that you’re in a talent pool that is much more conservative, much more, I would say risk averse, and much more corporate in a certain sense, whereas Geneva, I would say people here are more willing to take risks. They’re more of the startup mindset, and that’s another reason why we build our base here and I think we would actually be maybe more successful than we would be if we would have gone to Zurich.

Ben: Last question. We’ve talked about the company journey. What about your personal journey? So you’ve gone from scientist to CEO of what is now becoming a really big business. What has that transition been like for you?

Andy: Well, I think in terms of what I do on a day-to-day basis, obviously there’s a big transition because I had to learn new things, figure it out, learn from mistakes obviously, and also do things that maybe I wasn’t trained to do in my schooling. So that’s obviously a transition. But I think what is also very important is yes, you need to change and evolve, but it’s also important to stay the same in many ways, to hold onto your core ideals and remember and keep in mind what is important. And I think for our business to succeed, we must also keep in mind always that’s our main focus. Why are we here? Why do we get up in the morning? It’s not to create a huge fortune. It’s not to become the next Google.

We don’t want to become what we’re fighting. What we want to do is we want to stay true to our ideals — freedom, democracy, privacy, security, serving the world, serving the community and serving the users first. And that’s something that I think my role now in the company is to make sure that we retain that for as long as possible and to make sure that the new people that come in understand that that is what makes us different. And that is what makes us important and able to carry out the mission that we want to do.

Ben: Wonderful. No better way to finish than that, with your description of why this business is so important and why it’s so different. I think it’s inspiring and I’m delighted that you’re based here in Switzerland. Andy, I just wish you all the best with your mission, and I hope the trajectory carries on being as exponential as it’s been. And I thank you very much for coming on the podcast.

Andy: Thank you.

The Ethics of Technological Change (#9)

Structural Shifts with Jim LARUS (Professor and Dean of Computer and Communications Sciences at EPFL); Philippe GONZALEZ (Senior Lecturer in Sociology at University of Lausanne) and David Galbraith (Partner at Anthemis VC).

This podcast was recorded at EPFL Lausanne and in it we’re exploring the ethics of computer science and broader technological change. Ben Robinson is joined into the conversation by: Jim LARUS (Professor and Dean of Computer and Communications Sciences at EPFL); Philippe GONZALEZ (Senior Lecturer in Sociology at University of Lausanne) and David Galbraith (Partner at Anthemis VC).

Ben Robinson (BR): Welcome to Aperture. For this episode we are at the EPFL and we are discussing technology in society. And for this conversation we are with Jim LARUS, who’s the Professor and Dean of the School of Computer and Communication Sciences at EPFL. We’re with Phillipe GONZALEZ, who is a Senior Lecturer in Sociology at the University of Lausanne. The two guys teach a course together which is called Global Issues and looks at the social implications of technology. And then we also have David GALBRAITH with us who’s a Partner at Anthemis and a regular on the podcast and David’s here to make sure I get value from this conversation to ask him some provocative questions. So Jim, Phillipe maybe we can just start by you just explaining Global Issues, what’s the course and the subject matter that you cover in the course.

we want the students to appreciate that there’s a larger world out there, they are policies, they are social issues, they are legal issues, they are sort of unintended consequences from technology. That is fairly common, it’s nothing to sort of be surprised about but it’s something that a lot of students are just unaware of — JL

Jim Larus (JL): Sure happy to do it. Global Issues is a course that we require of all of the first year students at EPFL. Doesn’t matter what subject they’re going to take at EPFL and its taught in sections of about 100–120 students with a particular focus to each section. The section that Phillipe and I teach is called Communications, which we’ve interpreted to mean anything related to the Internet and that’s pretty broad, we cover a lot of technology that way. But the idea behind the course is that the students here are basically going to spend five years studying science, technology, engineering and it’s going to be a very technical, very focused education, they’re going to come out with a lot of skills but they may not sort of realize that what they’re building or what they’re discovering is going to be used in a larger context and people are going to take their inventions, are going to take their discoveries and use it for whatever purposes they want and so sometimes these uses of technology have larger implications. You know obviously something like the Internet has had tremendous impact on society in ways in which were impossible to predict 10 years ago, 20 years ago when it really took off.

And the idea of this course is to get the students to actually think about this during their education instead of finding this out when they get out into the real world and start realizing that oh, what I’ve learned, what I’m doing actually has implications for the broader society. Instead we want the students to sort of be aware of this and start thinking about it as principal, ethical people they need to think through some of what they do sometimes they actually need to be advocates for a particular position. They need to take a stand as to how their technology is going to be used because those are going to be the ones developing it you know particularly our school with computer scientists a lot of the information technology is going to be developed by students who have backgrounds similar to ours who will be working in companies. And so we want the students to appreciate that there’s a larger world out there, they are policies, they are social issues, they are legal issues, they are sort of unintended consequences from technology. That is fairly common, it’s nothing to sort of be surprised about but it’s something that a lot of students are just unaware of.

BR: And how did you decide or when did you decide to introduce this course. Was in response to some of the scandals we’ve seen around Facebook and others or was it just something you thought was missing from the curriculum?

JL: I didn’t actually introduce it. It’s been here for longer than I’ve been here so it’s longer than six years. I think keep EPFL decided on it probably about a decade ago, not in response to anything in particular. I mean a lot of the sections of the course have nothing to do with the Internet, nothing to do with the recent scandals so you know there are sections related to health related issues, there are sections related to climate change, sort of larger societal issues as well as information technology.

BR: So in a way, if I was quite early with this then because you don’t see many of these courses and yet a lot of people are talking more and more about the importance of grounding tech work in philosophy and discussing the broader implications.

JL: Yeah, I think that’s true EPFL was pretty advanced at it.

Philippe Gonzalez (PG): Yeah it started more than a decade ago with what we call Le Collège des Humanité (the Humanities College), which was the idea that engineers here at the EPFL would really benefit from being exposed to social sciences, history and it was a way of having this collaboration between the University of Lausanne and the EPFL. So we’ve been having social scientists, historians, people in literature giving classes here but the next level was teaching together and I think that’s the strength of this Global Issue Curriculum because every week we’re teaching together two hours and one hour the lesson about technology, about the specifics of the Internet or cryptography and then they get to listen for an hour about sociology and history. And I try to show them that some things in anthropology for example the invention of writing have a major impact on the way we can see the internet.

BR: So I guess in general you would argue that it’s important for any student to consider the broader implications of what they’re working on but I guess would you consider even more important when people are working with technologies that can have almost by definition such a massive impact? And maybe I can express it more provocatively? Do you think that the Facebook scandal might not have happened if philosophy had been taught more broadly?

JL: (Laughs) Depends of who they were taught to.

Media is a way of organizing a society, a way of defining what is a common good, defining who’s in and who’s out. So with the Internet we have new ways of relating one to the other, of creating collectives and nations and this has an impact because it’s redefining what is public and what is private in a way that’s never been thought before. — PG

PG: I think that first of all, what we try to tell the students is that some of the things we think are new and technology and are not that really and some of them have been around for hundreds or thousands of years and they already had implications on the way society was made, thought, enforced. So one of the first things I tell the students is for example that when writing was invented it was about kingdoms and ruling kingdoms and ruling the economy. And writing was not something for the people, it was for the elite and so it was a very vertical society and that goes back to the Neolithic, to the invention of agriculture. Well, writing has been around for 6000 years but it was to enforce all those old societies like the pharaohs in Egypt or Babylon for example. So I’m telling them see in the internet you have that old technology that’s there and at the same time we have things that come up, that come from the French and the American Revolution because we have this idea of the Free Press. So we have to deal with the Internet, it’s a complex tool and it inherits previous technologies like the printing press and like the invention of writing. The invention of writing was to control society. The invention of printing and the revolution was to think of a more democratic free society and we have the two things combined in the Internet and now we have to be aware of which side we’re pushing.

David Galbraith (DG): I was going to ask this one thing. So the invention of the printing press certainly had a role in the Reformation and it literally rewired society. And we can prove this by looking at let’s say the distance from mines to cities like Hamburg or Seville or places that literally invented not just new… new types of organizations: so joint stock corporations and the replacement of feudal societies and with what we today think of as corporations happened because of that. Now we have something which seems like a similar phase change. We have technology where everyone has a channel, everyone has a voice, and we’ve gone beyond the broadcast medium of printing. Is this going to have a similar phase change in terms of the types of corporations, the types of structures that we will see?

We start seeing the implications of applying new technology to politics, where you can narrow cast your political message to a group that agrees with you. And then we lose one of the arguments for free speech which was: you put your ideas out in public and if people disagree with them, if they think they’re wrong they argue against you. Well that doesn’t happen anymore on the Internet. — JL

PG: Absolutely, it changes. I mean in media you have always to think that a media is a way of organizing a society, it’s a way of producing goods, defining what is a common good, okay and defining who’s in and who’s out. So with the Internet we have new ways of relating one to the other, of creating collectives, collectivities, nations and this has an impact because it’s redefining what is public, what is private in a way that’s never been thought before. For example I can have… That’s one of the things you say in class, if you’re not paying for it you’re probably the product. So it’s redefining the way as a consumer, I’m entering the market and I feel like I’m the one buying something when in fact I’m the one who’s being sold through the marks situation and lighting on my path and I’m not aware of that. And that’s, well now regulations nations are taking a stance, for example now if you click on the Internet in Europe, so you have to be okay with the policy about cookies but it’s a new enforcement because we started to see the consequences of being tracked so we’re starting to be aware of where this new technology is pushing our democracies and we’ve seen that with the Cambridge Analytica scandal around the Trump election, we see that well… social networks are not only to find old mates from my school, there are also ways, powerful ways to influence geopolitics.

DG: And so is, the current debate is focused a lot on privacy but one person privacy is another person’s secrecy so how much is this not just an issue about privacy per se but an issue about balance. And for example in the Reformation we had… before society settled down into new configuration.

JL: So I mean, one of the points we make is that none of this is all good or all bad. And we actually explain targeted advertising to the students as part of it. And you know it’s frankly it’s amazing to me but it’s a revelation to many of the students that this is even going on and they’re very well informed they’re smart but they have no idea that they’re actually a product being sold. You know I think that you know Google gives them a search engine for nothing, Facebook gives them this connection to their friends for nothing and that you know they’re doing it out of the goodness of their heart. They don’t realize sort of the flip side of the market.

And at that level you know a couple years ago didn’t appear to be too bad but then you sort of take this technology which has been honed by a lot of very smart people and become extremely, extremely good. And you say what other use it’s going to be put to and you say politics is a very obvious one and then you start seeing the implications of applying this technology where you really narrow cast your political message to a group that agrees with you. And then we lose one of the arguments for free speech which was you know you put your ideas out in public and if people disagree with them, if they think they’re wrong they argue against you. Well that doesn’t happen anymore on the Internet. And so we try to sort of explain to the students as part of this that it’s not a simple process. The technology is there, the technology can be used for many good purposes. You know ads that are shown that are relevant to you are certainly more valuable than ads that are showing that are totally wrong for you and both to you and also to the advertisers. But there are things that can happen because of this that are not predictable. I don’t think anybody you know 10 years ago when Facebook was much smaller company would have said that it would have changed the election made Donald Trump the President of the United States. But clearly it has.

DG: Do you think it is Facebook? Or do you think it’s the configuration of the network? It’s in the, there is a possibility…

JL: If it wasn’t Facebook, it would be some other social network. I don’t think Facebook I mean, I think Facebook is particularly effective because of its scale and because it’s a social network. The value of it goes up with the number of people and it basically has everybody… so its value is much higher than other social networks. But if it wasn’t Facebook, there would be some other social network… before it was MySpace. You know, Facebook hadn’t come there would have been something else.

DG: But are there any types of social networks that you think in that configuration are more powerful. So the President of the US uses Twitter for example not Facebook traditionally.

JL: He uses Twitter because Twitter is a perfect medium for him to just broadcast without any filtering. But he’s not engaged in a discussion. If you’d like to prevent a discussion… you know he is fighting in Court to be able to throw people off of his Twitter feed. Which you know in the US seems to be sort of blatantly impossible but he’s still arguing that he should be able to do this so it shows that it’s sort of but not a social network in some senses, it’s a super powerful mechanism for doing press releases.

DG: Right, and anti-social network.

the old Anglo-Saxon philosophy that led to free-speech was premised on this idea that it would be beneficial in eliminating incorrect or bad ideas because they would be put into the public, into the market. That is definitely not the way the Internet works. You don’t see what’s going on, you have this amazing ability to find people who will agree with you because of the advertising technology. So I’m a little wary of taking all these arguments for free speech and just saying “yeah, they should apply to the Internet” – JL

BR: So as I said we will come back to questions of monetizing the internet and the trade-offs of privacy and we also want to come back to the ethics of data-science…. but we just want to get you a bit more on this subjects of discourse and the quality of discourse which is… to go back to this analogy of the printing press and the Reformation. Initially things got worse but eventually things got better right? You know the Demos became better educated and better able to hold society and democracy to account. Do you think I guess the question is are you guys optimistic that because as we’ve been discussing that in this new world the discourse has been challenged because everybody has a voice and people with more extreme views potentially have an outsized voice. And so we’ve lost this sense of balance but do you think that over time we will establish the checks and balances such that the discourse will improve and this will be a fundamentally good thing for society?

JL: But you have to hope so… Not as deadly as that time because we have a lot worse arms than they did back then. You have you have to be sort of optimistic that things that these transitions are painful but that you end up at a better place in the end. This goes back to one of the other points I make is that on the Internet people sort of underestimate the radicalness of the change. I mean until the Internet occurred to disseminate anything, disseminate information it had to be physical, if I wanted to sort of share something with you I had to write it down on a physical piece of paper. And if you wanted it I somehow had to get you a physical object. There was this cost to it and that served as a moderating function. Book publishers in general wouldn’t go publish sort of complete nonsense because nobody would buy it and they would lose was money on it and they would lose the money up front to produce the copies of it. So there was this moderating influence even though there were plenty of extreme examples. This is God right?

The cost of disseminating something, for better or worse and I think of it many cases it’s definitely for better is essentially zero, right? I can communicate with pretty much anybody, anywhere in the world at zero cost, I can send them large amounts of information at zero cost and zero amount of time. We’re not used to this, people, even before the Internet really took off back when sort of all the academics were using it so this sort of phenomena where people behave very poorly in the groups online. The norms of the behavior which they would, they would do things they would never do in person, never do face to face with another person. They would have no qualms about so calling people names, insulting people, attacking their motives, when they would just never do this in the real world. And we saw this and we sort of, it was kind of a strange phenomenon but it was marginal at the time of small groups. We now see it on a much larger scale and we see it being utilized for bad purposes I think for a lot of politicians and a lot of political movement. This is going to be a hard thing to get over because it’s clearly something in us where in an animoty or even just distance from another real person lets us behave in ways that, as members of a group we don’t behave in. And it’ll be hard to reconcile that with the fact that we now have a network which has global reach, which allows you to do this with anybody anywhere in the world. So in some sense the two movements are very much going the wrong direction.

DG: And do you have a view on the political dimension for this because the left and right have very different views of, if there is such a thing as left or right anymore but they certainly seem to be operating very differently in terms of issues such as diversity for example or whether people who really care about minority issues should have a voice and whether the majority should dominate the discussion or not… so these things are nuanced. There’s no straight answer but there’s certainly very, very different views on these things. Do you have any view on this?

JL: You know I’m an American so my view is sort of shaped by the view in the United States that free speech should be allowed and encouraged up to sort of fairly clear but pretty far limits like advocating violence or injury… but I have to say that I think a lot of the argument for free speech… a lot of this sort of old Anglo Saxon philosophy that led to it was premised on this idea that it would be beneficial in eliminating incorrect or bad ideas because they would be put into the public, into the market and that is definitely not the way the Internet works. You don’t see what’s going on, you have this amazing ability to find people who will agree with you because of the advertising technology. So I’m a little wary of taking all these arguments for free speech and just saying yeah, they should apply to the Internet.

DG: So the Chomsky idea that you can manufacture consent, there’s this idea that maybe you don’t even need to conspiratorially do that, maybe that’s a flaw in the way human beings are that we, we self-configure dystopia?

The printing press in the Reformation brought a new technology and a new society. We had war but we also needed strong institutions that showed us how to use that technology and I’d like to mention two institutions. One is what what we call the encyclopedia […] we’ll try to share knowledge and to invoke reason and that will be a strong tool against despotism, against absolute monarchy…. The second big institution is free press, a concept which started before the English revolution, but the ideas we have about privacy, they come from the French and the American Revolution so we’re still living in that world. Actually right now we have a powerful media that is undermining those ideas. And we’re still trying to come up with new concepts to understand the links, which I mean the articulation between publicity and privacy. — PG

PG: Well actually, that’s really interesting. I believe it’s something inherent to being a human, a human being the fact that you have to trust someone else to give you some information. Let’s say, if I want to know what it was like for my father to be a kid, I have to ask him and he’s telling me a story. So one of the ways we get knowledge is to listening to someone telling us a story. So that’s a default position but I think it’s a serious advantage of our species when it comes to evolution because then we can share, we can trade and we can transmit that knowledge from generation to generation and we can spread it across the globe. So that’s one of the things but of course trust can be twisted.

So I think one of the issues with the Internet right now because it started on a paradox. It started with; it was it was something military; it was a military experiment, who was taking over, in a sense, by libertarians who were advocating for free speech, for a new continent. So you have the two things in there, built in the Internet, you have something like a kind of centralized control even if the military experiment was about if a bomb is dropped in a place and I want a signal to go somewhere, I can bypass that place and I can send the signal… but it was this idea of a chain of command and we see that right now. All the fears we have about facial recognition and the way China is collecting information about its citizens.

That fear is a fear that democracy as we’ve known it might disappear. Right now the UK is investing a lot of money into cameras of video surveillance and now the issues to know if they’ll match the cameras with software that have facial recognition. And that’s a major issue because I mean, in a democracy to we have to film and recognize the face of a citizen walking in a public sphere and into public sphere and in a public place, that’s a serious question. So one of the things is that yes the printing press in the Reformation brought a new technology and a new society. We had war but we also needed strong institutions that showed us how to use that technology and I’d like to mention two institutions. One is what what we call the encyclopedia. So you have all these guys, Didaut, Voltaire, who were philosophers and said okay, we’ll have something like The Republic of Letters, the Republic of the academics and we’ll try to share knowledge and to invoke reason and that will be a strong tool against despotism, against absolute monarcghy…. okay and actually they won. I mean the world we’re living in, the society we’re living in, in Europe and in America. It was made by their ideas. The second big institution is free press. Free Press started before the English revolution but it truly thrifts under the French and the American Revolutions and you know the ideas we have about privacy, they come from the French and the American Revolution so we’re still living in that world. Actually right now we have a powerful media that is undermining those ideas. And we’re still trying to come up with new concepts to understand the links, which I mean the articulation between publicity and privacy.

DG: This is not something I agree with but recently in a conversation from someone who’d lived in China for seven years said that when the West are getting it wrong, that the Chinese see the inequality in Hong Kong and they do not want this to spread to the mainland and that the reason that they’re clamping down is because they want to keep you know… Xi Jinping wakes up every morning and has one thought on his mind, which is to keep the country together. And that the only way you can see that happening is not let inequality run wild and in previous periods of technological progress, we know it creates enormous amounts of wealth but it also and Piketty has argued this recently that it creates inequality, it seemed to happen in Rome and it created an enormously long period after fall of Rome, of people being interested in stories rather than technology and reason.

The Internet started off as a totally decentralized organization. There was very little control, it certainly was not top-down and it was by design. It was taken over by academics, by libertarians and it was actually very badly built because of this because nobody sort of thought about the issues of what would happen if you put all the world’s population on it. Just give a very tangible example… there’s no notion of identity on the Internet. — JL

JL: But you know, China has vast inequality; they have the second largest number of billionaires after the US. And if you just go away from the coastal cities where the money’s been created you know the 300, 400 million people that are still left on the farms there… are poor, they’re not as poor as they used to be but there’s a huge gap between them and their children or whoever’s you know gone to live in the cities but without the permanent residents of the cities without the benefits of being permanent residents of the cities. So I think China has an inequality problem and I think a lot of the control of the Internet and get back to that in a second is very much related to political… that they want to show the dominance of the Communist Party. And yet a lot of their concern with Hong Kong is that it’s a revolt against the authority of the communist party.

DG: In fairness is some evidence of the inequality within the big Chinese cities is dropping, whereas Hong Kong’s is increasing.

JL: Yeah. So the inequality across China is still fairly large but just going back to that, I would sort of correct a little bit of what Philippe said, I think that this is one of the interesting parts of the course is that the Internet in my view started off as a sort of a totally decentralized organization. There was very little control, it certainly was not top-down and it was by design. It was taken over by academics, by libertarians and it was actually very badly built because of this because nobody sort of thought about the issues of what would happen if you put all the world’s population on it so and just give you a very tangible example is that there’s no notion of identity on the Internet. So you can’t tell who you’re communicating with, I could put up an account name Donald Trump and you know except for the fact that I would be more reasonable you couldn’t tell me from the President of the United States. You know this is a fundamental flaw that people will look at it now and say, how could we have gotten it wrong? but at the time it made a lot of sense to build an internet that was simpler without the sort of strong trust, the cryptography that we needed to do it right, didn’t exist at the time. So would have been a very difficult task and for a long time people thought this is actually strength of the Internet that it’s decentralized. There was no control, it wasn’t top-down and you know that there was a sort of saying in the 90s, “the internet sees censorship and routes around it”. But China prove this totally wrong basically that if you have enough money, technological smarts and people, you can control the internet. And so China has one of the sort of most interesting internets but it’s a different internet than we have, commercially they use it for far more things than we do. It’s a mobile internet as opposed to… sort of more desktop based internet because they started later. There are a couple of apps there that are sort of much more interesting than the apps that we have on our phone, things like We Chat, which allows you to pay and sort of interact with people and sort of ways that we can’t do in the West, because we have more static banking system. So you know they have the benefits of a lot of the Internet but they have sort of absolute political control. You cannot use it for sort of expressing political opinions that are counter to the government. You will be found, you will be stopped, regardless of whether you have strict identity or not. So this is interesting. I mean 20 years ago people would have said no, you couldn’t have done this with the Internet and China proves them wrong. And it’ll be interesting to see in the future which internet wins.

BR: We want to come back to this question. So the different, the different manifestations, the different types of internet depending, Chinese internet, the US internet, whether there’s even a concept of the European internet…

… but I just want to go back to something that Philippe was talking about when he talked about the institutions that grew up in response to the last significant change in the way we communicate. Do you think that those, the balances and the institutions we need for the digital world will happen organically, will people configure their activities in a way that’s different or does it take the intervention of the state? Because, and how would that work because we’ve talked a lot on other podcasts about GDPR and whether that’s really, actually any help at all. So does the State need to get involved and if the state needs to get involved, what might that state intervention look like?

We’re coming to the Tragedy of the Commons. Before, two solutions have been offered. One is privatization or option number two is the state takes care of it. But then came up Nobel-prize economist Elinor Ostrom and she said “well actually we can think about collective action, the commons, the way we share a common.” And I think that’s one of the most interesting ways of thinking about building new institutions for the web. It’s neither the individual, nor the state. If I share a common purpose and enter into a collective and we start thinking together like in the wiki philosophy… then it’s something in-between and we can arrive at some very interesting solutions. — PG

PG: That’s a very good question.

JL: You know, I think China has already intervened and sort of decided what the model is going to be in China. I think the US is struggling with it because of the, sort of free speech makes it very difficult for the government.

BR: Can I ask a different question then, which is like if you take a platform like Facebook. Does it have a duty to protect the truth? And can that even be done at scale?

JL: So who’s answering that question, I think are hands on. So the answer in China is it has a duty to protect the truth as defined by the Central Party period. In United States there is no duty because it is a private institution. And private institutions are not covered by the Constitution… there’s a very permissive law, which allows them to sort of get away with taking no responsibility for what’s posted on Facebook. In Europe I think it’s different, I think there is a sort of wider range of opinions about the obligations of companies and sort of the ethical point of view that companies have to take responsibility for the consequences of their decisions, their actions, their behavior here and so there’s a lot more freedom to negotiate and we’ve seen this with Germany for instance, forcing them to take down all mention of Nazi positions quite quickly. So I suspect that the answer will depend very much on where you are and where you live and what governments you’re under. And things like GDPR which have this sort of far reaching implication or even the earlier version of its which was the sort of right to be forgotten that you could sort of remove information from the Google search index.

BR: The interesting thing I guess about the GDPR approach is it’s trying to give us as individuals more power over our own privacy and our own data. But you yourself said that you’ve got students and you know… these students that come to EPFL this the top 1% of students right in terms of academic ability and if they’re not even aware of some of these tradeoffs…

Another way to look at GDPR is that it didn’t go far enough — it basically just gave you informed disclosure and consent, so you can say yes or no, but really what was missing was the fact that you’re giving something to this company that you’re interacting with and they’re monetizing it and they’re not paying you anything for it except by providing a service which may not reflect its value. So one very interesting idea is this notion of selling your information to companies — it becomes an actual transaction, but then you can make this trade off “do I really need the money, or do I want my privacy”, because right now the tradeoff is “do I want to use this website or do I want my privacy?” because you do want to use this website. — JL

JL: But it’s even worse than that actually, you sort of look at experience with privacy or security. You get this pop up on the screen that says this is not a good idea of the following reason; do you want to do this? There’s a huge amount of research that says that 99 out of 100 people are just going to click yes, I want to do this because you’re in the middle of doing something and this is getting in the way and you just want to get it out the way, you’re not really sort of thinking about the fact that it may sort of open up your machine to all sorts of problems or attacks or something like this. And the same thing with GDPR, I would predict that probably same thing. 99% of people see this pop up that subscribe to the view of their screen. They just click yes without sort of seeing it.

BR: That I suppose the counters, if the answer to this isn’t to empower the individual to take more responsibility then the answer has to be for a larger body e.g., the state to intervene somewhere.

PG: Well actually what you’re saying is really interesting because it’s… We’re coming back to something called the Tragedy of the Commons and it is this idea that came up in the 60s that if you have some land and you have sheep feeding on the land and I have got my flock, you’ve got your flock and we don’t coordinate, by the end of the week there is no more grass and by the end of the month I mean we just killed the land. Okay, so the idea is what can we do, so two solutions have been offered. One is privatization, I buy the land, I put a fence around it and that’s it and I’ll make sure I take care of the land. Or option number two is the state buys land, puts a fence around it and takes care of the land. The thing is one of the problems with individual action is that usually we don’t pay much attention to consequences… but then came up, very interesting economist and she got a Nobel Prize for it and her name is Elinor Ostrom and she said well actually we can think about collective action, the commons, the way we share a common. And I think that’s one of the most interesting ways of thinking about building new institutions for the web. It’s neither the individual because if it’s only about my personal responsibility, I’m not aware or I’m not capable of having the craft or having the knowledge, the tools, but if I share a common purpose and enter into a collective and we start thinking together like in Wikipedia and everything that has to do with the wiki philosophy then it’s not only individual, it’s not the state. It’s something in-between and we can… It’s more plastic. It’s less legalized… but we can arrive at some very interesting solutions.

So now I have an example, I mean, it’s not only about the wikis, it’s also about the issue of law because it’s not we’re talking about a technology that has an impact on your life and for example if you get your credit card number stolen okay, what can you do? Of course, you can go to the police and file a complaint okay. And then where can you voice your problem? You can voice it on the forum of your bank you know, under a website but it remains an inevitable complaint. But if you start to merge with other people whose card has been stolen and you build a collective and then you start lobbying at the state to change the law. You don’t have the state meddling into the technology because it wants to overpower the technology; you have citizens concern by the consequences of a certain technology asking to change the law to fit actual situations. So I would think that that’s the best way to build new institutions around the technology. That means that we’re not building it from scratch, from a needle idea. We’re not building this technology from a perspective that is ideological; we’re building this new technology through practical situations and through the consequences of those situations.

DG: I mean these kinds of pacts between the mass and the people who are rich basically… so that in the industrial era, you had collective bargaining through unions and things like this. What happened? That was when one bunch of people instead of working in farms, were now operating the machinery that you needed people to operate the machinery. What happens when you have robots and basically the people who have the capital also own the machines who operate themselves?

JL: So before we go to that which I think is a great topic, let me just add one thing to this thing is that another way to look at GDPR is that it didn’t go far enough that you know, it basically just said, we’re going to give you informed disclosure and consent. So you can say yes or no, but really what was missing was the fact that you’re giving something to this company that you’re interacting with and they’re turning around and monetizing it and they’re not really sort of paying you anything for it except by providing a service which may not reflect its value. So one thing that has been explored and I think is actually very interesting idea is this notion of sort of selling your information to companies. And so it becomes an actual transaction where you say yeah, you can track me, you can sort of use it to target advertising, you can use it to sort of sell information to your advertisers but you have to pay me whatever the unit is per your interaction and then you know that obviously would be a contract, would have to be enforced by a government because otherwise it’s not likely to happen… but there the interaction becomes a lot richer, you can make this trade off you know, do I really need the money or do I want my privacy, because right now the tradeoff is do I want to use this website or do I want my privacy? And you’re given a sort of much worse choice than almost everybody will say I’m in the process of using this website, I want to use this website.

DG: Do you think that needs to be, the amount of money for an individual might be quite small, do you think that needs to be a data union like a collective?

JL: There is that view that you could sort of have larger aggregates of people that would make it much more significant, I don’t know if the amount of money is that small Facebook in the United States makes between $200-$300 per person for the advertising. I mean that’s by far in a way, the largest in the world. Many countries makes almost nothing but there you know, countries where there’s less developed advertising industry but you know is it going to make you rich? I don’t think so… because it only wins in the aggregates of any individual, that’s not going to make a huge amount of money for them but it would shift the balance quite a bit. And you would be able to make a much more intelligent trade off, is it really worth disclosing this information for a penny? Probably not.

BR: You might ask whether that trade off should be necessary in the first place anyway that right? So what you’re saying is that at the moment we give our data in exchange for a service and what you’re saying is we should be paid in exchange for giving our data, right and…

JL: Look at the verb you use, give like it was a gift of our data? And what I’m saying is no, it’s a commercial transaction; the company on the other side turns around this gift and sells it to someone else for a lot of money.

BR: So what I was getting at was like, so we wanted to come back to this side of the Chinese internet because there we’re many issues which I think we’ve already touched on with the Chinese model, but one of the positives is that because it’s built on a mobile internet and a much better payments infrastructure. People pay for the services that they subscribe to rather than having through some opaque transaction actually give that data away? Because…

JL: But they give their data there is a vast amount of information collection, much more than we would tolerate in the West.

DG: It’s almost the default is this, that data is exposed. But certainly if you look at the podcast market in China for example, it’s 10 times the size of the US one and it’s all people paying for content whereas in the US it’s…

JL: There is a micro payment system. So you can pay for sort of small quantities things without getting a credit card fee involved, exactly.

BR: Don’t you think that’s inherently the problem with the US internet, which is because there’s no concept of micro payments, what the next best model was and is probably for the foreseeable future, is advertising, which has two issues right. One is the sort of you dupe in a way the consumers are giving you their content and secondly it creates a fundamental conflict of interest right? Because are you ever acting in the interest of the consumer or are you always acting in the interest of the advertisers who pay you?

The press was very much like what the Internet is today. There were very partisan journals on both sides, and the level of invective and insult in them was extremely high. Maybe even worse considering the level of discourse in general than what we would accept today. Politicians were routinely insulted and accused of all sorts of terrible things and this was just considered to be a consequence of the free press. — JL

JL: So I don’t think it’s a tradeoff. I mean, I think advertising was going to happen on the internet regardless of whether we had micro payments, it was way too attractive a vehicle. They are too many smart people who saw the connection between what you were doing and being able to sell an ad that was related to that, to sort of have that not happen, even if there was a micro payment system. Same time, I think a micro payment system would enable a lot of really interesting things to happen on the internet because people could get paid for it without doing advertising.

BR: I think you’d see a large, much larger monetization of the internet with micropayments because advertising is, what, 2% of GDP? Therefore can have that surely put some sort of constraints on the size of…

DG: Certainly, so we have reduced the number of ads because they can, because there’s much more micropayment purchasing of content. So there is some evidence it does seem to change the model as to whether people pay for stuff.

We need institutions to deal with technology and to deal with the shape of our society. So yes, I’m optimistic but we have to work a lot. — PG

PG: And one of the things we try to say as well is that advertisement on the internet has an impact as well on the other media, starting with the press. And one of the problems we have with democracy is that the standards of the press are getting low and low and low because they can’t afford it right now to have investigative journalists, because it’s very expensive and it used to be funded by the money coming from the advertisement for the newspapers and now all that money is migrating to the to the web. And so we see all those huge newspapers really being pushed to rethink the way they do journalism to the point that they can’t afford except for the New York Times or The Washington Post or Le Monde. They can’t afford to have very good journalism done in the field, long term journalism. So that’s a real problem for democracy.

BR: Even those publications that you mentioned, like the New York Times that still have the money to do investigative journalism. hey’re still doing things like changing the headlines of their articles several times during the day to see what gets more clicks. Everybody’s drawn towards sensationalism because sensationalism gets eyeballs and eyeballs get advertising dollars.

JL: Yeah but you know I read the New York Times regularly, I would say that their investigative journalism has gotten better. They went into a trough but they were really worried about surviving and now the Internet seems to be their friend and they have some very interesting series. Now quite a bit, with a lot of resources and actually it applied in new ways, they’re able to sort of take very large amounts of data and then present the information visually in ways that you could just never have done on paper.

BR: So do you almost see the New York Times as a reason for optimism that these sorts of Industrial age institutions can adapt themselves?

JL: Let me bring something to the table and basically say you know the fact that you can do a podcast with technology in our office and you can make a business out of it. You couldn’t have done this without the internet. And you know this is another form of journalism, this is another way of disseminating information. It’s not about sort of current news but it’s about the bigger issues behind the current news, investigative journalism. I think that’s fantastic, I think that there’s a lot of opportunity for people to sort of put a lot more information out there in front of it and it goes back to the fact that there’s no cost to do that. So yeah, it has really hurt the newspapers, they had basically a monopoly on a certain type of advertising for a long time. And you know the TV came along and hurt them but they responded and, it’s painful the transition they’re going through and I don’t think all of them are going to survive clearly but you know, I think if you look at the big picture, there’s a lot of really interesting things happening now in terms of disseminating news, disseminating information and some of it is fake. There’s some of it has no standards. One of the things I was going to say earlier, which you reminded me of is that earlier in the history of the printing, the press was very much more like what the internet was today. You know, there were very partisan journals on both sides. And the sort of level of invective and insult in them was extremely high you know sort of you know. Maybe even worse considering sort of the level of discourse in general than what we would accept today. You know politicians were routinely insulted and sort of accused of all sorts of terrible things and this was just considered to be sort of a consequence of the free press.

BR: And what changed was that the populace became more educated and demanded high standards?

JL: At some point the press decided that they were going to rise above this and that they were going to be impartial but I think that’s relatively recent.

PG: Yeah, it’s in the 1920s. And you have this great figure Walter Lippmann; he wrote the book called Liberty in the News. And he was saying you see the press. It’s drawn between two opposites. It’s either the taste of the public, either the wealth or the richness that can be produced through advertisement. And he said the only way to counteract, to oppose, to fight against this current is to have a strong institution, which is journalism. It’s the ethics of the profession and since you start developing high standards of what’s journalism, then you will start to have a better press and he was correct. He was correct and the way he thought the press was enforced for 70 years until the Internet came and changed the rules of financing and as well as gate-keeping. Because what we’re changing here, what we have with your podcast for example is a different type of gate-keeping: who is allowed to speak, you come here to the EPFL and you select two scholars and you talk about them but you could do the same thing with two kids just doing skateboard in the street. So gatekeeping is one of the major tests of journalism to choose which information is valuable for democracy? And that’s what we lost in a sense because we opened the gates. Okay, so everyone now since if you have a mic and a connection you can produce and publish but the problem is that do you have the skills, you have the skills to define what good information is and that’s what we might lose?

DG: I was going to say that the… That point about gatekeeping that some of the evidence is that there are particular types of network topology that allow complex ideas to spread and because there is an infinite supply of lies, there’ an infinite supply of simple lies. Where’s there’s finite supply of truths, if you need a network that allows for complex ideas to spread for the truth to win out. Is that something you’re optimistic might happen with this form of gatekeeping?

PG: I mean Walter Lippmann, he wrote in the 20s and 30s. And he wrote three huge books about media. He wrote Liberty in the News, he wrote Public Opinion and he wrote the Public Phantom. And he was struggling exactly with the same problems and I’m talking about the 30s. And he was saying one of the ideas he has in Public Opinion he says, okay journalists should work with academics and especially with think tanks, followed by the government for providing statistics, very good surveys, inquiries. So I think the problem we have now with the Internet is the same. We have to imagine those kind of institutions and I think it’s not an accident if the Swiss television is talking about coming here to the EPFL and having their bowling’s on the campus, I think it’s a way of I mean, perpetrating that idea that comes from Lippmann and I mean, the strong idea from Lippmann, he was a columnist, he was a very good journalist and a philosopher. His idea was we need institutions to deal with the technology and to deal with the shape of our society. So yes, I’m optimistic but we have to do that, we have to work a lot.

BR: And just come back to this point about, so you still think that these are institutions, still don’t think that teaching philosophy to people to make better decisions about what they read, could be the answer?

JL: Why are the two distinct?

BR: Maybe they aren’t?

JL: I don’t think so. I mean institutions change because of popular pressure, public demand; people don’t sort of impose their will. And if they’re not even aware that there’s an issue, that there’s a problem, there’s no something that could be better and I think our class is a good example. We don’t sort of propose solutions to the problems. We don’t go and tell the students you should do this or you should do that. We basically sort of talk about the problems and the complexity of the problems and the historical precedents and historical basis for the problems and let the students make up their own mind as to what they think should happen in the future. The final project on the course is a presentation on a topic of their interest and part of it is they’re supposed to propose solutions to the problems and they think that it’s interesting because we don’t really do that in our presentation but we asked them to do it in theirs, and I think it brings out how hard it is to do it because their solutions in general are not very good and they’re sort of very partial. They don’t really cover the complexity of it. I think struggling to come up with these solutions is very good because it makes you realize that these problems don’t necessarily have simple solutions. But yeah, I mean going back to your question I think that you should, we have to do both.

DG: Do you think I mean, we’ve talked a lot about privacy in social networks and things like that but in terms of AI and machines talking to people? Is that worse in the sense that we’ve now got black box mechanisms where we don’t know, we have potentially biases that are encoded? Do we need to look at the institutions required for the provenance of the training of these things or?

JL: It’s a really interesting questions obviously a very hot topic in the whole field of machine learning these days. The answer is everybody who’s doing machine learning seriously and sort of applying it to any kind of problems should be concerned about the bias in the data and the bias in the training because otherwise, the answers you get are not necessarily the correct answers. So yeah, you know even separated from the sort of ethical, which I think comes about a lot because of the application domain. But even if you were solving sort of a purely technical problem with machine learning, you should be very concerned about the biases. You know, there are lots of examples were sort of, you train on a particular set and what you end up with is a machine learning system that works for people that are very similar to that set. You know, white people for face recognition because that’s all you had in your set, people who speak English like I speak English because you didn’t train with people with accents and you end up with this tool that basically doesn’t really work right. And so again that’s another form of bias but bias is also used for sort of much more serious consequences of using tools for judicial purposes and you basically are based that on sort of data that’s collected, that reflects people’s biases and sort of racial prejudices is a big one in the US. But you know again, it’s a problem that you’ve haven’t built a very good system, if you’ve sort of not accounted for this and not done your best to correct for it and not easy to even know what happens and it’s not easy to correct for it.

BR: And would you advocate for data science being a regulated industry?

JL: No, how can you regulate?

BR: Because there are a lot of responsibilities that come with data science so you’ve talked about bios but what about using people’s data against them? You know what about if on one, a bit, you’d be like were discussing earlier you know, on one level is great that we can pull data because we can diagnose illness faster. But what happens if we discover somebody has a preexisting condition and we want to charge them more for their insurance you know like the responsibilities around data science?

JL: Yeah but that’s not a data science problem, that’s an insurance problem. And I think the problem with regulating data science is that data science is just a technique that is going to be used by pretty much everybody in every single facet of industry or is already in the process of being used. You know, we have a Masters in data science, very popular masters for students because it’s an important area. But the sort of consequences of it depends on what the data is doing and what you’re producing from it and then what decisions you are making from it. So the one about health is a very good one in fact, I use it in the class because you can provide two perspectives, one of which is that the Insurance obviously want more accurate information about the people they’re ensuring because then there’s less uncertainty and they can price their product more fairly. But the flip side of it is that if you’re one of these people with a congenital problem then pricing it more fairly means that it would be much more expensive to you to have insurance. And the other idea of insurance is to distribute the risk among a larger group of people so that no one individually is forced to bear the entire cost of something they can afford.

BR: In keeping with that conversation, that’s a follow, some sort of middle ground is the way to go right because premiums that apply to everybody mean that we’re not pricing the risk correctly but allowing people to use data against their interest is not so it would seem like there’s some sort of halfway house.

JL: There is definitely a huge tension there and I mean EPFL and Switzerland in general has a lot of research going on and personalized health and that’s actually one of the sections of Global Issues because there are a lot of issues related to that and the professor in in SB science TV, who is leading the personalized Health Initiative. So it’s very clear that sort of believes that Switzerland needs to change the laws related to insurance before personalized health becomes widely deployed, that you just cannot allow this information to be out there in the hands of insurance companies applied to individuals as opposed to applied to groups of individuals for pricing.

DG: And a lot of the issues that we talked about information technology can be solved through openness and meritocracy, things like that. Isn’t the elephant in the room that the life sciences issues, eliminate all of that possibility in the sense that people actually buy advantage, permanent advantage is that is there…

JL: And what do you think?

DG: Genetic engineering, being able to choose that your children are going to live longer and be healthier and things like that through money.

JL: Rich people’s children have always lived longer and been healthier.

DG: Yeah, exactly. So this does seem to exacerbate that potentially, is that.

JL: It potentially could right, I mean I’m not sure the technology is really there yet to make an appreciable difference, except for sort of gross effect factors like sex selection, which is clearly a huge factor in certain countries.

BR: I wanted to ask you Jim, about this idea of because you mentioned earlier you said when the Internet was conceived, it was an incomplete conception because it didn’t have any notion of identity for example, there was no notion of a native currency for paying for things on the Internet. You in your course you talk about Bitcoin, in the in your course you talk a lot about technologies for establishing identity on the internet. Do you think these things can be almost sort of reversed engineered back into the internet? Are you confident that technology beholds? The answer to some of the problems we’ve been talking about?

JL: Yeah, I think they will eventually be added to the Internet. The problem of course is that because the Internet is everywhere and it’s embedded in so many different systems and has so many different players building it, running it, making any kind of change is extraordinarily difficult. Change occurs slowly, much slower and it would have been much easier to sort of fix these problems in advance but you know as I said, people didn’t perceive them as real problems. And in fact you know the technology of the, particular cryptography which is the foundation of Bitcoin didn’t really exist when the Internet was invented; this notion of public key cryptography was from the late 70s. Before that there was no way to do cryptography across the network. It was a brilliant idea. Brilliant observation but it took many years to develop it to where we are now. So yeah, I mean it’s great to sort of imagine what we could have done if we knew what we have now but this, that’s not the way things work.

BR: So just to make it more practical for… So here at the EPFL, you’ve got this Center for Digital Trust? What kind of technologies is that employ? What kind of stakeholders is it bringing together, how well is that working?

JL: Great question! So the answer is that we have a lot of people here who do cryptography and a lot of people who are very interested in sort of applications of blockchain which is the technology that underlies Bitcoin. It’s not necessarily related to financial information. So we’re interested in how you can share information more reliably and securely across the internet so… give you an example, one of the companies that participates in C4GT, and I won’t name it but they basically collaborate with a number of other companies who have information that would be very beneficial to share among this group of companies. But it belongs to the company. So they don’t want to sort of just put it there. They don’t want to give it to the central company because they want to maintain ownership. But all of the parties would benefit. I mean just basically, the tragedy of commons again, is that if they all act individually; they sort of are worse off. If they act together, they would all be better off and so the technology that we’ve been developing is basically.. allows the information to be shared but controlled at the same time.

So you know, you can maintain strict controls over who can see it, what they can do with it and you can maintain your ownership of the information but at the same time you can share it with a group of people, group of companies that are sharing with you. And so this is the type of technology we’re building. There are a number of companies that are interested in it. Even international organizations like the International Committee of the Red Cross is got a lot of information sharing problems, because a lot of their operations are done in very hostile parts of the world. Whereas there a lot of parties that would like to know about the people that they’re protecting as well and so they need to be able to preserve information.

And so technologies like blockchain have a lot of uses beyond just building things like Bitcoin and we’ve been exploring that and working with companies to take the core technology which was developed here and then applying that of applications of the companies come along with to demonstrate it and then another example is actually hospitals in Switzerland, all of which have information about their patients. And they protect it and they should protect that kind of information. But you know, if I’m a medical researcher doing research on a particular disease, it’d be great if I could access the medical records of the patients all across Switzerland in many different hospitals that have that disease, I could gain a larger set of samples with informed consent, patients could give me permission, but how do I get access to all this information? How did the hospitals at the same time maintain their responsibility to protect it? And to make sure that I’m not doing something with it that I shouldn’t, I’m not giving to insurance companies, I’m not selling and I’m not being a Cambridge Analytica as well? And this is the type of technology that we’re developing and trying to demonstrate by building applications and collaboration with other parties.

DG: And does this perform better than having a trusted third party being the gatekeeper of distributing that?

JL: Better in what sense? I mean there’s a lot of different dimensions right? You know having a trusted third party is usually more efficient. You know if you really do trust the third party just giving them all the information that they can aggregate it, they can then distribute it, they can control it. But coming up with a trusted third party is very difficult. And you know, you got to wonder about what their incentives are. The hospitals in Switzerland have never come up with a trusted third party for sharing medical information.

BR: Yeah. And in general, I’m sorry to be possibly pessimistic but and normally applications of blockchain like this are better when they are permissioned right? They are more efficient and I’m just wondering, are we trying to tackle it… because the issue, the interesting global thing, the ethics of AI is a global thing, the ethics of data science is a global phenomenon, all these problems that we’re facing happening in a time that’s… are we facing these problems at a time when the world is of fractionalizing and balkanizing? Does it make it harder to solve these problems?

JL: No. I mean, I don’t think so. I think the sort of research that goes on it’s become, if anything much more international over the last 10 years than it used to be, it used to be you know in this field the US dominated computer science research, you look in the conferences, the journals, who was basically US centric publication, it’s a lot more international than it ever was. Europe, China, various other parts of the world are participating and there’s still fortunately a culture of open publication. There are a lot of ideas, they move around very quickly. They’re the sort of people that are innovating in this area and then this whole sort of startup culture of this entrepreneurship, a lot of these ideas are now moving very fast from just being developed to actually having people to exploring them and trying to see whether there are commercial applications of them. Well they fail because a lot of the applications don’t work. They’re ahead of their timers. They don’t have the right market. This is great.

I think climate change is much more likely to wipe us out than killer robots. — JL

BR: So I just want to with the time we have remaining… just want to enter into kind of a final section where we indulge into some futurism and some future applications of these technologies and so I read on Wikipedia Jim, that you you worked on the Singularity Project at Microsoft. So let’s maybe let’s start there, how close to Singularity do you think?

JL: You got to remember that the only thing in common with that vision of singularity is the name. The Singularity Project was a project that Galen Hunt and I did at Microsoft…

BR: So it had nothing to do with the…

JL: No, to rewrite an operating system in a high level to try to attack the security problems by going and sort of redoing everything from fresh and It was a great project, we did a lot of impact of our follow on projects inside Microsoft, we had zero impact on the rest of the world…

BR: Okay, so having exposed that I didn’t do my research properly. I’m still going to get you on that question though. How far away, do you think the singularity is?

JL: Extraordinarily far away. I do not worry about machines replacing humans in general. There are two forms of artificial intelligence. There is sort of let’s call general intelligence, which is sort of a machine being as smart and as capable as a human being. And now this is what the Singularity worries about this, that there is going to build machines that become as smart as and even smarter than humans and decide that they don’t need us.

But if you look at what artificial intelligence machine learning is doing, they’re doing very individual human skills. They’re doing sight, they’re doing vision recognition, they’re doing hearing, they’re doing language recognition, they’re doing natural language processing but you put them all together and they’re not even as intelligent as a little year and a half year old baby. And they’re extraordinarily difficult to build; you show a baby, sort of a stuffed dog. And then they see a real dog and they realized that they’re the same thing, believe me, no machine learning system would ever make that kind of influence and we’re talking about sort of a little infant, we are a long way from general intelligence.

BR: What about killer robots?

JL: Killer robots? Yeah. With these technologies that we have you can use them to build sort of very lethal weapons whether they’re robots, they will be, you can call them robots and it’s sort of militarily attractive to do that kind of thing. And I think there’s a lot of pressure in a lot of countries to go down that path and I think it’s really quite dangerous.

BR: And how do you stop them?

JL: How do we stop any weapon? It has to be a treaty between countries where we agree that this is sort of beyond sort of reasonable and that the country’s all sort of step back and say we don’t want this to happen. We agree that this is an area of weaponry, we will not explore. You know, that’s happening for certain types of weapons. Not all, unfortunately.

BR: Philippe you’ve been quiet, what worries you about the application of these technologies, what’s the most worrying use case that it keeps you awake at night?

PG: Well one of the things that we kind of lost some of our access to pluralism, we’re not exposed to such a diverse community with you knows different kinds of opinions. Today with the social networks we’re well in this slope. That’s one of the main losses. The other one is that maybe we lost some of the institutions and I’ve been using a lot this word and I’d like to say that an institution for me is kind of a socially enforced collective habit. So what are the collective habits we are choosing for our society? And you were asking before about philosophy, do we have to teach philosophy or social sciences or build institutions and I think they’re quite different and they go together and what we try to do in the course with Jim is we try to have a space for reflecting on the implications of the technology and I come with my words, Jim comes with his and also we come with our passions and we have this space to reflect on those issues but at the same time it’s a training, it’s a kind of institution as well because we’re training those engineers, those very gifted students to think in another way about the work and its impact on society.

BR: Jim, do you worry about runaway functions, if I’ve even got that terminology correct? This idea that you optimize something for something but it just over time it does it at the expense of all else so you optimize the machine to make paperclips and then eventually sort of kills all human beings because it wants to make as many paper clips.

JL: No. I don’t think that that’s the way we’re, the minds of the human species is going to come. I think climate change is much more likely to wipe us out than killer robots. Sad, but I think climate change is very real. Yep.

BR: Okay, so I think we’re almost out of time. Dave, do you have another question any on the topic of futurism?

DG: : Just to close off on the climate change. Do you think for example, with fires in the Amazon that it would be morally justified to let’s say infect the cattle with BSE to stop the sales of beef and put out the fires? I mean, what level can we combat issues?

JL: I think that’s probably the wrong approach.(laughter) You know that if anything, history shows that just sort of aggressive action leads to aggressive counter actions and that you would be better off trying to do what’s been done, which quite effectively, I think is to shame Amazon and inform the world that this problem and yeah that is again, one of the sort of advantages of the Internet is that the sort of availability of technology, cheap satellites, made this imaging possible and then the Internet has made it possible to disseminate the information about these fires in essentially real time throughout the world, get the celebrities to jump in and play a lot of pressure in a very short amount of time to Brazil. Wether it changes, I hope so but you know, technology has good sides as well as bad sites too can be applied to fix problems.

BR: I think that’s quite an optimistic note to finish on but I think we can do better. So I’m going to ask each of you to conclude with an optimistic comment about how we can take the really positive foundations of the internet and use them for good.

JL: So one of the things I do in this course is I finish on a positive note. The last lecture we do is about the whole open, whatever movement, the open source movement for software, the open Science movement, open data movement, talking about how the fact that you eliminate the cost of disseminating information has opened up all of these different ways of doing software development, doing government, doing science, that was just would have been just totally impossible when even when my career got started. And to me now, this is an exciting thing is that you know the other bad things that happened and there are a lot of, sort of people make money off of but they’re also sort of a lot of really good people out there, taking advantage of the internet to do things that they really get benefit out of… and that the rest of us benefit out of and if you look on Facebook or Instagram or something for people that are sort of really benefiting from it. I mean, there’s another statistic which I don’t know if you’ve seen is that you know Facebook makes $200- 300 per person in the United States but the value that people get out of it is much higher. There was a very nice study where they basically paid people to stop using Facebook for a year and they had an auction and the price for people in the United States was about $2,000 to not use it for a year. And so, the benefit is 10 times what the revenue is to Facebook so that you know the difference is really what you know Facebook is giving you are giving you $1900 to $1800 dollars’ worth of benefit.

BR: I think that is a podcast in itself, this idea of how much consumer surplus the internet creates which isn’t captured in any of our existing measures of the economy but I think that is we’ll get you on the podcast for that discussion. We just really have to leave; we want to get Philippe’s optimistic conclusion to the podcast.

PG: Well I’ve talked about Walter Lippmann and he was very optimistic in the fact that democracies thrive with very good knowledge, very good media, and a link between the two. And I think the Internet has given us this opportunity. When I see Wikipedia, that’s it’s quite fantastic actually. And I would say, I would say that the way internet enables collective activity and research. That’s the thing that really makes me optimistic. Thank you.

BR: Amen to that. Okay so I think all that remains is to thank you very much for your time. Thank you, Philippe, thank you Jim, thank you David, for taking part in this podcast.

All: Thank you.