Pod Insights a16z Podcast a16z Podcast | Good Bubbles, Bad Bubbles -- and Where Unicorns Come from About this episode Venture capital and investing in startups is the (modern) classic case of decision making under uncertainty. As new players and sources of capital enter the market, however, how do we hedge against that uncertainty? For one thing, startups have more reasons than ever to focus on "the 2 Cs": cash (flow), and control.
Or so argues Bill Janeway, an early venture capitalist and partner at Warburg Pincus, visiting lecturer in economics at Cambridge, and author of one of the definitive books on the history and evolution of the VC industry. In this segment of the a16z Podcast, Janeway -- who can best be described as a “theorist practitioner of financial economics” -- offers his insights on where unicorns come from (hint: it has to do with IT "disappearing"); how big companies need "absorptive capacity" when acquiring new tech; and why bubbles are "banal" -- including the difference between "good" bubbles and "bad" bubbles.
Transcript 0:00 Speaker 1 Hi, everyone, welcome to the A6CNC podcast. I'm Sonal, and I'm here today with Michael, and we are talking to Bill Janeway. Bill is a senior advisor at Warburg Pincus, where he had headed up its high-tech investment efforts starting in the late 80s. So he was actually one of the early venture capitalists. And he is also a visiting lecturer today in economics at Cambridge, where he also received his doctorate in economics. So the best way to probably describe him is as a theorist practitioner of financial economics. A couple years ago, Bill wrote the definitive and award-winning book on doing capitalism in the innovation economy. And he has a really unique perspective having tracked the history and evolution of venture capital. So we thought it'd be great to have him on the A6 and Z podcast to share his perspective on the industry today.
0:45 Speaker 1 Welcome, Bill.
0:46 Speaker 2 Thank you very much. So we want to tap into your brain if we can today and talk to you about what's top of mind. And I know you've been you write a lot about innovation, you talk about it. You've had some tweet storm battles with our own Mark Andreessen around the subject of unicorns and and how to think about it. So maybe we can jump in there. Trevor Burrus, Jr.
1:08 Speaker 0 Well, sure, we can start there. The my tweet storm conversations with Mark have actually begun with what I think was a misunderstanding, but I think clarifying this particular issue helps understand where the unicorns come from, though not necessarily where they're going. So my thesis is that, like electricity in the 1920s, from the point of view of the user, not the engineer, not the guys behind the scenes or under the table. But from the point of view of the user, IT is in the process of disappearing. Now, that comes in the context of open source software, web services from the cloud, which means that the cost in terms of time and money of launching a new service has declined asymptotically to zero.
2:08 Speaker 0 Right. Not building it out into a full business, but launching it. Back in the 20s, when you could plug into the wall into a standard socket, uh, instead of worrying about cycles and impedance, all of a sudden you had an explosion in applications. They were called household appliances from toasters to washing machines to dishwashers. Today we're having the same kind of thing in the web services economy. Now what also goes with IT disappearing, uh, as I say, not under the hood. It's really hard to run cloud services. It's really hard to provision mobile applications that capture real transactions.
2:58 Speaker 0 But from the point of view of the user, it is as if you no longer worried about what was going on behind the scenes. And what is distinctive about the web economy is the extraordinarily low friction in bringing new services out and people signing on to those new services and starting to use them. This means, on the one hand, a step function increase in the number of startups, in the number of Darwinian hopeful monsters. Right. Most of which will fail, most of which will fail pretty fast and without costing an enormous amount of money, so that at the front end of funding these, the spray and prey style of investing is rational and practical.
3:53 Speaker 0 When one or more of these services actually, quote, gets traction, we're in an environment which is not going to stay like this forever. We're in an environment of extraordinarily low interest rates, of enormous quantities of liquidity, looking for return. There is the phenomenon you guys out here know a lot more about than I still have to explain it back east, let alone in Cambridge, England, the phenomenon of FOMO, fear of missing out on the next Facebook, the next Twitter, whatever it may be. Although the next Twitter may be the last Twitter.
4:39 Speaker 0 I don't have an opinion about that. So we're now seeing, and this is where Mark and I are in complete agreement, and we're in agreement with people like Bill Gurley that the provision of large amounts of capital at premium valuations to private companies by investors who historically and culturally are public market investors, are enabling both these private companies to stay private much longer, and also to fund losses to have no concern for what at some point in the not terribly distant future is going to become a very, very different environment.
5:26 Speaker 0 I grew up in the world where my original mentor in how to understand a business used to say corporate happiness is positive cash flow. Positive cash flow from operations means that you, the entrepreneur, the venture capitalist, own options. You own the options about whether to sell or not sell, whether to raise more money or not raise more money. You're not a prisoner of the capital markets. Now, when you see investors, hedge funds, mutual funds paying premium prices to buy illiquidity, to buy securities with no liquidity and with no governance rights.
6:13 Speaker 0 Trevor Burrus, Jr.
6:15 Speaker 2 In the private market.
6:16 Speaker 0 In the private markets. The only thing you know is that when it ends, it will end painfully. Uh-huh. Now there's a last piece of this, which I don't, my own view is that the common opinion is misses a critical point. And that is, well, what about the IPO market? Now it's one thing to say, right, if I have Wellington and T Row and Fidelity and the Russians throwing hundreds of million dollars at me at multi-billion dollar valuations. Why would I go public? The answer is of course you wouldn't.
6:46 Speaker 2 Right. Money's available and it's and it's cheap, right?
6:48 Speaker 0 Trevor Burrus, Jr. It's very cheap. But on the other side, if I wanted to go public on the way up, it is much, much, much harder to do that. This is one of my that now I'm wearing my uh my academic, my theorist hat, if you like. There's been a huge amount of work that's been done on the returns to venture capital, going back to when there began to be a venture capital industry, which is roughly 1980. Um it's always been very closely correlated with the state of the public markets and particularly with access to the public markets. From 1980, this this is my own data, data published in peer-reviewed journals, from 1980 to 2005.
7:33 Speaker 0 The average number of venture-backed IPOs per quarter was about 30, with a standard deviation of 10. So you were capturing most of the IPOs that were 20 to 40 per quarter. When it was more than 40, you were in a hot market. Right. When it was more than 40, and most of the companies weren't yet profitable, then you were in the bubble. You were the dot com telecom internet bubble of the late 90s. Since then, we've had all of about four quarters in 15 years when there have been as many as 20 venture capital-backed IPOs. That's at the bottom end of the normal range. Most of them have been below more than a standard deviation from the average over the previous generation.
8:21 Speaker 0 And increasingly, the IPOs that have taken place have taken place for guess what? Biotech and life science companies. Because now you're speculating, not just on as you used to about biotech, will the molecule get through clinicals, you're speculating on will big pharma, whose productivity of research has collapsed, buy them before we have to find out whether the molecule gets through phase three clinicals. So the number of venture-backed IT companies, which historically was two-thirds of the total from quarter to quarter to quarter, has plummeted. And the total number has plummeted.
9:04 Speaker 1 So what's going on there?
9:06 Speaker 0 One clue is that the mean and median value of an IPO has more than tripled since the dot com in real terms.
9:17 Speaker 2 Right.
9:17 Speaker 0 You have to do a deal that's over 100, maybe 150 million. I'm not talking about Facebook. But you have those the mean and median are way up from where they were. Well, why is that? Trevor Burrus, Jr.
9:30 Speaker 1 Well, it's actually important that you're pointing out the median as well, because obviously it's not accounting for the outliers like the Facebook. No,
9:36 Speaker 0 absolutely right. It's really the medium that I look at. The institutional change that most people who think about the IPO market are not taking into account is the incredible consolidation of the investment banking business. What that means is that there is a hole in the capital raising process. And To get the attention of the global major dealer banks, you've got to be offering them a transaction on which they can make at least five million dollars. And you start doing the arithmetic on the economics of an IPO, and you're up in the nine figures to do an IPO. Now, why does that matter? Well, it matters because the value of being able to go public is liquidity for your early investors.
10:29 Speaker 0 There is no more important world in the world of finance, word in the world of finance than liquidity. Because again, it's the flip side of positive cash flow from operations. It's your control over your situation. You change your mind. And you it's there's a price that you pay, but the price gets you back to cash on your own. So we come all the way back around to the unicorns, 85 or so companies, private companies with valuations of above a billion dollars under no pressure to generate positive cash flow from operations because they're being funded at very, very, very low cost of capital by people who, after they've made the investment, they're there.
11:18 Speaker 0 They are stuck, and they have no control, no control whatsoever over what to do if something goes wrong. Trevor Burrus, Jr.
11:28 Speaker 2 So how does this then play out and how does that also relate to venture capital returns?
11:33 Speaker 0 The first way I think this plays out is that the relatively narrow band of companies that are creating disruptive web-based services are being able to explore the new economic space that's been created in a much more rapid way than they otherwise would. That's a good thing. Right. That's a good thing. But as you do see IPOs for these companies at prices materially below the last private rounds, that's going to feed back upstream. It's one thing to say, well, I've got some anti-dilution protection. If it goes out, I get a few more shares. But you're still trying to explain to your boss, to your public market investors, to your limited partners.
12:24 Speaker 0 Just tell me again why you did that deal. Tell me again why you paid that price for the privilege of owning a high-risk business which has demonstrated no ability to make money, and which is dependent on there being more people like you to continue to feed at capital without discipline and without control.
12:48 Speaker 2 I mean, is it is that without discipline and without control or and or is it mutual funds and other of these players coming into the private market looking for growth? Trevor Burrus, Jr.
13:00 Speaker 0 Well, they're looking, they're looking, they're f they're looking for super growth, and they're taking to acquire access to that growth. They're making, they're taking investment positions, which are irreversible even if the price could be modestly adjusted after the fact. Right. It's a category error. When one or more of these companies actually fails, when it finds that there is no way to get deposit cash flow, when it finds that it is dependent perpetually on influx of cheap capital, and they fail, then the message will come back. And by the way, as long as this capital is available, there will be more and more companies coming to buy that capital.
13:55 Speaker 2 Right.
13:55 Speaker 0 Again, the people on the on the other side of the market aren't morons. So the quality inevitably will decline. Inevitably.
14:04 Speaker 1 So how do you actually get into evaluating that quality, though? Because one of the conundrums of VC, obviously, is that we're in early, but it's it's you have to figure out ways of assessing if a business is going to be successful, but yet at the same time, you really don't know because of the type of innovation it is. Trevor Burrus, Jr.
14:20 Speaker 0 Right. Well, clearly the classic place where you're making decisions under conditions of radical uncertainty is when you're funding a startup. Right. So you think about how do you evaluate a startup? Well, the first way I actually begin is by looking at the biographies, of course. I have a I have a heavy propensity for people who have survived failed startups, near-death experiences of big companies, and along the way have demonstrated some degree of creative discipline in and had some experience in in helping to build a real business. And by a real business, I mean a business that is self-sustaining from positive cash flow from operations.
15:07 Speaker 0 I'll keep coming back to that. But the second thing I'm really interested in is how do they read the market, particularly today when we're talking about SaaS, IT-enabled businesses delivering value, whether to consumers or to enterprise customers, how good are they? How compelling and analytically clear is their evaluation of the market they're trying to address. It doesn't mean that that market they're going to wind up addressing, but it tells you an awful lot about how they think.
15:58 Speaker 1 That's correct. We don't know even how to think about SAS. When it's Mark Benioff came up with the concept of Salesforce, people were just like, what the hell is that? Trevor Burrus, Jr.
16:07 Speaker 0 We have to be very careful about survivors' bias.
16:10 Speaker 1 Yes, definitely.
16:11 Speaker 0 As I said, one of the really interesting, maybe the most important phenomenon of the startup universe is the combination of open source software and cloud computing services. It used to be that whether you were addressing consumer or business markets, it took 20 million dollars to produce whether it was something that went out and retail through a disk or whether it went out on a tape in a truck to a big company. Because you have to buy the hardware. You have to buy the software licenses. I tell my the kids in my class, you know, to do a startup guys, you don't even need your parents' credit card. You can do it on your own. That means so many more startups.
16:58 Speaker 0 So we really have to be careful about survivors' bias. And as I say, I think if if what you want to do is be investing in the next Facebook, then you got to be kissing an awful lot of frogs and hope that none of them are going to leave you with warts because most of them are going to die. And it's a way of investing. It's not the way I temperamentally, you need a particular kind of temperament for that kind of investing. It's not temperament that I share. So I can't go much further down the road with you on those. I have myself historically very much been an investor at Warburg Pinkers and before into business into new companies delivering innovative functionality to business customers increasingly where it's an intersection of functional capability and information that previously could not be aggregated, analyzed, and used to drive decisions.
17:59 Speaker 0 Trevor Burrus
18:01 Speaker 2 Given what you describe, given the money that that is sort of this constant flow of money that's needed to fund some of these, as you say, unicorns. But also given the fact that there are so many companies that are started that some of them will, you know, follow through and become what they, you know, everybody wants them to become.
18:30 Speaker 2 Trevor Burrus,
18:30 Speaker 0 I have no doubt that sooner or later that money will disappear. Right. No doubt.
18:35 Speaker 2 And what happens then? That's my question.
18:37 Speaker 0 No, here's the the real point I want to make here. Um, in an environment where the IPO market is very hard to reach, and you have to build a business under conventional valuation metrics of more than 100 million dollars to be able to go public, to have the $500 million valuation that lets you do the $120, $140 million IPO, which will get Goldman or JPM interested under those conditions, from the entrepreneur's point of view and from the venture financier's point of view, the rational strategy is to keep asking the question at each stage at each round of investment, do we sell now? To think about what is an honorable and economically useful function for venture capital startups, which is funding research and development for big companies.
19:33 Speaker 0 Now that shifts the terms for the innovation economy at large, and this is kind of what's evolved. I I wrote about this a little bit in my book, doing capitalism. Uh I would I would emphasize it more today that the rational approach is to be thinking from the beginning, not the default option is we build this business to an IPO. The default is that we see Sell after we've A plugged it in and it lit up. We proved that it works. B, I'm doing this in terms of venture rounds. B, we have two or three customers who will stand up and say, I paid real money for this stuff. I'm talking about enterprise, not retail.
20:18 Speaker 0 I paid real money for this stuff, it works, and I'll buy some more. Do we sell now? Or do we then take on the much bigger risk of actually trying to build that self-sustaining, cash positive business, which over time has the potential to generate 10 times the return, but you're taking on orders of magnitude more risk.
20:43 Speaker 2 Right.
20:43 Speaker 0 Now, coming back to uncertainty, the central theme of the whole first half of my book is that there's only one way to hedge against uncertainty in venture startups. You don't have credit default swaps. You know, you you don't have instruments that you can buy to hedge your position. It's two words cash and control. Enough cash, unequivocal access to enough cash so that when something goes wrong, you can buy the time to find out what it is and to assess what you can do about it, and then enough control, not necessarily 51% ownership or a majority of the board, enough strategic influence so that you can mobilize the team that can shift gears and restart so that you can actually have the space in which to sell the business sooner than you hope to sell it, in which, and this of course is something that venture capitalists do.
21:50 Speaker 0 You can change management and have another go at the original business plan because it was a failure of execution. But cash and control is the only, it's the retrospective hedge against the inevitable uncertainty of playing this game in the early stages. You can diversify across the portfolio, but you know there's a there's a a non there's a systematic risk across all of our early stage portfolios. And the risk is it's really execution. You can't diversify away that risk, but you can hedge against it by holding back more cash by being very careful about what the governance structure is.
22:38 Speaker 0 And you know, uh Warburg Pink is when we were building the businesses we built in the 1990s, like Veritas and BEA, we were in a position where we could have sufficient ownership so that we could help support and guide, but also respond in real time to the inevitable uncertainties of this way of life.
23:03 Speaker 2 So does what you describe and do you expect, and this is something that um our managing partner uh Scott Cooper has written about. Do you expect to see more MA sooner rather than later when you talk about at each stage you need to sort of step back and think about do I sell?
23:16 Speaker 0 Right. And if you look at the data, look at the NVCA data, the um the proportion and the and and the value and the value per transaction of MA has kind of gone up as the number of IPOs for venture backed companies has gone down and stayed down. So that's the that's the flip side. If you and it does create another issue, and this is an issue which I think is is systemically important for our economy. The constraint on successfully bringing this kind of innovation to the marketplace where it will make a real difference to the economy. The constraint becomes the absorptive capacity of the acquiring companies. It's really easy, and we have so much experience of the big company, the big stagnant company buying the hot startup and burying it, smothering it.
24:06 Speaker 0 I'm very I have to say I'm extraordinarily impressed by Google's style of being able to acquire and support and create space. Now, if you're as rich as Google is, you can afford to do that. But it's a cultural phenomenon. Right. And we have many other examples, including a number right out here of big companies that have proven to be terrible acquirers. Learning how to acquire, protect, honor, support, and then build out innovation from the outside is a huge challenge, and it's one that's really important for the whole economy, not just for our little segment of it. Trevor Burrus, Jr.
24:44 Speaker 1 So, Bill, one question in your notion about the role that VC plays in RD. I mean, basically you're saying that VC is distributed RD. And I'm curious to hear some of your takeaways based on that
24:54 Speaker 0 view. Aaron Powell Well, what I'm saying is that because getting public is so difficult and so rare for venture-backed tech companies that practically speaking, we should be thinking about that what we're doing is funding distributed research and development for big companies. And the big companies typically have lost much of the capacity for innovation, and they have to learn how to acquire innovation from the outside and not kill it. In biotech, that's kind of the model because big pharma's productivity of research has declined. There have been decades now through which they've been acquiring early stage biotech companies and they're doing it earlier and earlier.
25:40 Speaker 0 And it's relatively easy for them because they're just buying a molecule, and they've got all of the uh machinery from toxicology to sales and marketing by way of compliance and regulatory stuff and clinical trials management. It's much harder in the world of IT for big companies to be able to acquire and not kill. But with respect to what the BCs are doing, the danger, the danger can be going too far upstream.
26:08 Speaker 1 What do you mean by that?
26:09 Speaker 0 I mean funding science rather than funding funding the commercialization of technology and finding commercially valuable new innovative applications of technology. Back in the mid-auties, um, some billions of dollars were wasted in venture guys investing in nanotechnology, venture guys investing in science for ultimately clean tech opportunities, but that were way upstream that took much too long, much too much money, and had a kind of scientific risk, not just technology risk. This is where there really is a role for big companies, and if they're not going to play it, nobody is.
26:59 Speaker 1 So you're saying basically that the science that's um fundamental science, funding fundamental science, VC is doing that is a category error. Correct. Whereas commercial science that's maybe accelerated by
27:09 Speaker 0 Well, commercializing technology, science to technology to commercial application, where VC successfully has intersected that is somewhere between proof of concept in the lab and defining the applications that have immediate commercial, commercial opportunity for success. You know, we only have three to five years to demonstrate that we've done something smart when we've made a startup investment. Trevor Burrus,
27:40 Speaker 2 Jr. Yeah, and it's been, you know, to your point about nanotech, it's been whatever, 15 years, and we're starting to sort of see it creep out there, except that nobody really calls it nanotech anymore, it's just there.
27:52 Speaker 0 It begins to, yeah, it's beginning to emerge as features and functions, but building a business. You know, there's a lot of history here. Lasers, right. Okay. The first laser was it took 10 years to replicate the lasing phenomenon, but it took 20 years to discover what the hell the killer app was, you know. And who would have believed that the killer app was at the supermarket checkout counter, right? And then consumer electronics.
28:21 Speaker 2 Right. CD players. Right. Right. I thought it was laser rock shows, but apparently that was laser.
28:25 Speaker 1 Well, laser printers, actually laser printers and the semiconductors. That
28:28 Speaker 0 laser printer came along after the supermarket, but you're right. Um
28:32 Speaker 1 back on the theme of just, you know, this notion of experiments. So one of the things that's interesting about your background as an economist is that you're sort of an outlier in that industry because you're talking about trial and error as a philosophy of the other. Well, that's
28:45 Speaker 0 absolutely right. That's absolutely right. And and you know, I did my doctorate in economics at Cambridge in 1970, I left academia. I left academia because frankly, I found it impossible to teach bright young kids that the way to think about how the economy worked was through a model that said resources are allocated efficiently to their highest return uses, because I'd already begun to appreciate what I then really learned as a working venture capitalist, that we advance by trial and error and error and error, that waste, necessary waste is built into the system, that virtue is not efficiency, efficiency is the enemy of innovation.
29:33 Speaker 0 Then post-2008, all of a sudden the recognition that maybe the markets aren't that efficient. You know, maybe if we've had a global financial crisis and a world class economic freeze, it's time to start thinking about markets in a different way. So it's just been a tremendously exciting and stimulus, stimulating and youth youth enhancing experience.
30:01 Speaker 2 Yeah, there's been plenty of stimulus lately, but that that too. That too. What what we
30:05 Speaker 0 have, and by the way, we've we've seen something like this before. Um what we have now is uh an environment in which exactly it's designed for experimentation. And that's why I say back in the 20s, that was the case with electricity. Back 50 years before that, after we'd built out the railroad networks, we had the opportunity, we I say metaphorically had the opportunity to discover the killer app for the railroad. It couldn't work until you had railroads everywhere. It was called retail mail order. It was called Montgomery Ward and Sears Roebuck. In 1970, every town in America had a shoemaker. By 1920, all the shoes in America were made in Brockton, Massachusetts.
30:51 Speaker 2 Right.
30:51 Speaker 0 When the railroads were built out, we could have the kind of experimentation that when the electricity grid was deployed, we could have now we're having the same thing in the digital world, and it's fascinating. Trevor Burrus, Jr.
31:04 Speaker 1 How does trial and error get financed?
31:05 Speaker 0 Trevor Burrus So when you've got so much uncertainty up front, you know that there are only two sources of capital that can invest without regard for the immediate economic value that they can define. One is the government, when it's got a mission, whether it's national security or the war on cancer, and the other is speculators, financial speculators. Now, wherever, wherever you have markets and assets, you're going to have bubbles. Bubbles are banal. The objective speculation, everything from tulip bulbs to beach houses in the Nevada desert by way of gold mines and silver mines and commodities, occasionally, every once in a while, the object of speculation is one of those technological innovations, which, if it's deployed at large scale, changes the world and creates a new economy.
32:00 Speaker 0 Canals and railroads, electricity, aviation, wireless, aka radio, and yeah, the Internet. So the way to think about financial speculation is to recognize that bubbles are banal, but every once in a while, they're not only necessary, they're productive, because they bring together the magnitude of capital you need to build out the new network infrastructure and to finance the exploration of the new economic space that's being created. So you think about the difference between 2000, when the bubble was limited to the public capital markets. So when it burst, the damage wasn't that great. Yeah, the people who are invested lost their money, but they weren't leveraged, and it didn't infect the whole system.
32:48 Speaker 0 When the bubble infects the credit system, the banking system, and when the object of speculation are those beach houses in the Nevada Desert, when it bursts, it's catastrophic. It freezes the whole economy. So good bubbles, bad bubbles. Productive bubbles, ugly bubbles. Productive bubbles, if it's only in the liquid markets and the stock market, the junk bond market, and the object of speculation is some promising technology, let it run for a while. So
33:19 Speaker 2 to be clear, 2000, good bubble.
33:22 Speaker 0 Yep.
33:22 Speaker 2 2008. Bad bubble. Bad bubble.
33:27 Speaker 0 So the it in it was the bubble was located in the core of the credit system of the world, and the object of speculation was real estate that was never going to increase the production capabilities of the global economy or make those uh uh production capabilities less damaging.
33:47 Speaker 2 Well, so which brings us to 2015.
33:50 Speaker 0 Right. Micro bubble. Micro bubble in limited scope, focused, understand. Every bubble, every bubble begins with a plausible story. The plausible story today, as we've discussed, is zero cost of launching a new web-based service, zero almost zero friction in the explosion of that service if it hits a need, and consequently the potential for almost limitless growth in an incredibly short time. So
34:25 Speaker 1 if there's always a plausible story, why do you say this is a micro bubble?
34:29 Speaker 0 It's micro. This is micro because the scope of the bubble is much narrower. The scope of the bubble on the one hand, the scope is narrow, but what's weird and different is that it's actually taking place in the private market. Right. It's being funded differently. Now the good news about that is that when it bursts, the some of these companies that have learned nothing about business discipline will disappear, and the people who have invested in them will lose their money, maybe their jobs, that usually happens as well. But it's not going to implicate the entire economic system. It's not even going to implicate the remarkable dynamics of the economic system known as San Francisco Silicon Valley Bay Area.
35:14 Speaker 0 It will be looked back upon, and people will say, on the one hand, what were we thinking? And on the other hand, say, wait a second, out of those 85 unicorns, there are five great companies. Right, right. And by the way, that's the necessary waste of trial and error. Well, it
35:31 Speaker 2 sounds like there is pain ahead and yet there is much progress ahead, too. So Bill Janeway, thanks so much for joining us.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated August 25, 2026 · transcript and summary are machine-generated