Keith Teare on THAT WAS THE WEEK in tech

Speaker

I've got sunshine on a cloudy day When it's cold outside, I've got the month of

Speaker

August the 18th, the Friday, 2023. If it's Friday, I'm keen on it, must be, I've got sunshine on a cloudy day, must be, that was the week. My weekly show with Keith Teer, CEO of SignalRank about the state of tech in general and often the state of venture. The last two couple of shows haven't been exactly cheerful about the future of tech. Two weeks ago we had the bonfire of the unicorns and last week dry powder, which suggests that there's no logical reason why venture capitalists should be keeping their powder dry. They may just never invest again in the market. It's been a miserable time as an investor, but this week Keith comes out firing as an optimist. Venture is good is the title of his editorial, a convincing riposte to people who believe, perhaps like Kyle Harrison, whose article is in his newsletter, that the market and the venture market is a kind of Bernie Madoff style scam. Keith, the greater full theory you suggest is wrong when it comes to analysing venture capital. Yes, I think it's a fairly cynical theory. The mid-journey picture on the front of it is Bernie Madoff. It isn't actually Bernie Madoff. It's mid-journey's take on Bernie Madoff. I thought it was familiar. I couldn't figure it out. It's a sort of cross between Bernie Madoff and some high-powered Silicon Valley venture capitalist. Yes, and the contrast with the title is meant to be slightly ominous. Venture is good, but if he's saying it, then there needs to be a question mark. I picked up on this. It was in two different places. I don't think it's yet a meme, but once something's in two places, we can probably predict that by next week... Keith, once you're talking about it, it's a meme. You are Mr. Meme, aren't you? Isn't that your middle name? Keith Meme Tier? And the meme is... William, actually. The meme is that venture capital only works if some foolish investor pays more for the same company that you just invested in. This is Carl Harrison's point. I get the sense from you is you weren't entirely sure where Harrison was. Yes, I read him a lot and I don't think he's particularly cynical, so I was a bit surprised he means it literally. But I do think he's playing around with the idea that maybe there's an element of that. And I think what he's doing is overreacting to 2021. And in 2021, as always happens in a pylon, I think of it as a pylon, P-I-L-E money exceeds the demand for it and prices go up. And lots of investors come into the market that were not previously there. And suddenly the gross amounts invested dwarf prior years. That happened in 2021. It was at least double any prior year in terms of amount and velocity. You historicize it, Keith, in the editorial. You say that it all began... You're not blaming it on Yuri Milner, but you're historicizing the period between 2012 and 2021 in terms of Milner's investments, particularly in Facebook. Why is this significant? Well, because what we saw in 2021 was that was the end result of what is known as growth investing, warping the entire venture market. Growth investing is characterized by writing very large checks that value a company more than it really should be valued in order to get in. And that does correlate to the greater fool theory. That's why what Kyle is saying is, you know, there's a reason why he thinks it. Because normally those late stage growth investors are relying on an IPO. And the people who are the greater fools are the general public who buy at the IPO. As always. Although had they bought the IPO of Facebook, maybe not Twitter or earlier tech companies, they would have made money. Right. So I think there's a fair argument that some of the growth stage investors invested based on the greater fool theory. But to characterize venture capital as a whole as that is the mistake. So what happened in 2022 to change all this? Because it doesn't seem as if anything's fundamentally changed that there are always these speculative ventures. You invest, VCs invest, then late stage VCs invest, and then the public invest. And mostly they lose, but sometimes they win. Why is it different now? I think it's hard to unpack because there's a lot of variables. But I think you start with interest rate rises to combat inflation in the began during COVID due to supply chain issues in the world. So suddenly prices started to go up. Cars, for example, is a good example. You're often paying tens of thousands of dollars above the asking price for a car. Or mortgage rates, which change the real estate industry. Mortgage rates later. But this whole thing set in motion. It's the price of money that you and I have talked about. The price of money went up. And so public markets started to lose money to other asset classes. And at that time, the public market reaction was to lower the value of public companies to a much lower multiple of their revenues and earnings. And that set in motion a kind of a chain of events that ends up in venture capital where valuations have to go lower. And, you know, there's this kind of being washed up on the tide, washed up on the shore as the tide goes out. All those companies that raised money in 2021 and early 22 are high valuations. The tide's gone out and they can't get back in the sea because no one's going to pay the price. But again, that's happened before. It happened after the dotcom boom. It happened in the late 2000s. So I don't understand why it's so different. You're saying because of the more macro economic issues of inflation plus interest rates. Those are the key drivers. And COVID, of course, is a massive catalyst for that. Although COVID drove the market up and created enormous speculation. Well, actually, it drove prices up. Yes, you're right. And in some cases it drove up company values, but not entirely, by the way. Some companies lost value during COVID. So the other thing that happened is deglobalization. As deglobalization happens in the China-America conflict. Yeah, you got a piece. We'll talk about Biden and Chinese de-investment later. Right. So that also creates inefficiencies in the market. So it's all correlated. The idea that this is all to do with venture being a Bernie Madoff-like asset class is too superficial. Yeah. One of the things that occurs to me is how unusual this period was in the last 20 years, say between 2002 and 2021. It was an interesting piece. I don't know if you saw it in the Wall Street Journal this week, a couple of days ago on mortgage rates. And they had a chart of the mortgage rates over the last 50 years. And what's really astonishing is we think that 7 percent is high now, but it actually isn't that high. It's only high compared to the last 20 years. And if you go back to the 80s, the mortgage rate was at 18 percent. I remember that. Could it be argued, Keith, that we're returning to normal now? I don't think there's such a thing as normal. I think that's a historical. I think there's a set of unique variables in every moment that lead to outcomes and any attempt to read history backwards. And, you know, like there's a lot of people in venture that say, oh, it's like 2016. Well, that might be 2016. Well, roughly the number of funding rounds, the amount being deployed was equal to 2023. And that's roughly true in terms of just pure numbers. But there's nothing in common between 2016 and 2023 when it comes to the likely future. You know, so the likely future is... I take your point, but looking back, and this is not a show on real estate, who was buying houses in the late 70s, early 80s when the mortgage rate was 18 percent? Somebody must have been. You know, I bought one in London and I can tell you that you got... you had zero down, interest only mortgages. That meant that it was affordable. So the market adapted so that the consumer could still play. And then when I sold my house, which I did when I moved to the States in 1994. Sorry, 96. I only got as much for the house as I paid for it. Those glorified rentals. One of the companies that comes up in the Harrison piece, which we've talked about before, is Hopin. You mention it in your piece. What does Hopin tell us? Why does it still make you confident ultimately about venture? As I said, the title of the piece is Venture is Good. Not morally good, but economically good. You conclude your editorial suggesting that long live venture capital. And you still believe that it's a viable sector. What is Hopin? Why is it both a warning and a suggestion that things can still work out? Yeah, so I don't know if it's Hopin or Hopin. I call it Hopin, but you... I think it's Hopin. Hopin, Hopin. So obviously Hopin is opposed to Child for Failure because it just sold its assets for $15 million, having previously been valued at $7.5 billion. And Kyle's piece is triggered by that, as was one of my editorials last week. That's a barbecued unicorn, isn't it, Keith? It's a barbecued unicorn. So it's opposed to Child for Failure, but it's also opposed to Child for Success because within two years, Hopin raised a lot of money, produced a lot of virtual events, made a lot of revenue, and then failed to transition from the COVID to the post-COVID area to create a sustainable business. But the fact that investors were prepared to risk close to a billion dollars on its success tells you that the appetite for risk and reward is there. It won't last forever. If generation after generation loses their money, in the end, they'll go somewhere else. I did a show earlier this week on the value of art. Wealthy people will put their money into art or real estate or government bonds or non-tech stocks. I think all of the above is true and has always been true. So if you just focus in on venture and you look at the amount of money invested in venture, gross money, and the gross output from that money, venture is a massive value creator. A failure like Hopin is dramatic and newsworthy, but actually is a sideshow compared to the asset class as a whole. Now, it gets really tricky when you get down to single funds. Instead of looking at venture as a whole, you look at, let's say, I don't know, benchmark capital. Benchmark capital can win or lose in any given fund, and so can Sequoia. But the asset class as a whole is very valid. Now, a pension fund like Calpus, for example, invests in lots of funds and has diversified risk across the entire asset class. So it's pretty hard for it to lose money.

Speaker

Some people might say, Keith, you're the CEO of a startup, SignalRank, which is an interesting AI play on investment. You have to say this by definition. I mean, you have to be a believer because you're in the space. Well, there's facts, isn't there? I mean, I don't agree with the idea of alternative facts that Trump's assistant, I've forgotten her name now, the blonde-haired woman from back in the day, who's married to a Democrat. I know what you mean about alternative facts. Alternative facts. I don't believe in alternative facts. I mean, you can measure how much money was invested in venture funds, and you can measure the mergers and acquisitions and IPO value of those companies, starting with Microsoft and Apple, by the way. And the value produced vastly outweighs the value invested. So as an asset class, the facts are just the facts. Now, it's hard to invest in an asset class. That's one of the problems with venture capital and SignalRank. If you could invest in venture capital as an index, like across the entire asset class, you'd do very well, but you can't. There's no instrument that you can buy that does that. So CalPERS have to go negotiate fund by fund by fund to get into each one, one at a time. What SignalRank is building is an index approach to venture capital, based on Series B rounds, where a lot of the risk has already gone out of the system. And I do think an index of venture capital would be a super attractive asset class to any investor. Wealth managers at places like UBS, if they could have a product that was a private market product, which is what venture capital is, that returned what venture capital returns, they'd be super happy, as their clients would be. But it doesn't exist. It all comes back, and we'll get to this with the tweet of the week. It all comes down to the emotional aspect, the unreasonable aspect of investment, which is both the problem and the charm, the up and the downside of investment. I'm going to take a short break. Everyone needs to pay for their breakfasts. Keith, no free lunch. My new sponsor on the show is Liberties, a quarterly journal of culture and politics. And I think our kind of viewers and listeners who are interested in venture capital and the future need to have a better understanding of history, too. I'm going to run a short ad for them, and then we'll be back. And we'll talk about AI of the Week, Essays of the Week, and the rest of the show. So stay tuned, everybody.

Speaker

of cultural exploration, of intellectual delight, of immaculate prose. It's invaluable. Subscribe now or find Liberties at your favorite bookseller. It's full of excellent essays. And in fact, all guests on Keynum will get an annual subscription for free. Keith, the essays of your week are pretty interesting. One in particular was the Harrison piece. Generally, though, things are still a bit miserable when it comes to the news of the week. Any reason to be a bit more cheerful?

Speaker

Well, look, I think we're in a period of emotional turmoil in venture capital, and that's my world that I live in day to day. Well, that's your euphemism. What's my euphemism? Emotional turmoil. Yeah, I think it's a euphemism for people running away from the ball. You know, a bit like Spurs when they play soccer. Oh, thank you. What about Manchester United? Speaking of running away from the balls, you've got your own problems at Manchester United. Exactly. Running towards the balls, I think. That's another question, another show. There isn't a lot to be cheerful about. What about on the AI front? Even your essays of the week, when it comes to AI, we've got Gary Marcus suggesting that, what if generative AI turns out to be a dud? He's very influential. I know you did a debate with him on the future of AI. John Thornhill at the FT, very respected tech commentator, also did a piece on that. I think it's the top rated, the most viewed piece now on FT. Is everything getting sucked into all this pessimism, Keith? You know, there's a kind of a cousin piece by Ian Bremmer that contrasts. Yeah, which is less miserable in its own way. Yeah. Now, if you look at what's going on, Gary Marcus represents almost a one man campaign against large language models. He's a zealot. He's clever, so he can put an argument together. Yeah, and he's an AI expert. I mean, to be fair, he's not some journalist. No, no. He's absolutely an AI expert, but he's an expert from a different field of AI that doesn't like large language models. I believe that. And if you wanted to be kind, unkind to Gary, you might say that he's a little myth that he wasn't the one who invented generative AI. That might be true, but I don't think he would ever have gone down a large language model path. He doesn't believe in it. By the way, he's completely right. Probably, I'll add the word probably, that artificial general intelligence will not come out of what open AI is doing. It's going to need a different path. I agree with him on that. But I don't agree with him that it's either a dud or scary. I've used open AI this week for coding. Yeah. I could not have written the code I wrote. It's the stuff I showed you before. I agree. I mean, I don't see how he could call it a complete dud. But you talked about the Ian Bremmer, Mustafa Suleiman piece in Foreign Affairs, the AI power paradox as a cousin of the Marcus piece. I'm not sure it was. I mean, I don't think that Bremmer thinks of AI as a dud. He just thinks it's going to result in profound political change. Well, I think he's, I would characterize him, if you think of the spectrum, Gary Marcus, stop it now. And me, leave it alone and let it blossom. Bremmer's in the middle and says we need to regulate it, but we need to regulate it in a clever way that involves all of the developers who are building it, as well as all the nations. That it must be a global regulation, not state by state, because that won't work. Yeah. And I would have thought you would have liked the Bremmer piece. I do. Because you've always been someone who's argued that 21st century represents the end of the nation state. I mean, what Bremmer's saying is that tech companies like Microsoft and Google now are great powers and that they have to be acknowledged and that there needs to be a seat at the table for these companies when it comes to discussing AI. No, I completely agree with him. Funnily enough, he doesn't reference ICANN. And ICANN is the one global institution that is structured in the way that he describes, where stakeholders, including nations, are all sitting at the same table, including the tech companies, the people who sell domain names, the people who run the infrastructure. They're all sitting at the same table. And ICANN, whatever you think of ICANN, it is successful at running the domain name system, which is its remit, and the IP addressing system, which we all sit on top of as we do restream here. So, yeah, no, I do agree with Bremmer that that's the right approach, a multi-stakeholder global conversation with some self-reflective rules that create parameters. I do agree with him. That said, it may be premature even to do that because AI is at such an early stage and there's so much change happening. I don't even think we know what it is yet. Yeah. Well, we may not know what it is, but we know what it is. And I think the Bremmer piece is interesting to read alongside the Marcus piece. I'm not sure they're related. They're obviously about the same subject. I was pleased, Keith, and this was not at my bidding, that you included my conversation with Dave Weiner on Keynone, on his origins. Dave remains a legendary figure in the field. What would that conversation teach us about today's moment in 2023, do you think? Well, look, I think Dave Weiner is a super interesting guy. I've known him since the end of Real Names. He helped me create a blog on his radio userland platform in 2002. So I've known him a long time. I helped him sell his company Weblogs.com to Verisign in, I think, around 2005-ish. So I wouldn't say he and I are friends, but I think we like each other and we respect each other. We don't know each other well enough to be friends. I thought your interview with him was really good and timely. Dave basically has been the cheerleader of distributed publishing and subscribing from the very, very beginning. He invented some things. He championed other things. But he's been a key driver of everything and still is. He's writing code today. He's releasing products pretty much every month. He releases a new product. So it's a wide-ranging look at publishing and subscribing to content over about a 30-year period. Yeah, and in some ways, the conversation suggests that not much has changed. He reminded me of his engagement with Bill Gates in the mid-'90s, which triggered a lot of his work. His hostility to top-down media, of course. We still have the same issue in many ways. Elon Musk is the first cousin, maybe even the brother, younger brother of Bill Gates. What about, Keith, more concrete things? I think there's some interesting news that Google's DeepMind is testing a personal life coach AI tool. I think it would have been news if they weren't. Isn't everybody testing one of these devices? What's the problem with that? I think it's noteworthy just because Google has hesitated to do that for such a long time based on the fear that this coach might not be a good coach. In that sense, by the way, Gary Marcus's concerns are correct. Most of the large damage models are not reliable coaches yet, except in very specific areas like coding. It wouldn't be any worse than human coaches. They're usually wrong too, right? Yeah, exactly. Human coaches can be very damaging, just like therapists, if they reinforce negatives, which can happen. It still seems the logical next step when it comes to AI in terms of a real product is a personal assistant, a really easy-to-use personal assistant. To me, that's the Netscape moment, when we can all use it, not just you and Dave Weiner, but real people who don't understand tech. I agree with that. Do you remember Clippy? Yeah, but it didn't work, did it? It didn't work, but the idea of a personal assistant that can help you goes back a long way. Yeah, an algorithm that puts all my shows together. I go about it in exactly the same way in terms of posting and performing and all the rest of it. It shouldn't be that hard, but I don't know how to start. Yeah, I have tried to use OpenAI to produce That Was The Week newsletter, and I can give it all of the links and ask it to organize them into sections and do short summaries. It can kind of do it one by one, but it actually takes longer than me doing it myself, so it isn't there yet. But the dream would be I read stuff, I favorite things I read, it picks up the things I favorite, produces That Was The Week with the right subheadings and putting things in the right place, and all I've got to do is write the editorial. That would save me about two hours' work. Yeah, and for me, the AI would recognize emails that confirm interviews and do all the rest of the heavy lifting, which I have to waste my time. I mean, I guess I could get a personal assistant, but by the time I explain everything to them, I might as well do it myself. Yeah. Calendly has a negative side to it, but actually one of the good things about it is it can do that. You just get people to pick a slot that you've predefined and it goes in your calendar without you doing anything. Well, it's going to happen. I think that's one thing we can say for sure, whether it's in the next five or 10 years is the question. One interesting piece of news, more down to earth, is that the VC firm Blackbird is selling its stake in Canva at $25.5 billion. Canva, in a way, is maybe a sneak preview of an AI world, is it, Keith? Well, it's integrated AI and it is doing more and more to be able to automate designs for things like slideshows. So it's a threat to the Adobe's of the world. Well, Adobe is doing it as well, but in different ways, like you can put a photo in Adobe that's maybe square and it can fill it out to be landscape by artificially creating additional organic extensions to the photo. And you can put a house in a field in Adobe and tell it what kind of a house. So Adobe is starting to do things. But this Canva news, it reminds us that there is some normal out there and that not everything is dark and dire. Yeah, no, I think that's exactly right. Most things are good, by the way. The only bad thing is the emotional reaction to the market change in 2021. Everything else is good. And the funny thing about us human beings... But that's like saying after, you know, the Abraham Lincoln assassination. And how was the show, Mrs. Lincoln? No, it's a bit broader than that. It's how is the future of America, Mrs. Lincoln? And the answer would be good. Yeah, after her husband just got assassinated. I'm not sure what she would say. Well, in a way, that's my point. As humans, emotions dominate our short term brain. But that's who we are, Keith. And we're never going to change that. Even your algorithms isn't going to change that. You have to have an emotion to make the decision to use it. But the whole traditions of critical thinking, philosophy, economics allows us to abstract and to think historically about the moment we're in and where things are going. And that is, you know, above emotion. Now, of course, you have opinions that shape outcomes because human beings... History doesn't just happen. Choices are made by humans. So emotion is super important. To misquote someone, we make our own history, just not quite in the way we think we are. What about when it comes to Africa and China? You had an interesting piece on Biden's restricting U.S. investments in China's military might. Is this a big deal? Are we slipping and sliding into a digital Cold War? Look, I think if you look at Biden's conference that he's having with Asian leaders in the next period, and you look back at, even go back to the Trump regime, the one constant that both Trump and Biden shared is a trade war and a kind of a Cold War with China. Which is, you know, on the one hand, you could argue is rational because it's slowing down your likely competitor. On the other hand, it's entirely irrational because it's shrinking the world economy and making life harder for most human beings. It's reason and unreason, which we need to disentangle. It's certainly significant. We've talked about it before. And probably that news needs to be read in the context also of Bremer's piece on nation states and AI power, because that's the reality on the ground. Just one thing, Andrew, just to retrieve one thing. There's really no good outcome for us humans if nation states fight. I mean, there just isn't. It's technology. Unless the nation state wins, like the United States or even China. You're part of that nation state. Even that isn't good. Would you argue that it was good for the UK to fight World War Two, even though they won? I mean, look at the bad. That's a red herring. I'm not going to go that way. We could have another show on that. When it comes to Biden and Trump, it's interesting. I don't want to talk about Trump, but in many ways, Biden continues in that Trumpian critique of China and globalization. So in some ways, we get obsessed with Trump for various reasons, but perhaps he's a bit more normal than we think. Finally, when it comes to Startup of the Week, you've always been very bullish and particularly interested in Africa. Your Startup of the Week is Mastercard's decision to purchase a minority state in MTN's 5.2 billion tech business, which is an Africa-related story. Tell us about this one, Keith. MTN is a digital platform in Africa, mobile phones and the like, and also media services. It has built a peer-to-peer payment system, not unlike PayPal was originally, and it's now huge in Africa. There's another big one in Kenya. Africa is quite far ahead when it comes to digital payments because it didn't have the old infrastructure, so it built the new one first. Mastercard is now recognizing Africa as a first-class citizen for its investments, which I think is huge. MTN is all over Africa. When I go to South Africa with my family, MTN is everywhere. When you go to Nigeria and Ghana, it's there. When you go to Kenya, it's there. And there's other big companies like Safaricom, which is a huge African cellular and payments company. I put it at the start of the week just because I don't think people realize how advanced Africa is becoming in infrastructure. And is this mobile infrastructure? So it allows them, as some optimists believe, to bolt over any kind of wired infrastructure? Correct. Satellite and mobile are the two big techs. There's no cable networks, it's satellite. Yeah, so that won't be the last we hear of that. Finally, Tweet of the Week, and it's not Tweet anymore, it's X of the Week, appropriately. In this sense, usually the X of the Week is not connected with the rest of the show, but this is central. It's an X from Visual Capitalist on the rollercoaster of emotional invested borrowed from Fidelity. What does that tell us, Keith? And it comes back to this issue of reason and emotion when it comes to investing. Yeah, well, the graph in the middle there is very telling. The graph basically says how much money you would make if you were in the market permanently during the ups and downs. And that's the left-hand pie chart. So this is for people just listening. We have two graphs. One that looks like a rollercoaster, makes you feel a bit queasy. And the other, which is a much more conventional, slow rise. And that's the one you like, isn't it? Yeah, so the left-hand bar there is you didn't use your emotions at all. You just stayed in the market. That's how much money you would make. The next one, the next lower one, is what if you missed the 10 best days in the market? Just 10 days, if you're out of the market for 10 days. And then each subsequent decline in money made is staying out of the market longer. So I don't know about you, when I invest in stocks and shares, when shares go down, I want to sell them. And when they go up, I want to buy them. And that's the exact opposite of what you should do. So emotions kill value growth. That's basically the point. And can we get rid of emotions? I mean, I guess Fidelity are in the business of trying to do that. All these investment wealth management companies try and convince us to overcome our emotions. That's what you're in a way trying to do for investment at Signal, right? Do you think we can ever overcome emotions when it comes to investment? Only by outsourcing decisions to unemotional agents, maybe co-pilots. And can we build algorithms that don't include emotion? Well, ideally, yes. Algorithms should be just math. But doesn't math contain emotion? Usually not. It doesn't depend on which math you take and how you present it. Well, there's a deeper conversation there. What you're talking about is heuristics, which is math that seeks an outcome using a predetermined formula. That's a heuristic. And yes, that has emotion in it. So my heuristic at SignalRank says good investors stay good for a long time. And if you score them and understand who they are and track what they do, you will end up in good places. But that's not the same as machine learning and AI. Machine learning and AI just say, here's the data. Tell me what you can derive from it, where the goal, you give it a goal. The goal is to make more money with my dollar than any other way. Read the data and tell me how to do it. And that algorithm should not be emotional. It should be just logical. Well, speaking of goals, Keith, tomorrow is Manchester United versus Spurs or Spurs United at Spurs. The 5.30 p.m. kickoff in London, 9.30 on the Pacific Coast. What does the algorithm, what does a dispassionate analysis tell us? Who's going to win tomorrow? I would say Spurs let in two goals against Brentford and United didn't let any in against Wolves. But Spurs scored two against Brentford and United only scored one against Wolves and struggled to do that. So I think everything about it says a nil-nil draw.