Dry Powder?

Speaker

When it's cold outside, I've got the month of May. Everybody say, I guess you say. What can make me feel this way? It's my girl, my girl, I'm talking about my girl. My girl, I've got... Hello, everybody. It is the evening of August 10th, Thursday, 2023. It's almost Friday. Almost time for that was the week. And I'm traveling tomorrow, so we're doing it early. Last week, we had a show, Keith and I, about what he called the bonfire of the unicorns. The unicorns are burning and things are really bad on the investment front. This week, he has another metaphorical theme, dry powder, Keith. Why dry powder for this week in that was the week? Are things getting even more miserable on the venture front? So in my world, dry powder was the topic of the week. And what dry powder means is venture funds have raised a lot of money. Therefore, one would assume have the ability to write checks, but are not doing so. Hence, their powder remains dry. So it's a way of talking about venture capitalists going on investment strike in a way. Bill Gurley wrote a very well shared and read tweet now from now on with an ex post talking about this. In response to the information saying that there was 271 billion of dry powder. And the implication in that headline is dry powder is going to be invested. Therefore, things are much better than you think. And Gurley says, well, no, not really. Things are not better because dry powder can remain dry. That's the essence. And everybody weighed in on both sides of the argument. So what's his argument? I mean, you just assume logically that venture capitalists, that's what they do. They invest money. And if they've got funds, why wouldn't it be dry powder? At some point, they've got to put their money into the market. Otherwise, they might as well just shut up and go and get real jobs. Well, there's some perverse incentives in venture. Venture is measured by something called IRR, which is a mathematical measure of how fast money turns into more money. So time is a key determinant of what that number ends up being. It's roughly speaking, it's how much you pay per year to your investors compared to, say, the interest rate of 5 percent. And a good venture capitalist will invest, will return about 15 percent per year. But if they invest their money at a time when things are slow and later rounds are not happening for the companies they invest in, time extends and their IRR declines and they get scored against in the league table of investors. So it's much, much harder for them to raise the next fund. So the incentive is not to invest unless the companies they invest in are going to be invested in by other people and as fast as possible. So this is a vicious cycle. Things are getting worse and worse. And as they get worse, the logic is that they must get better. But actually, you're suggesting that the logic drives the thing down. The logic, almost everything drives things in the other direction when it comes to risk. Now, that doesn't mean nothing's happening. There are still really good early stage investors who are investing, although at a slower rate. My wife has an article this week saying that seed investing is down for the first time significantly. That's to say the average check size has been down in terms of number of rounds for a while. So every single part of the venture capital ecosystem is slowing and declining. So in that context, it's pretty hard to get it unclogged. To get it unclogged, there has to be massive readjustments on valuation. There's quite a few articles in this week's newsletter that speak to that. But isn't that inevitable, Keith? I mean, everybody has to do that. And what's the choice? Yeah, but it's pretty dramatic what they have to do. So Jason Lemkin from the conference SASTA has a piece this week where he talks on Twitter about a company that is doing about $1.2 million in annual run rate. And last round was valued at $60 million. And in his opinion, today is probably worth $20 million. So an investor won't actually look at that company unless the company reorganizes itself prior to seeking money. So what does that mean? Layoffs? No, not necessarily, although it might mean that if they're burning too much cash. What it definitely means is it's called cramming down. They basically have to reduce the value of everyone's shares by two thirds and then go to market with a $20 million valuation instead of a 60. And that's on the CEO. So he has to persuade all of his old investors to go along with that. What's the choice? You can say that, but people don't face reality. The reality is either cramming down or bankruptcy. It's not much of a choice, is it? There's cramming down, there's bridge rounds, which is favored, by the way, to cramming down. Bridge rounds is where you write enough of a new check to keep the company going for another year without doing the cram down, which means you cling on to the fake valuation for longer. Now, if you're a VC... Is there any evidence of things improving? I mean, what has to happen to change all this? What about AI? Well, that's the key. I think what has to happen is the new investments that have been made this year, mainly at the pre-seed and seed stage, will mature into A rounds and B rounds and will not be burdened by these crazy valuations that happened in 2021 and 2022. And it will be a new generation, if you will, a new cohort that is free of that legacy problem. That's the most likely thing that can happen. And that will come out of the seed stage. What's missing is the answer to the question, well, who's going to write the big checks later? You know, because there's no liquidity in that part of the market, where's the money going to come from for the big checks later? That's the biggest open question right now. Well, I can't say I feel sorry for VCs. I don't think many of us do. You may, but they've made so much money in the past. Talking about AI, you had an interesting piece in the Harvard Business Review in the newsletter this week. And this makes sense. AI won't replace humans, but humans with AI will replace humans without AI. What does this piece suggest? Yeah, so this is a professor from Harvard. And he reacts against the idea that AI is dangerous or will destroy all jobs. But he makes the point that many, many jobs will require a knowledge of how to use AI. And if you are in a job that is likely to require that, unless you learn the skills required, you'll be replaced by someone that has them. So it's about upgrading the workforce to embrace the new skills. And, you know, I've been playing with AI for quite a while now. It's not that hard. In fact, it's easier than not using it once you break the ice and dive in. So I think he's right. I think he may underestimate the job replacement effect. I think it's probably bigger than he thinks. But he's certainly right that those people who are in jobs will need to have AI skills, either building AI, training AI, or using AI in their job to get the job done faster and better. And when are we going to start seeing some of these AI products, these apps? They're really already emerging now in most fields. You know, they'll be invisible to you a lot of the time. Like Apple announced this week that they're going to be replacing Siri with a more sophisticated AI-driven version. They're also going to be using AI behind the scenes in lots of things. Cars use AI, and you don't see the AI. But what about an AI assistant, a personal assistant, which is what open AI has been promising? That seems to me the most obvious and key for us there. Well, that's already there. You just have to log on to ChatGPT and ask it to do things for you. But doesn't it require it to know you? It's not just you asking questions. No, it doesn't require knowing you. But they have now built in, in the settings, you can tell it about yourself. And you can tell it whether you want serious linear answers with logic or whether you want more playful answers. For example, if you're in the area of fiction, you might want that. So there are ways to teach it about you. But it'll work without that, depending on your job. And, you know, if your job is copywriting, and you give it some copy, it'll edit it. If your job is, I don't know, subtitles, and you give it some French that you want to turn into English, it will do that for you. And if you're working at Crunchbase, not only will it do it for you, it might result in layoffs, another piece. And I know you have particular interest in Crunchbase because your wife, Shanae, works there. Five months ago, according to Fortune, Crunchbase vowed to avoid layoffs. And then AI has changed everything. In what sense has it changed everything? Well, what Crunchbase does is it pulls in data about company fundings from various sources. And it puts them into a database. And I know that they have a lot of people, more than 100, whose job it is to make the data read well and be accurate. So, obviously, AI is very, very good at copy editing. And it's pretty good at writing and reading from databases. And so the engineering focus at Crunchbase, more and more, is going to turn to automated publishing of information about funding rounds. Which means that they don't need as many editorial writers. They probably don't need as many engineers over time. And they definitely don't need as many data managers. Has this resulted in the same thing at your startup, Keith? You're in a similar kind of business in some ways at SignalRank. Did you promise no layoffs? And has AI changed everything for you? It hasn't. I mean, we're very small. And we started life as an AI company. So we never hired people to do jobs that could be automated. So SignalRank's entire end-to-end, from getting data out of places like Crunchbase and Prequin and PitchBook, putting it into our database, scoring it, producing outcomes from the scoring, and even making investments in the companies that qualify, the entire flow is automated. And five of us are enough to do that work. Actually, I forgot to give you this piece, but earlier this week on Keynone, I interviewed the professor of computer science at Oxford University. I'm going to tell you his name, Michael Waldridge. I posted on the 7th, the beginning of this week. And he told me something that was interesting. I don't know whether this is common information, but he said that generative AI or chat GPT was a bet and that the very idea of it was invented by Google. And it was only open AI and Microsoft that actually bet on it. Does that make sense to you? It does make sense because all of the early attempts to act on the bet backfired. If you probably remember those stories where Google's AI and Microsoft was accused of being racist within a couple of days, not just racist, fascist even, going way back before chat GPT. So there was a lot of fear in those big companies about being the first to do it. And chat GPT's bet was present inside Google. In fact, the famous scientist who left Google recently saying that large language models were dangerous. Geoffrey Hinton. Geoffrey Hinton was originally an architect of large language models and has gone to the other side. Now, there is also, you know, AI engineering is as conflicted as Christianity or religion in general in that there are lots of different points of view that share in common some cause. And the differences, however, outweigh what they have in common and they fight along. So large language models have enemies from other AI engineers. And Google is one place where that's very strong. Yeah. And Google, of course, is doing rather well. It stocks as Microsoft and Apple and Amazon, all the big players. And we've talked about this before, Keith, but I mean, one of the reasons for the crisis in venture and the drying up of all the funds is that it's the big companies who are shaping this economy. It's very different from Web 2.0. Well, that's true in, I think, data centric machine learning models. It's less true when you look at some of the other big things happening in the world, like in biotech or attempts to re-engineer power generation from fusion. Yeah, I'm doing a show on fusion, but fusion is a slightly different category, isn't it? I mean, the kind of venture that you cover and are interested in, they don't invest in fusion. Well, some of the people I know do. Sam Altman being one of them, by the way. Right. He has his own fusion startup. Yeah. I'm trying to cheer you up. You can cheer me up just by, I mean, just point out that there's still tens of billions of dollars being invested every quarter in new startups. So that's a big idea. Maybe we need to just rethink what a startup is. I mean, Google is, and Microsoft, they're startups. They have startups within them. They're investing massive amounts of money in driving the economy forward. So whether the innovation is coming from within Google or Microsoft, especially as I suggested earlier, I mean, Google came up with the generative idea, which OpenAI bet on. So what do we lose when the venture ecosystem withers away, if innovation is being driven now by these large tech companies? Well, I don't think you get a solution to the innovator's dilemma within the big companies. You know, Christensen's famous declaration that it's super hard to kill your existing business. So the biggest news around chat GPT was the threat it represents to Google search. And Google will, as Microsoft will, desperately try to make AI search enhancing as opposed to a replacement because they depend on search for a lot of revenue. So the innovator's dilemma will never be solved inside big companies. They've all read that Christensen. They know this argument and they understand that ultimately they need to blow some of their own businesses up. But they can't. They can't because they're public companies. Look at Google's dependency on this little thing called CPC. It stands for cost per click. It's about 90 plus percent of all Google's revenue comes from cost per click. So if people aren't clicking on search results to go to Web pages, that business is gone. So Google has to slow down the transition from today to whatever tomorrow brings. And before it can allow that to happen, it's got to have a replacement like HP had inkjets before they killed laser printers. They didn't just kill laser printers. They had an inkjet replacement, which was a good business. So the innovator's dilemma relies on innovation and innovation. Almost always still today is coming from individuals who are outside of those big companies. Well, we'll have to see. I hope I can cheer you up in future shows. You're I'm cheerful. Your misery is. It's catching and other bad news this week. I was in New York a couple of weeks ago. I went to a WeWork office and actually seemed to be buzzing. But the WeWork news is also very bad. The as a going concern. There's a warning that may not be viable in the long run. What's happening with WeWork? I mean, it's not really a tech company. I mean, it's a real estate play. Yeah. Well, you remember that WeWork was taken over by SoftBank. Pretty much. It then became a public company. It was in the past valued at $42 billion. Today, its valuation is about $500 million. And what it's saying is that based on the rules of accounting, it's trading insolvently or close to it and may have to file for bankruptcy. And that's because the core business is capital intensive. And the unit economics for every dollar invested, the possible profit that comes out of it is super small. So it's a low margin, high cost business, which the headline there is a reminder that venture capital shouldn't invest in low margin, high cost businesses because that isn't what venture is about. So there's a lot of hubris in those investments, clearly. And then there's the media ecosystem. That might be a glimmer of light this week. Your Startup of the Week, and I think they've been Startup of the Week before, is Substack, a better Internet for readers, according to three of Substack's senior people, Chris Best, Hamish McKenzie and Jayaraj Sethi in their reads this week. What is Substack doing that makes them your Startup of the Week for August the 10th, 2023, Keith? Well, I think it's really good that they've acknowledged that building a place for writers, and they've broadened it beyond writers, really. It's also podcasters and video podcasters. You know, that creates a potential destination for readers or viewers or listeners. But it doesn't actually build a destination. In order for it to become a destination, there has to be features that appeal to the consumers of that content. And, you know, you and me on the show, and it must have been more than two years ago, it was very early in the life of Substack when they announced a lot of the new features for writers, I made a point that was not well understood, which is they've got to do the same thing for readers. And now, two years later, they're announcing they're going to do it, which is why I made them start up the week, because it's... But what exactly are they doing for readers? They talk about... It's clear what a Substack writer is. What's a Substack reader? They don't define it. That's the interesting thing. It's very vague. But I think the reason that they're thinking about it is that Twitter, especially, you can now write a long piece on Twitter if you have the blue checkmark, or should we say X if you're a premium subscriber, which is the new language. And so writing at length for an audience, that is not going to be enough for Substack to be differentiated. So they're going to have to do reading just for the very fact that they're opening up to the fact they need to do that. I think it's great. But isn't the unique thing about Substack, certainly in comparison to Twitter, is there's a business model for creators. The subscription model is an alternative to advertising. That's what attracts them to me. I find the advertising model, especially for podcasting, it's increasingly hard, especially in this downturn. And you don't want to litter your show with trashy ads for stuff that's either embarrassing or worthless. Totally agree. But readers think the same. I don't know about you, and I disagree with some of my colleagues on the Gilmore Gang on this who like advertising-based video, but I hate it. Disney this week has just increased the price of Disney Plus, Hulu, and ESPN to $25 a month for the non-advertising version. The ad-supported one is, I think, $10 a month less. I would never take the lower price and get the ads. I hate the ads. So what chance do you think Substack has of defining a space which is really viable in this, what they call the media ecosystem? Well, I think it's already defined a space that's viable for, let's call them broadly speaking, creators. You watch TV and you'll see on CNBC or CNN or MSNBC or HBO, people are saying my Substack is the following. So Substack is now a home for a lot of intelligent, creative people who write or record really good content. And it's decent when it comes to readership. I don't know how they can mess it up. All they've got to do is improve the features on both sides, and it should keep growing. And it can be the place where you go for the good stuff. I mean, the real question is how many Substack subscribers, I mean, how many subscriptions are readers going to pay? I mean, it's one thing to pay $50 a year for one or two of your favorite writers. But when it comes to 10 or 15, so you have to assume in the long run that there must be some sort of broader Substack subscription package that will allow you access to, I don't know, 15 or 20 of your favorite writers and podcasters and videographers. Yeah, I think that bundling has to come. I actually think sponsorships have to come as well, not ads but sponsorships. Yeah, but then you don't need Substack. Then you can do sponsorships anyway. You can do them on podcasts. There is a site called Beehive that's doing sponsorships. To all intents and purposes, it's copied Substack except it's engaged with sponsors to create sponsorships like exists in the podcast world. But I think that has to come as well. How are they doing, Beehive? Quite well. They're younger than Substack, but they're doing quite well. And finally, as always, Tweet of the Week. We should call it X of the Week. And it's another Twitter tweet. What's this one, Keith? It is, and it is X of the Week, Andrew. I made the change. Oh, good. So what this is is a little bit funny. Twitter has decided to auction off all of the memorabilia, including the neon sign that's on the front of the building in San Francisco. They're auctioning all the memorabilia off that refers to Twitter. So they're cashing in on the old brand as they retire it. And there's a long list, the links in the newsletter, of all kinds of stuff you can bid on. I think it's September the 19th, if I'm not mistaken, when it goes live. At what point are people going to stop using the word Twitter, do you think? I would guess never because, you know, Meta is still Facebook and Alphabet is still Google. Yeah, but Zuckerberg never replaced the word Facebook or the brand Facebook with Meta. It's an umbrella term, a holding company, and the same is true of Google and Alphabet. I mean, Musk is actually replacing the word Twitter, the brand Twitter, with X. He is, and he's succeeding. I mean, I would predict that a couple of years from now we'll just think of X. I did a show today on black Twitter. Apparently 70% of African-Americans are off Twitter now, 70% of women as well. So I'm not convinced. Do you mean 70% of people who were formerly on it? Yeah. I think that's very unlikely to be true. So you think that X can succeed? No, what's the reason to leave other than Elon hatred? There's really no reason to leave. Everything you want is there and hasn't gone away, and if anything, it's even better. Right, and actually one of the other news pieces of the week that you cover is something about Linda, whatever her name is, the new CEO, who's a classic advertising person. So on the one hand, you've got a very traditional advertising person running Twitter, making it just as it's always been, and then you've got Musk, so maybe he can do both simultaneously. Yeah, she was on CNBC this morning at some length. I interviewed her in the headquarters in San Fran, and she made the point that she's running the company and he's running next-generation product development, and he doesn't interfere in the advertising stuff. She listed a whole list of really well-known advertisers that are now back on Twitter, including Pepsi and Coke, and it feels to me as if the death of Twitter is massively overstated. I think it's fine. So innovation is security. Things aren't quite as bad as you suggest, even though the powder will remain dry on the venture capital front for a while. Yeah, it's business as usual, but there's going to be no great swing back to the good old days of 2021. I almost said 1921.