Speaker unverified: belief

Transcript context

…ve heard. I'll tell you why it's simple because you're going to make me write no checks. So I guess I'd say most of the AI companies I’ve invested in probably three or four. Three of the four I'd say. The only one I might quibble with there, I think, that's very good for staying out of trouble. And this is where I tend to push back when people are saying, oh, it's an AI bubble. It's like, I dunno, I think people are happy with the tokens they're buying. I think a lot of tokens are happening and just generally they seem to be delivering useful outcomes… because they're actually delivering value on the customer service side or people enjoy their Midjourney adventure, but people are getting value from the products. So it's a pushback that doesn't… I was going to push back on number four, which is positive unit margins because just aren't the underlying costs… Again, when you guys started Fin, it sounds like you were underwater… Yeah, we were losing money on it… But then just pretty quickly it right-sizes as you optimize it, and so couldn't one be too focused on the current implementation? Yeah, I mean this is a conversation we have internally with our CFO quite a bit actually because, we’re good, I can imagine it’d be the kind of thing a CFO would want to… Hey, Des, do you have five minutes? That's exactly, yeah, “quick chat.” I can't help but notice the team have done this preemptive loading or whatever. It's causing this shitload of money. So what's my counter? I guess I prefer it if the path towards profitability isn't just OpenAI is going to figure this out for me, right. An interesting way I'd say this, with Fin for example, obviously our profit goes up when we are firing less dead tokens. A dead token being we've generated an answer and it wasn't right, so we can't charge money for it. If you're, say, guessing the next line of code or tab to autocomplete the next line of code, if five of six of those is wrong, I don't know if you're ever going to get bailed out because you’re basing five-sixths of your costs is not something you can resell. So there's a questionnaire of how much of your tokens are actually generating a thing that a user wants independent of what you charge for as long as the user wants it. I think you're always in good condition, whereas if you're burning a million tokens to find one, and that one, you're never going to be able to recoup your costs or at least I'd love to see your telemetry to make sure that you actually have thought this true. I suspect you haven't. 45, that’s not bad. Yeah, I was very impressed by the multiple, clearly deliberate twenties. I mean we're definitely all different. A lot of key things we all agree on, Eoghan would be like a, first and foremost, he's a very strong CEO. He is very decisive and he is very brave, is the best way I could describe it. An interesting thing, when he returned to Intercom, one of the things he did was basically rebuild the culture, and one of the things he focused on was resilience and open-mindedness. People might be offended. sting thing, when he returned to Intercom, one of the things he did was basically rebuild the culture, and one of the things he focused on was resilience and open-mindedness. People might be offended. Yeah, of course. All of the things, all of the downsides you'd possibly guess are all there. I also think that there's no path to … there's no way. The phrase I've settled on when I look back, is, sometimes you have to go too far to know you've gone far enough. And we have an AI assistant in the product. Yeah, exactly. And we've updated our homepage and say we're AI first, so we're good. And I think you need to be willing, genuinely willing to make brave hard to undo bets. And I think you need obviously having this sort of moral authority of a founder and being CEO kind of gives you some of that, but still it's a huge decision to make. And I think I am much more of an, my default DNA is I'm more of an operator in the sense of, “All right, what are we doing? OK, well I'll make it work. ” Whatever it is. And I think if it was a company of people like me, what you'd see is probably predictable, reliable, sustainable performance or whatever, but probably not enough actual, well, definitely not enough brave big swings, which is actually where you need to get to. You've now worked with so many different companies externally, you've seen a lot. What is predictive of success and what is predictive of failure? The biggest thing I'll always come back to when I'm talking to anyone who's trying to pitch me to invest, or me to induce John to invest, is it's always some version of: Do you have a real product that solves a real problem that really exists and people are really already trying to solve by paying money or time somewhere? It sounds so trivial, but you'll be shocked how many times you'll fail or you'll get some sort of jazz hands-type routine somewhere along the way where it's like, “Don't look too much at this, but just trust me.” The areas that I end up being blind to is the extremely market-expanding type things. As in if someone said to you, “Hey, all companies are going to have a chatroom and they're going to all hang out in an all day and have unproductive conversations, it's going to be big.” I'd be like, “Oh, I don't see it.” Whereas that’s how you would've missed out on Slack, or whatever. But I think I can almost hear from the what are you building and why and who's it for, and show me what the product does if it's not a real solution to a real problem, I'm kind of already out. Soho House. I think I can almost hear from the what are you building and why and who's it for, and show me what the product does if it's not a real solution to a real problem, I'm kind of already out. Soho House. Exactly all of that, right? To some degree, like remote working, into the mix as well. But I think the amount of people who are chasing the trinkets of being a founder of a startup, even if they're quite smart and they can actually kind of go and build something, if their actual motivation isn't the problem or isn't just some deep desire to be quite successful, but it is instead to be perceived the whole, “I could have been a contender. Rather than I could have contended.” If you don't really, really want to actually play the game, instead you just want to be seen to be playing the game. I think that's probably the single biggest thing that tells you you're probably, best-case scenario you'll sell at 5 million. But more likely you'll still be alive in seven years, all your investors will wonder what you're doing, and you'll be basically sending one investor update every now and then. Yes, I have noticed investor updates with metrics don't predict success, but investor updates without metrics that tell a really fancy story, but don't have metrics, are actually quite predictive of failure. Those companies always fail. I basically 100% agree. And honestly, you can even tell where the metrics are in the update because oftentimes my favorite updates, I mean this company actually probably should … And no o investor updates are fine. There's a bunch of successful companies that just never … We never sent investor updates. I'm sorry for all the investors, but we're bad communicators. But if you go to the trouble of writing an investor update and then make a proactive decision to not say how your business is doing, that suggests some deep denial about what we're running a business means. So there's one company—I can't say—but we're both an investor in it, but their updates, just one of the most recent ones was just like, “Here's performance ARR plus 17%, blah plus this, blah plus that.” Something like, “I hope you can see from the numbers we're doing great. Best of luck to see you, see you next quarter.” And I was like, “Yep, brilliant archive. I'll mark it up.” Something like that. I think in general, the degree to, I think it was Paul Graham said, the ratio of numbers to words is usually the actual thing you're looking for, which is if the numbers speak, then the words don't have to. What else is predictive of success? Numbers is one. I almost want to say the inverse of all the things that I hate seeing. I hate seeing founders who invest massively in their personal brand instead of their company brand. I hate seeing people who are obsessed about … if the first three or four updates I get are begs for retweets and tweets and all that sort of stuff, that's always not a great sign because it sort of says to me you haven't worked out how to market or whatever. Anything around what are the customers saying? Whenever I reply and say, “What do customers think of this feature?” They're like, “Oh, we're going to ask them.”…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.