Deel’s growth overshadows SpyGate, Claude learned new skills, and the great “momentum is moat” debate | E2194

26:44So, you know, if you want to like send people their laptop and make sure you understand what they're doing with their computers and all that stuff. Yeah, all that is kind of in this. Well, I look forward to you taking that company public and getting them off the bench, Jason. Both Twist 500 companies. Go to twist500.com and see the top 500 private companies. And for producer Marcus, if you could let us know what percentage of those companies have been on Twist, and then I want to view on twist500.com. So twist500.com slash, let's call it twist. Then I want to see just who's been on twist, who hasn't, and just make a running list of that and a link so people can easily see in the twist500 who's been on the show.

27:35Okay, let's keep going through the docket. Yes, so there was a really interesting interview with Roloff Botha over on Uncapped. I think it's Jack Altman's show. And you talked about this on Twitter. So I went ahead and watched it. And I think it's actually a fascinating kind of meditation on venture capital as an asset class or not, Jason. Mind if I play the clip? Sure. I'll see you on the other side in two and a half minutes, everybody, unless Jason says stop, in which case. Two and a half minutes. Okay, here we go. We can cut it off a little bit. Okay, or maybe I'll say pause if we need to pause on it.

28:05Precisely. I don't think venture is an asset class. Why not? It doesn't support the numbers. So there was a lot of analysis back in the 1970s and 80s with the capital asset pricing model, and people figured out that there's this asset class that supposedly has uncorrelated returns and a bunch of asset managers deem that they need to invest a certain percentage of their endowment or foundation or pension fund into this thing called venture capital. If you look at the data, they're basically 20 companies per year on average over the last 20, 30 years that have ended up being worth in realized exits a billion dollars or more, just 20 companies.

28:40Despite a lot more money plowing into venture capital, we haven't seen a material change in the number of companies that are outcomes that are that large. And I think part of that is that there's a lot more talent than really interesting ideas or interesting companies to be built. And I think we're spreading a lot of that talent thin right now, similar to what happened in 1999, by the way. So when you look at the data, the amount of money going into venture capital right now in America is in the order of$250 billion a year. And the numbers, you know, these are all estimates. Let's just say it's$250 billion a year.

29:11Need a lot of exits. Well, let's do some very simple arithmetic for a second. $250 billion going in every single year. If you assume that the firms generate 12 % IRRs net, net of fees and carry, which isn't that great, by the way. I mean, over the last three or four years, the NASDAQ has compounded at 16%, 17%. Let's just say 12%, not spectacular. You're basically just average performance. You'd need a 3.7x roughly. On 10 years. On a seven-year exit horizon. So I'm being a little bit aggressive. I mean, maybe it won't even be that good. So 3.7x on 250 billion. That approximates to a trillion dollars a year.

29:44Coming out. Coming out. By the way, that means, and that's what the investors own. So let's say that the investors own two-thirds of the company, to make the arithmetic simple. That's 1.5 trillion annually in company exit value. Yes. Just think about that for a second. Where's that coming from? Well, Figma's worth what? It's a lot, but it's... 40-ish billion? Let's say it's worth 0.03 trillion. So, you know, if you start thinking in trillions and Figma gets you 0.03 trillion, you need 30, 40, 50 Figma's every year. Okay. Okay. To make that arithmetic work. You founded a company, you launched the MVP, but now you need customers.

30:22But how do you get in front of an audience of real people, not bots, who might actually have the budget and wherewithal to buy what you're selling? Well, I'm here to tell you, if you're doing B2B advertising and marketing your product to businesses, you need to be on LinkedIn ads. It's that simple, folks. This is not a spray and pray tactic. No, we're not trying to put your ad in front of everyone. No. With LinkedIn, you know how to find the real decision makers, and you can target them based on data that we all put in our LinkedIn profiles. You know, things like our title, the city we're in, the name of our company, how many people work there.

31:00And that data allows you to get in front of decision makers. We all know LinkedIn has a billion members. Incredible. What an achievement. But 10 million of them are C-level executives, right? These are the people who are the CEOs, the CFOs, the CTOs. These are the real deal. The kind of people who, these are the people who can approve a million dollar deal,$10 million deal,$100 million deal. And LinkedIn's going to help you refine your marketing strategy. They're going to build campaigns that work for you and target the audience you need. They're also going to help you measure your ROI, so you're never wasting your resources on ads that aren't converting.

31:33I've got a special offer here for Twist listeners. Wow, this one's better than the last one. Launch your first campaign and get$250 free when you spend at least$250. That's a great deal. $500 worth of ads? Half price. Go to linkedin.com slash thisweekendstartups to claim your credits. No spaces, no dashes. LinkedIn.com. That's already in your browser, right? Because you're there every day. And then just put a slash at the end. This week in startups. No spaces, no dashes. Thank you to our friends at LinkedIn.

32:06Yeah, so - Roloff goes on to say, oh, please, Justin, go for it. Well, no, I was just saying, he, at the All-In Summit, gave exactly this same assessment of it. And a capital asset class as defined by Claude, producer Claude says a capital asset class or simply asset class is a broad category of investments that share similar characteristics, risk return profiles. That's the key. And behave similarly in the market. Traditional assets include equities, fixed income, cash, cash equivalents, real estate commodities. And, you know, I asked Claude, is VC actually one? And he said yes and no. Great for splitting the difference.

32:47I think that's the debate here. So got that right. And sometimes it's considered its own distinct asset class due to its characteristic. But more people believe it's like a sub-asset class of private equity or alternative invest, you know, alternatives. And the distinction matters because it's not super predictable. You have the J curve, as we've talked about, this long investment. It's illiquid. You have this very, you know, feast or famine, you know, power law and active management is a key piece of it. So anyway, Ruloff is right. And he makes a funny joke in there. He says it's a return free risk.

33:32Yeah. Yeah. Return free risk. Yeah. He says you're better off investing in index funds or holding T-bills. Brutal. Well, and so then we get into the gamesmanship here. So we're all talking our books. We're all playing a chessboard. So Sequoia is trying to keep the industry small, focused, and elite. Because that's how they made all their big wins, right? You don't want to chase. So they're trying to keep their fund size the same so they can have that IRR. They're not trying to do what Andreessen is doing, which is making giant funds that, you know, revert to the mean and have lower return profiles, but more money, more capital under management.

34:11So here's your tension, Alex. which is going to get have higher IRR it's probably going to be Sequoia because they're trying to keep it more boutiquey who's going to have more assets under management probably going to be Andreessen Horowitz and these people just going big but when you go big you got to write big checks you write big checks now you're in competition you own less of the company you got a worse deal and so the industry this relates to our opening preamble here where we were talking about, hey, this could be the biggest ham. It could dwarf hams we've previously seen. So there's a big opportunity happening here.

34:47So who's Roloff speaking to? He's speaking to LPs. I think he's doing LP management in this discussion of, hey, you are an LP in venture capital, but we also have this Sequoia Heritage kind of Sequoia Fund. That's the public companies. So, hey, let's make a delineation here. There's venture, there's growth, and then there's us holding public equities. You are better off just owning the MAG-7 or owning the NASDAQ QQQ and just moving on with your life and not worrying. And you can just sleep at night because all of the stuff that comes out of venture, this huge funnel of venture, drip, drip, drip, created MAG-7, drip, drip, drip, created the broader NASDAQ index.

35:35So just buy a tech index, maybe you're better off. So there's too many players. I was talking to somebody who's been in venture for 20 years plus, and they said, it's a different game now. I'm doing seed investments, you know, put one to 3 million in. So now there's so many competitors that we're one of four, five, six people in a hot deal. And the founders are great at pitching and hyping things up and creating a marketplace. In fact, Y Combinator, Launch, Techstars, we're good at teaching them how to do that. So it's a key skill. They create this marketplace and then boom, you have seed funds sometimes coming in at a 30 million, 40 million valuation.

36:20The math does not math, Alex. You can't build a 3, 4, 5x fund. You're going to get ground down to a 1.5 to 2.5x fund. So he said, I think what I'm going to do is on my next fund, I'm just going to be strategically looking at what's breaking out and then just buying secondaries and just, you know, waiting for the breakouts and putting money into those because he's got a stellar reputation and knows everybody. And I was like, you know what? Game on the field changing, smart move. You know what I did? I was like, a lot of competition in the accelerator space. I'm going to start a pre-accelerator.

36:56Right. Get in earlier. Play a different game. Play a different game. Found a university starting in Saudi Arabia, in Riyadh, November 3rd, 4th, and 5th. Mina.launch.co. We were supposed to do 25 companies in the first cohort. They asked us, because there's so many applications, to go to 50. I said, sure. Then I got asked yesterday, hey, or two days ago, my management team, can we go to 60? I said, absolutely not. They said, oh, come on. My team's like, we got this 10 other companies. I said, absolutely not. And then I said, okay, fine, 60. this um it's a lot of companies man it's a lot of companies but it's also it's um you know what we do is so much work that most people aren't built for it you have to be willing to work 60 hours a week to work in what i'm doing here like with the team if you're going to work 40 or 50 hours a week you're just we can't pull it off because there's too many companies we need to talk to them so it is what it is jason help me understand that the tension between there is sharp elbows because there's so much opportunity right now inside of startups and venture to roloff's point that there are only so many breakout companies and then also the fact that you know across every data set that i can see the amount of money that vcs are raising is in pretty much precipitous decline so it feels like there's either more money than i think in the market or more opportunity than i think in the market but i was seeing things kind of through a pessimistic light in the near term but then your opening riff was it's the best time ever to be in involved in technology.

38:23So what am I missing between those two poles that describes reality? So your entry price matters. And if the entry price is too high on average, it's hard to do the portfolio management. One of these firms, let's say a$300 million firm has 30 names in it, you know, and you wind up investing, you know,$10 million and you own on average, you know, 10 % of each company, you get diluted down to 5%. Okay, so now you got 30 names, 5%. Let's say one becomes a$10 billion company. It's incredible. You just made$500 million. So your$300 million fund is not even 2x. You could have just put your money. Now, you know how hard it is to hit a deck of corn?

39:13I mean, people are like, oh, J-Cow, Spray and Pride, blah, blah, blah. hey man, number one company, 200 billion. Number two company, 100 billion. Please say more. Y Combinator, you want to dunk on me? Like, has Y Combinator had a company that broke 100 billion or 200 billion in valuation? I don't know, maybe. I think it was Airbnb's, their biggest. So it's really hard. And there's a lot of luck involved in this. And like, you have to recognize it and realize the power law just rules your entire existence. And in that case, I just described a firm making 30 bets at market rates, their dilution, and then where they wind up.

39:52They have to hit. Now, let's say they do hit a$10 billion company and a$5 billion company. So you got$15 billion, and they have 5 % of$15 billion. So now they have$750 million in returns. Congratulations, you're 2.x. You still didn't beat the index. That's if you hit a$10 billion company and a$5 billion company. So you hit Instacart, and you hit Reddit, and you're like, oh my God, I'm a hero. Look at that. Like every 15 companies that hit a five or$10 billion thing, I'm a genius. It's the math doesn't work. Now, if you had invested and you own 10, you own 20%, like you used to, and then you got diluted down to 10%.

40:29Now you'd be 10 % of 15 billion. You would be, you know, sitting pretty incredibly at 1.5 billion and your$300 million fund would be five exit, you'd be a legend. So entry price matters. And too many people chasing the same deals at too high a price means VCs lose, LPs lose. Who wins? Founders. There you go. Hard to be too sad about that. I mean, you know. I mean, here's the reason why it's a problem. The reason it's a problem is we need to keep the business sustainable. And it really is Darwin. You ever, like, you know, there's too many fish in the pond, and the big fish start eating all the little fish, and then, like, you know, all of a sudden, the pond gets out of whack, and some algae grows, but, you know, everybody's eating each other.

41:21It's too violent, and then all of a sudden, the ecosystem collapses. Collapses. And then everybody dies. All of the dead carcasses go to the bottom of the lake, and then the little fish and the minnows start growing because there's nobody eating them. And then you have to restart the entire cycle again. It happens in nature sometimes in lakes and bodies of water. And that happens in our industry too. So you'll see a lot of firms that are zombie firms that are walking dead right now. They're not making new investments. They're just managing their funds. And so there'll be a lot of that happening, a lot of turnover in the space.

41:57And I can tell you, you know, there's two reasons why these VC firms end. Number one, the principal, like me, or let's say SACS or whatever. They don't care to do succession planning or whatever. Maybe he does. I'm working on it. It's a 10-year plan. But maybe they just say, you know what? Enough. These bets I'm working on have to shut the firm down, and I'll just invest out of my own book, which is what Shmop did, right? Famously. He's been very public about it. And the other reason is you can't raise the next fund because you return 1.5x or 2x, and people are like, eh, I don't know. I don't believe in venture anymore.

42:32The end. There's a great clip from Andreessen investor, Brian Kim. He's a part of their AI apps investing team, Jason. And he talks - Just talked to him yesterday. He's coming on the show. Oh. Nice guy. I'm just pulling clips of people you just talked to. Love it. And this is all about momentum and what he kind of expects to see today and what's impressive. And then I want to get into the value of momentum and moats in startups. But first, let's play this 25-second clip from our man. Here we go. okay two years to be friend of the if somebody came to me and it came to us a 16th and said oh our company grew from zero to two million dollar in run rate in three months are you interested the answer is holy that's great let's dig in understand etc nowadays we grew from zero to two i'm like how how fast like a month so i'm like nah okay it's being a little facetious but okay it came yeah so that's that's the gist of this now This little riff from Brian got some notable pushback.

43:36Liz Wessel, she's a partner. First round just was like, momentum is not a moat. Momentum is not a moat because his argument is that it is. I'm just curious about moats and momentum and where you stand on this conversation and argument. This is a really good poll because I hadn't seen any of this discussion. So we talked about earlier the size of Andreessen's fund, right? And then putting a lot of work together. And then first round is the antithesis of that. They were designed by my friend Josh Koppelman, Howard Morgan, and the team over there to put 250K checks in the first round. That's why they named First Round.

44:10Pretty brilliant branding from the team over there. Long story short, they are probably patient capital. And I think Andreessen is probably impatient capital on the spectrum, just historically, philosophically. They literally named their incubator Speedrun. I think we've got five or six of our founder university and accelerator companies have gone to speed run, having good experience. It seems like a really good program. I gave a shout out to Paris program over the summer, my friend Mar over there, and Pejmon, and then Sequoia's Ark. I think Stephanie and Ruloff are working on that. And then you have 16Z, speed run, all of these, I would say, including our program's launch, accelerator, all a better option for founders now than Y Combinator.

44:57and that's not a dig at Y Combinator I just put them in like the fifth slot now because I think you get a better deal economically and you get more bespoke experience and a lot more energy on your startup. Again, not a dig to Y Combinator but it's a bit of a factory and I think that's kind of why people are saying hey, maybe there's a better way and I think that's why Y Combinator frankly is being a little defensive. We had a couple of these brouhaha's coming up. They're kind of circling the wagons and attacking people and you know, it's kind of their vibe is to be a little cantankerous, but they're facing serious competition.

45:31And the reason they're facing serious competition, and I don't have a horse in this race, we're always filled up. We do 100 investments a year. We have a hard time picking them because 20 ,000 applications, 100 investments, one in 200 chance of us investing. Like, gosh, I really want to invest in like one in 50. I think there's like one in 50, I would have just as good returns. So I have to build a bigger fund and a bigger team to do that. putting it all aside um i talked a little bit before about it being a little chippy out there the opportunity space expands the velocity expands competition expands i was talking about this vc who is like i just can't compete or i don't want to compete in a zero-sum game where everybody loses or a game where everybody loses that's not an infinite game there's like a concept of an infinite game the more you play the more people who play just everybody just does better and better and then there's finite games venture in these seed rounds or series a is a very finite game but before that and then in the growth stage it's not a finite game it becomes an infinite game because there's an infinite number of people who want to transact the shares when they're public and then there's an infinite number of people who want to start companies so if you look at it like a barbell alex you have infinite games on either end when you're public anybody come by the stock or even for andrel right now on secondary markets or spacex on second market stripe you can just there seems to be a lot of ability to get in and out of those shares with spvs and then in the beginning it's like founders like please pay attention to me can anybody just return my call right the finite game in that middle section here so there's a finite game going on here between liz and brian the finite game is who gets that 500k check-in who gets that one million dollar seed check-in and so one person is saying hey i um i believe x and the person saying i believe y and they're in competition um but the momentum point because brian says right now momentum is the only moat and liz says momentum is not a moat and then we have also nikun kothari from fpb fpv ventures and he says that the only moats out there are retention and a mission worth spending 10 plus years on now in one hand momentum in a revenue terms jason allows you to raise more capital hire more better people do more marketing and maybe went a little bit faster but it's not a panacea yeah so this is a really good point thanks for getting me back on track here so i was explaining the game on the field now we'll get to the actual issue is it is momentum a moat uh capital as a weapon you know the i saw it firsthand with uber people just kept wanting to own uber shares i the number of people who tried to buy my uber shares at 5 billion 10 billion 20 billion, 30 billion, 40 billion, right up the stack was crazy.

48:19And then Travis's ability to raise capital. And then that became a blocker. You invest in Uber, you can't invest in Lyft, you invest in, you know, deal, you're not going to be an investor in Rippling or vice versa. So you can use capital as a weapon. And how that works is you can hire people at higher rates, You can spend more on marketing and your customer acquisition costs can be higher because people are like, okay, you're going to be in the J curve. But when you get on the other side, you will be able to raise your prices, cut your costs, et cetera. And the war will end, which is exactly what happened with ride sharing and other spaces.

49:05However, we had Lovable on. The founder's smart, right? He's got a lot of competition. so you can make the argument with lovable and i'll have you make this argument make the argument for me alex of lovable if their momentum and cursor take those two just think from first principles here for a second you've heard both sides of the argument where where would you say hey yeah lovable and cursor have enough momentum that people can uh stop them and that's worth investing in or sustainable growth and having lock in other ways and building a company based on product market fit and reducing churn and having that discipline is a better way to go go ahead so the question is would i rather back cursor and lovable which are pursuing essentially a very fast momentum-based strategy or perhaps a more measured company in the same space that was taking a slower less cash intensive approach.

50:01More methodical. Yeah. Looking at churn, you know, not trying to grow too fast, not spending$10 to make five, you know, spending$4 to make five or whatever it happens to be. Not selling$100 bills for 50 bucks. Yeah. I think three years ago, me would have said methodical. I think me now says I would back cursor and lovable compared to the other. Okay. That space. Yeah. There you go. so this could also be as bill gurley always explains hey what's the game on the field right now yeah so in the nba in the 90s uh defense won championships and antony seven seconds or less people steph curry shooting more threes the knicks coach the new knicks coach mike brown um tibbs who preceded him and got us to the eastern conference finals and the reward for that was getting fired was one of these defensive stalwarts.

50:54No offensive strategy. You know, very little offensive strategy. Mike Brown said, we're going to shoot 43s a game. So kind of getting to that Celtics level, threes per game, which I'm not crazy about watching, but I do want to win a chip. Very boring to watch. It can be like a little Chuck fest or whatever. And I say that as a Warriors fan, to be clear. Like, I know why we're doing this. Yeah, but the Warriors, you know, they just had such a beautiful style of play that I kind of forgive it because it wasn't as Chucky and you had Clay Thompson, the Splash Brothers and Steph who were just, oof, because they could also drive the lane.

51:25Anyway, putting it all aside, you had Draymond there, the Draymond show. That's, I think, what you have to think about. So it might be that Brian's right right now and our friend at first round, Liz, and Nukunj Kothari are right last year and five years from now. So, you know, both things can be true. Nick Coons, by the way, if people aren't familiar with him, was a previous investor at Kostla before spreading off onto his own fund. And Jason, just because I know you love to see charts about Uber, here is one that I just put together for you. And this right here is the break-even free cash flow mark.

52:09And you can see that it started off their life as a public company way down in the red and have since been consistently going on. So this is the end of the J curve when you do begin to rip cash out of the business. And oh, Lord, isn't that a beautiful chart? Yeah. And it gives you the it gives you such optionality that I was talking to Dara at the All-In Summit. And I said, hey, listen, you got you're doing a 20 billion dollar buyback. Why don't you put the 20 billion dollars into buying a car company or, you know, I don't know, Wave or Neuro or, you know, some self-driving company by Volkswagen?

52:43you know, you could buy Volkswagen for, you know, some amount of money. You could buy another brand and then just build your own cars and build a full stack like Elon's doing. And it's like, we can do both. We can chew gum and walk. It's not, it's not an, or it's an end. And I was like, Ooh, Ooh, I love that answer. I love that answer. And you know, he can, he can decide to change his mind if he thinks the stock's worth, you know, most people think the stock's worth 110 right now or 120 top analysts think 140 150 the low analysts think you know 85 to 95 where it sits now um putting it all aside uh he can just decide okay i in my mind i think it's 120 stock i'm buying under that number and when it gets above 120 i'm selling and it's very interesting like warren buffett had and charlie munger had a ratio in which i think i saw this discussion on in downtown Josh Brown's compound podcast.

53:41They would buy it if it was like a.3 or something, some ratio, and it never got there because they publicly said, we're buying our own shares at this number. So I put in a floor where - Right, because - Self-fulfilling prophecy. That's like a stable coin, like bounding, like people trying to keep it at the - So, I mean, literally one of the great things that Dara could say is anything under$115, we're buying the shares. Yeah. You know, it's going to go right to 115 or 110. Does that count as, can you do that? Is that legal? Isn't there something around? Okay. All right. Yeah. I mean, listen, it's just, we have a share buyback program.

54:20Here's where it gets triggered at the ratio of our earnings, you know, our price earnings ratio of our price earnings ratio is under 25 forward looking. We're a buyer when we're above 25 price earning ratio. We're a holder. That's it. Yeah. By the way, just for people who are curious, Uber's worth nearly four times as much as Volkswagen, in case anyone wanted to know the ratio there. Yeah. But I think if you buy Volkswagen, you're buying a challenged business, and it could be distracting, as opposed to having 20 partners putting their cars into the network on a global basis. What else is on the docket?

54:53Next up is Anthropic introducing something called skills. Now, usually, Jason, when an AI company drops an update or new technology, We have to read academic papers. We have to look at charts. It's all a little bit technical, a little bit jargony, and a little bit far away. In this case, what Anthropic has built is deceptively simple. So a skill is essentially a little place where you can put information for your AI model to know. So if you're a company that has a certain style guide, you can put that into a skill. If there's a bit of data you want to have, you can put that into a skill. They're very lightweight, almost like signposts for AI models to be better tuned for you.

55:29and what Simon Wilson said about this is that they're very, very token efficient. So if you want to steer your AI model one way or the other and you don't want to have to consume tons and tons of tokens to do so, this is a great way to do it. Now, we don't cover every single update from the open AIs and the XAIs, but to me this was pretty bullish because it seems like a very simple way to possibly make applied AI in the enterprise quite a lot better. And I've been thinking about the split between Anthropic's growth, which is mostly predicated on its API, which is business sales, and OpenAI's own, which is mostly through ChatGPT, which is much more consumer-weighted.

56:11And as we look at these two companies, Anthropic is currently at a roughly$7 billion run rate now, expecting to get to about$9 billion by the end of the year. And today, OpenAI is about$13. And so it seems that Anthropic is somewhat quietly through its improvement to its product and making things better for companies, catching up to OpenAI's historic growth. And so I think skills is a good example of how they're thinking outside of the box to try to find new ways to make this stuff work better. And I expect us all to use this sort of stuff in the next year. Yeah, so this is an interesting feature.

56:45It's kind of like memory, if you remember that feature in ChatGPT 3.5, but much more advanced. So you might want to say for every time I do a research report, please give me, create a Google sheet with the data in it, share it with Lon and Oliver and do this and email me a copy. So I'm reminded and put it in my notion table to do this, whatever it is. And then you don't have to remember to put all that stuff in. Or I work at a venture capital firm that has the podcast This Week in Startups. We do three episodes a week. Please make sure you understand all the topics we discussed on the show. Yada, yada, yada, yada, yada.

57:32You get the idea. So that's, I think, an interesting way to look at this, which is saving time and saving money because it will cost less tokens. Not that that matters, because I feel like all these products, you get more than enough tokens for business uses. So yeah, well done to Anthropic. I think it's more of an API question. I think if you're pulling Claude from their APIs, then I think the token is just a little bit more important. But I'm just impressed that we're still seeing this kind of gains, but there was a funny Hacker News comment. I love to read the Hacker News conversation about these things, Jason.

58:09And someone said, quote, I'm kind of in stitches over this. Claude's skills are dependent upon developers writing competent documentation and keeping it up to date, which most seemingly can't even do for actual code they write, never mind a brute force black box like an LLM. So perhaps we're going to introduce new and novel ways to have tech debt. Yeah, if you think about this product, you know, it would be good if we had like some solid examples from Anthropic of it working, but because we're kind of like coming up with the ideas here, but you might have, like you said, a style or you said, like, I think a style guide for writers.

58:45You might also have like a design guide, design language. So, okay. You know, we're launched. We have this design and this we can start to have this design language here are all those files we have um a corporate uh policy you have a corporate culture here are those documents and then anytime we do something in the world like so you're writing a press release make sure the style guide the design guide and the design language and our culture document analyze it and that we loop in the heads of each of those departments and just let them know what's going on and if there are any really it's just you start thinking about it it's kind of like building an operations team into the product and then they're always vigilant it's like having this vigilant group of centuries is that the term for guards that like adore centuries yeah it's like having these centuries in each part of the kingdom the castles and they're saying, hey, yeah, here's the standards for accounting.

59:49If you submit something, you can take clients, think about expense policy. You can take clients to lunch at up to$50 a person, but you can't take them to dinner unless they're already a customer who spends over a million dollars a year and then you can take them up to twice a year, up to$500 a person and then when some expense software runs You have it run against the expense policy methodically. So are skills to AI models a bit like to-do lists for humans? Or policy lists, like employee handbooks? I guess it could be both. I guess is the way to look at it. It could be both. But all of this is to say collaboration inside of LLMs is still lacking.

1:00:38They're still building out the framework of how, when we're like, Notion has all this kind of dialed in and, you know, it's a collaborative by, you know, from the bottom up. But LLMs were not built to be collaborative from the bottom up. And I was talking to the CEO of Grammarly and Coda, which bought Superhuman. And they're going to be on the program. Yeah. And he's going to be on the program again to sort of talk about some of their new features and architecture. but what's super interesting about the design of these products sorry I lost my train of thought but anyway I was talking to him about this sort of whole concept here oh it was about multiplayer mode Grammarly was always about making you a better writer and I kept saying to him like and I've been saying this for two or three years and I pay for Grammarly for everybody on the team I'm like it's great that Grammarly follows us around and we have the style guy and you can be professional, you can be good science and everybody's a better writer and less embarrassing stuff.

1:01:35But really what I want to do is be working on documents together and I want the Grammarly to become the central hub of all knowledge because essentially it's like a keyboard logger. So if Grammarly, you know, is watching you write the document, somebody edit the document, another person putting the charts in the document, me doing the strikethrough or asking questions in it, we would have this knowledge base based on our keystrokes that Grammarly could go from being, since it's following all of us all over the place, it could give us some grand view of the work happening in real time. Imagine if I could pull up Grammarly Pulse.

1:02:15I'm just making something up here. But it would tell me like, hey, by the way, people are working hard on the docket. They've got these four stories done and these two are worked on. You might want to consider that there was just breaking news about those six stories you know story number four there's breaking news and a fact's wrong it could be running those what we used to call cron jobs or running those agents against what's happening at the keyboard level which is super creepy in some ways well it's a it's a work computer i mean i if you told me to put it on my personal computer i'd tell you to go f off but like i mean on a work machine yeah we we installed some software on our work machines i get into details about that and for the first like two weeks it was a little freaky yeah because because i mean it does feel like i've got a miniature jason on my shoulder asking me if that's two minutes of youtube during lunch is actually necessary and the answer is yes but then i realized that i don't care it's fine i don't mind i'm busy yeah here's here's where it's become super helpful um and the this software i won't give a shout out to them anyway um they're on sponsors or whatever.

1:03:19And I may switch the software. So I don't want to create any hard feelings here. We just tried it for a year, but you have to lock your computers down. So I told everybody, keep your old computers. If they're under 500 bucks, I just gifted them to the employees. So we're going to get everybody new computers faster, which is good because if you have this kind of tracking software, it's really designed for finance companies, compliance-based businesses, let's say doctors with HIPAA, don't use your work computer for anything that's not appropriate what it also does is it obviously is tracking what you're doing and you can do like crazy granular stuff like if you were using call center people you could track down to the key stroke we don't do that it actually bogs the computer down but you could um what happens ultimately is the people who are using it know how much they're working and what they spend their time in and then as an organization it rolls up to my management team i don't have the time to look at this obviously but it was a very interesting thing when we installed it it was to lock everything down because we have all these global partnerships i've been talking about we need to lock this down because we also have every company's documents get sent to us like you know what if somebody's cap table leaked like that does happen it's a really bad look if a venture firm is responsible that finance companies have to have a higher level of duty put that aside now you understand the organization is spending 42 of their time in slack 30 of their time in notion 6 of their time in grammar whatever it is you know superhuman what they're spending their time on you can measure it you can manage it what also came out of that was i was being told these three people are the hardest working people in the company we need to promote them and give raises.

1:05:03And those three people were not even in the top half. And the people in the top three, I knew one of them and the other two, I didn't know they were the all-stars who were really making solid contributions. So it made me rethink compensation and it made me rethink bonuses. So I did something the last two years in our bonus structure, and it's a bonus. It doesn't get paid if I don't want to do it. It's just my money. Sometimes I'll just pop a little bonus on people. And I said, you know what? Bonuses are no longer a percentage of your salary. That makes no sense to me. It should be based on you know, when you work at a company, like, oh, you get a 2 % bonus, you get a 4 % bonus, cost of living increase, all that stuff.

1:05:43It's kind of like... I've lived and died by those numbers, yeah. I feel like it's so antiquated and antithetical to a high performance culture. So, you know, if you're running a basketball team, if somebody has the hot hand and they win you the game and they play better defense last night, you're going to give them more minutes on Friday night. you're not going to give them more minutes next year. Like, what's that about? So I said, okay, every month, take the bonus pool, the anticipated bonus pool for the year. And every month, I want you to give me the three hardest working by effort, which the software can tell you, which is the people who just put in the raw number of hours.

1:06:23And you can kind of check it. There's no gamesmanship of it, to be totally honest. But just amongst the top third of the company, who are the top three hardest working people are. And then based on what they did, their impact. So that would recognize somebody who just works an average number of hours or even a below average number of hours, conceivable. But they did something that really moved the needle forward. So you're recognizing the two things that actually make a difference in careers and businesses, effort and effectiveness. And people want to argue, well, I'm getting paid to do a certain output.

1:06:57If I do it in 10 hours, it should be the same if I do it in 100. And it's like that argument you hear when people basically F off at work and they're, you know, working an hour a day. They're like, well, I hit my numbers. That's all that matters. You're not paying me for 40 hours a week. It's like, well, we kind of are. It's in the documents. Literally, I am literally paying you for 40 hours a week. You want a piece job. That's fine. But you don't get benefits. Which is why the work from home thing is being unraveled and returned to office happens because there were a group of people who felt that way.

1:07:24And then they would then say, well, then I'll have three jobs because if you go down that I don't know if I can get that logic as long as I give them what they wanted as output that matters. And man has, and then I just do it monthly. Ah, so it's so everyone can get rewarded in weeks versus quarters. That's that's good. Literally we started doing, I think in August or September and the people who got picked were like, thank you. This is incredible. And I got a, you know, a couple of grand instead of waiting to the end of the year to get five or 10 grand or whatever it is like pretty dope. And then it sends the message I want for the culture of the company, which is give it your full effort.

1:07:59Make sure you're effective. And it's not just busy work. And I leave it up to you to decide if you want to do that. But we don't have to punch a clock. And if you're doing busy work, who gives a shit? I had people who were like, instead of automating stuff, just doing stuff manually. So you have somebody who's incredibly expensive in the company doing customer support emails. And then there's people who make half as much and they're like doing this incredible stuff. And I'm like, okay, that's my job as the manager then to figure that disparity out. Now, Jason, speaking about doing half as much, I know you're carrying about half the weight that you used to.

1:08:37Oh, yes, that's true. And you know, it's all because of Roe.co and my embracing GLP life. Look at you, Alex, with the perfect segue. Listen, I'm Svelte. I just got off the scale today. That's 172. Nice. The amount of joy when I hit that scale, it's the under 175, I feel. It's just amazing. And you know what? The food noise went away, which means I'm crisper and better on the pod. I have better arguments. I'm more in the zone. When you're fat, and I was a fat bastard. Now, Roe.co gives me the ability to say whatever I want, within reason. I became a fat bastard. That's Brooklyn for obese. And it was very depressing for me.

1:09:28I don't get depressed about much. When ski season ends, when the Knicks season ends, or when my weight goes up. Those are the three things that make me sad. And what a great couple of years it's been for me because I lost 40 pounds thanks to GLPs. Go to row.co, and if you got insurance, you don't got insurance, they're going to help you either way. It's incredibly affordable. I didn't talk about it for the first three years, but just last year, I was talking to the CEO of row.co. We were at an event together, playing a little cards, talking, chewing the fat. He said, you look great. How'd you do it?

1:10:03And I said, GLP is here. Oh, you know we're doing GLP. I said, that's great. I have just started speaking about it publicly. I said, hey, I got an idea for you. Serena Williams and Charles Barkley are becoming spokespersons for the GLP movement. I said, that's great. Congratulations. What a great idea. He said, would you be a spokesperson with us? And I said, you know what? I've never done the spokesperson thing except for Athena, athenawow.com to get yourself an assistant and get a great discount on it. It's a crossover here. Sure. And I talked to my wife about it and I said, hey, I think this would be like a noble thing for me to do.

1:10:37Just chit chat about my weight loss and be honest about it. I sleep better. I ski better. I just feel better about myself. So if you are sad, embarrassed about being a fat bastard, or you just want to lose the extra 10, I'm not telling you this is going to work or not. But I am telling you it worked for me. I am telling you that Road.co is a place you can explore it with a great team. I trust them. There you go. Road.co. That is the first time I think someone has ever successfully nested a promo inside of a promo. Ladies and gentlemen, Promoception. Well, you know, it's a good thing, too, because we just had to run payroll on gusto.com slash Jason.

1:11:22And I'm so focused because of my Athena assistants taking all this stuff off my to-do list. And I got such great sleep last night because I'm not snoring and I don't need a C-cap machine because I'm so fat. All of these things work in a positive cycle going forward. Yeah, this is the Jason stack, if you will. This is how to go full Jason. Speaking of the stack, I should probably do something for Whoop. I stopped wearing my Apple Watch because I got into Whoop. And then I've been doing my labs, like function health and superpowers. Whoop now has labs in it. And then there's another place over here, a local place, that everybody goes to to get labs.

1:11:58So that's the next thing I'm going to start talking about is labs, blood work, self-directed healthcare. The thing we're going to talk about next on the show, however, is what startups should build in the realm of AI. I have a question for you based on a little bit of data. So first of all, Jason, do you recall we had ramp on the show just the other day? Yeah. Mr. Eric Gleiman on. He was great. Oh, one of the best guests we've, um, going to have him back on for sure. Okay. There, just found it. Sorry. I've got three screens and I lost a window. Yeah. So Jason, uh, I found this really interesting bit of data from the ramp team and what this chart shows is different retention levels.

1:12:36Oh God, here we go. Content marketing. They get free ads all the time. This is the last one. Go ahead.

1:12:45You know, this content marketing strategy, it just kills me. It's like Carta's trying to do this all the time. They're like, we don't want to buy ads on the program, but we do want you to talk about all this Carta data. I'm just like, enough already. No more Carta data. Enough. Okay, well. But we'll sneak this one in. I'm going to just do one more of these. Go ahead. All right. This shows differing retention levels for AI products in the enterprise. Oh, this is interesting. Thank you. Back in 2022, retention was roughly 50 % by the end of the year. That improved to 63 % in 2023, and then 2024 up to about 80 % by the end of the year.

1:13:25And then this year, it's tracking even better than that. So we could see much less churn in the enterprise. And I think Ramp is big enough in the startup realm that they have enough companies using them to make this be reasonably directional data. I'm not trying to say that it's perfect or that you should all go use Ramp. But this, to me, really spoke to an improvement in the ability for startups building enterprise AI products to actually retain customers long term. This contrasts with some information we've seen from Deutsche Bank about how ChatGPT, the leading consumer AI product in the world, has seen its growth rate essentially flatline in Europe.

1:14:06And so I'm curious from your perspective, talking to founders as you do, if you recommend that they're going to be building in their own AI, they should focus more on the consumer side of things or more on the enterprise side of things. Six months ago, I would have said consumer, but now I'm not so sure. There's going to be a market for both. what you're seeing here is two phenomenons. One, there's less sampling going on and more product market fit happening, stronger product market fit. So when new products come out, people get promiscuous. When you first got your iPhone, you went into the app store when the app store launched, I think in the second or third version, and people were frisky that, oh, buy this for a dollar, buy that for$2.

1:14:46But then things became subscriptions, things became$60 a year, and you kind of lock into, hey, I really like Financial Times, The Economist, Wall Street Journal, whatever your bag is. I really like Calm.com. Oh, I use FitBod for my fitness stuff, whatever it happens to be. Or I got a subscription to Whoop. I got a subscription to Function Health. I got a subscription to Fitbit. You kind of lock into your Peloton, whatever it is. So that's pretty natural over time. And that's what you see in the retention chart. Now, in terms of Europe, that's a weird place. so there could be cultural things going on here but putting it aside there's also a natural audience for paid products and when you hit it it can be very hard to go above that natural audience you have to do um new things to expand the audience netflix perfect example netflix added games remember they put like a strip of games on there then they started doing live and they started doing like boxing matches ufc matches they're doing all kinds of interesting live stuff and then yesterday the ringer did a deal to bring the ringer podcasts which are really shows that are video they're just shows that are like the video like here is like maybe half of broadcast quality or 80 percent of broadcast quality but nobody cares because it's about the content and the freshness of it and the uniqueness of it and the depth of the experience of the host so when you hit your natural lights you try to do new things and sometimes they work sometimes they don't.

1:16:13And so that's what we're generally seeing here in these, in this data. And I've just made a new policy. Here's the new policy. New policy. We're going to use all this great data. We're going to use ramps. We're going to use corridors. We're going to use Zillow. But going forward, we're going to clip out who the data is from. And then we're just going to say, a leading SaaS company released some data. Here it is. Because I know what they're all doing. They're all hacking advertising. It's infuriating. It's enough already. Let me steelman that for you in the other direction. One of the hardest things to get in the private markets is information about what's going on.

1:16:59And companies that I think take the time to hire economists, as companies like Ramp and Carta have, and do actual rigorous analysis on a large enough data set i think can at times put out useful data and we've we've talked about carla before and you'll note that we haven't had car to date on the show in a while because you were opposed to it but i do think occasionally there is a chart that shows something so interesting even with a slightly biased source because i know they're they want all right i'll cut the difference okay let's put the difference with you you're a mensch I'm a business guy.

1:17:32I get it. You know, I was a journalist now. Like, I'm more like the publisher, so you're like the editor, the writer. I'm like the publisher. It's like, rah, rah, shi-hi, get my cigar. We need to make money and monetize this. Here's what we're going to do. Okay. We're going to take the logos off. If it's absolutely great, I might, might say, reveal who gave us the data. We're gonna hack them back. I'm hacking. It's a hack back. You do know they're gonna text me, not you, right? Absolutely. And they should. They should text Alex at 415-555-1212. Please. All right, everybody. End the show. Come on.

1:18:16Wrap it up here.