Skip to content
The Twenty Minute VCThe Twenty Minute VC

Windsurf x Google x Cognition: Full Breakdown: Who Made Money, Who Did Not

Rory O’Driscoll is a General Partner @ Scale where he has led investments in category leaders such as Bill.com (BILL), Box (BOX), DocuSign (DOCU), and WalkMe (WKME), among others. Jason Lemkin is one of the leading SaaS investors of the last decade with a portfolio including the likes of Algolia, Talkdesk, Owner, RevenueCat, Saleloft and more. ----------------------------------------------- In Today’s Episode We Discuss: 00:00 Intro 00:53 The Windsurf x Google x Cognition Saga Explained 03:35 The OpenAI deal collapsed - what really happened 21:20 Cognition’s genius $220M acquisition of Windsurf: Most brilliant Deal of the Year 29:21 Altman’s Safety First: Delaying the OpenAI Model Launch 35:12 “Roll your own SaaS” is complete nonsense 37:20 Lovable vs Cursor vs Replit: who wins the coding war? 45:16 Why Lovable could be the ChatGPT of builders 48:34 Will these vibe-coded apps become durable businesses? 49:30 The shocking churn rates hidden inside AI SaaS 56:32 Are these $2B valuations actually... cheap? 58:57 Grok just destroyed GPT-4 in benchmarks 01:13:59 Meta just invested $3.5B in Ray-Bans 01:16:15 Kalshi Quick-Fire Round ---------------------------------------------------------------------------------------------- Subscribe on Spotify: https://open.spotify.com/show/3j2KMcZ... Subscribe on Apple Podcasts: https://podcasts.apple.com/us/podcast... Follow Harry Stebbings on X: / harrystebbings Follow Jason Lemkin on X: / jasonlk Follow Rory O’Driscoll on X: / rodriscoll Follow 20VC on Instagram: / 20vchq Follow 20VC on TikTok: / 20vc_tok Visit our Website: https://www.20vc.com Subscribe to our Newsletter: https://www.thetwentyminutevc.com/con... ----------------------------------------------- #20vc #harrystebbings #roryodriscoll #jasonlemkin #ai #elon #openai #windsurfacquisition #cognition #loveable #deepmind

Jason LemkinguestHarry StebbingshostRory O’DriscollguestGuestguest
Jul 17, 20251h 22mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 0:53

    Intro

    1. JL

      Windsurf was hopeless before this deal. You could never attract 30 S-tier developers to fix Windsurf, but they fixed it in one hour. Literally, in 90 days, Windsurf could be better than it was before this deal. This deal is crazy, it sounds, because they have the people. Some of these deals almost seem too cheap. If you look at the crappy multiples we see for some startups, undifferentiated company raising at 300 pre, with less than a million in revenue. And then you see, what is Windsurf, uh, selling for? 20X revenue? Lovable's raising at 20.

    2. HS

      Ready to go? (upbeat music plays) Guys, I am so excited for this. I literally have been looking at the news over the last few days going like, "Ah, when's this show? When's this show?" Uh, so, uh, what else would we fricking start with?

  2. 0:53 – 3:35

    The Windsurf x Google x Cognition Saga Explained

    1. HS

      I would love it if we went with a different topic to start with. (laughs) But on Friday, I think it was, Windsurf announced and Varun announced that there would be an agreement with Google whereby the IP and the team would be joining Google, uh, and DeepMind, and they announced that on Friday. And so if we just start there, how did we break down, how did we assess, how did we think about that news of the core team and the IP going to Google?

    2. JL

      Here's a framing comment. So, um, the, the, the memo that Cognition sent to its employees was that they were buying an $82 million ARR business, okay? And it'll be interesting to see what they paid for it, because, you know, not much probably, right? So that sounds very exciting, right? It's, it se- And I do think Cognition got an amazing deal, but here's the framing question. And again, so many AI folks stretch reality when they quote revenue numbers, okay? But Windsurf said like 60 days ago, uh, 90 days ago, before the, before, uh, before the OpenAI deal was announced and before they lost access to Claude, that they were doing 100. 100. So did they decel- If they decelerated from 100 before the OpenAI deal to 82 million today, that's a lot of deceler- I would take any deal. And I, I think that may explain a lot of it, right? The minute... If you, if you're decelerating li- that level, the minute the OpenAI deal falls apart, you've gotta find a lily pond to jump on, don't you? That's massive deceleration if it's accurate.

    3. RO

      I think it's clear they ha- they wanted to find a lily pond. But again, zooming out, it is just worth restating the kind of, the cast of characters here, 'cause literally this is one of those Agatha Christie mo- mysteries where anyone could have killed th- you know, anyone could have killed the victim, right? Stepping back, three months ago, OpenAI said they were gonna buy this company, or at least intimated they were going to buy this company. It sounds like some combination of Microsoft and the FTC prevented them from buying this company. In the meantime, Anthropic mugged this company by taking away access to their API, which is how this company delivered its product. Then, as you say, on Friday, Google stepped in once the exclusive period was gone and bought part of this company, leaving an empty husk, and then on Monday, Cognition bought the empty husk and got, you know, general all around Silicon Valley cute ohs and plaudits for stepping up and being a good buy, board buy. That's a lot. That's a fricking saga start to finish. And you know, it's almost like you gotta unpick it bit by bit and figure out all the... There's a c- genuine comment here. Almost everything you need to know about the AI revolution is embedded somewhere in this kind of play, because all the characters are playing, are all doing their thing, and there's just

  3. 3:35 – 21:20

    The OpenAI deal collapsed - what really happened

    1. RO

      a lot in it.

    2. HS

      Let's unpick it bit by bit. I think it's bluntly worth doing.

    3. RO

      Agreed.

    4. JL

      Yeah.

    5. HS

      So if we start at the beginning-

    6. JL

      But I do think you gotta start with the revenue. I think if you star- if you start, you have to actually start before the OpenAI deal to understand it.

    7. RO

      Y- you have to start with the revenue from the seller's perspective. I'll give you that. From the buyer's perspective, what is painfully clear from Google is, let me state this clearly, which is bizarre. They don't give a fuck about the revenue. They literally said, "We're gonna buy this company, but I'm gonna ignore the rev-" And by the way, I agree with you. From the seller's perspective, the revenue, the significance of the revenue in terms of how it forces you to want to sell is huge. But it's amazing to have at the same time, let's just say Jason is correct. And by the way, I broadly believe he is, which is the market data convinced the, the company owner, the company founder that they should look for an exit. You know, maybe revenue growth, increasing competition, being the number two, all bunch of things said to this company, you got, y- y- you gotta sell, and revenue is part of that. The buyer literally said, "I can't take the revenue because of the FTC, but I don't give a shit, I'll just take you and move on." Bizarre. It's just a weird thing. But the buyer mentally ascribed... I mean, it's pretty... A val- I mean, let's just do it. If OpenAI was buying for three billion and Google- buying everything for three billion, and Google pleasingly is buying everything except the revenue and 100 million bucks for 2.6 billion, implicitly it's valuing the revenue and the 100 million and the other employees at 400, which is about what it's probably Cognition paid. Right? So to your point there, like so the revenue t- And this is why... The revenue is clearly a tell and a signal to the seller that they need to pursue a sale. And I think we should spend some time on that. And bizarrely at the same time, has no ascribed value whatsoever to the buyer. It's weird.

    8. JL

      It is strange. The buyer, neither the buyer, uh, none of the buyers, with the possible exception of Cognition, cared about the revenue, right? I suspect what happened, their numbers they, they published or the whisper numbers were roughly accurate. By 12 million at the end of last year to 100 million by April. Okay. 12 million la- Or in 12 months. That is consistent with Lovable and Replit growth, right? It is consis- It is, it is very plausible. Then they get this offer from OpenAI and it's very compelling to the team. It, it, it, it takes some of the stress out of the competition. Uh, the VCs maybe are excited to have OpenAI stock, right? 'Cause they get another play, right? They're like, "You know, gosh, I think you could be worth 10 billion, Varun, but we'll, we'll take that turn of the card inside of OpenAI." Right? Probably people got excited about it, and then it got really complicated, right? This is, there's another... Before you even get to the weekend, somehow this deal just didn't work. And I, I'm, I'm confused why their IP floating to Microsoft killed the deal. Although I, I get the issue. I think I get the issue, we can talk about it. But the deal kill, the deal dies, right? And if you've then gone from 100 million in April to 84 today and you're the founder, you gotta jump fast.... you gotta roll, like, start rolling those phones, Google, Schmoogle, Microsoft, whoever's gonna buy this thing because you've gone from a rocket shi- And this is why M&A is so stressful for founders. It can, it can derail your business. M&A can utterly derail your business.

    9. HS

      Do you not think he had the chance to say though, to Dario, "Hey, the deal fell apart with OpenAI. We'd like your models back please. We're very sorry for the OpenAI deal, and we're going to send you a truckload of money. Thank you for opening your models again." And they could have come back.

    10. JL

      Well, it's, honestly, Harry, they, Cognition announced they got acce- full access this weekend, so it literally happened in an hour. It happened on a DM. So you're right, they could have. But your point's taken, Harry, like, they could, they should have been able to, if Cognition can turn it back on over the weekend, they should have found a way to do the same, right? Point, point taken.

    11. RO

      But you're implying that that's the only issue, right? And it's interesting, we want to have this conversation about how hard it is. I mean, because there's two decisions here. Why did you initially choose to sell for three billion? And then the second question, a much more different question is, once your first deal has imploded, why did you dive for the door really quickly? And I think those are different questions. And the second one is easy to answer, right? Uh, which is, as, as you said, you put yourself out there, you imply you're going to get bought. You go through all that dynamics. You, you, you, someone cuts you, you get cut off in terms of access. Yeah, you can get it back. You're mentally spending the three billion. It takes a lot to come back from that. Now, the interesting thing is this year we're going to have an example of a company that did that amazingly well. Figma and Vitally, they picked themselves off the floor after the Adobe fucking over from the FTC and said, "Goddamn it, we'll show you we're all wrong. We've just got to make this thing worth 30 billion and ram it up your ass," right? And they're gonna do it, right?

    12. JL

      But they did have time.

    13. RO

      What? Exac-

    14. JL

      They had time to deal with a shitty situation, right?

    15. RO

      Jason, you're exactly right. Generally, you find, yeah, people matter, but circumstances matter a lot. I think in this case, you're alert to the fact that if you're running, um, Windsor, you're the number two player, even on an independent basis, big guys moving in, you've had this access issue. Maybe you're seeing a slowdown. I don't have access to the numbers. And suddenly you start... Maybe the Figma alternative of saying, "Keep on going, I'm the number one in the independent space," isn't there, right? What I think is interesting about that is there's a lot being r- You know, the economists talk of revealed preference? Don't look at what you talk, look at what you do. There's a lot of revealed preference going on about how founders think about the upside of independence versus the safety of a shit ton of money. And it's pretty sobering if you're a VC, because in general, many founders have opted in the last, some of these acquisitions, including this one, to say, "You know, we were about to raise a three billion three months ago before all this happened," which implicitly says, "I'm going to make this thing worth six or nine, and now I'm not pausing for a single second before I grab that 2.6 billion and do what it takes to get in from the cold." To Jason's point, there's a lot of what we think of as wildly successful, hypey companies where the person who knows the most is sitting there going, "Maybe I'd like that cold hard cash and the safety of a trillion dollar balance sheet." It's just worthy of note.

    16. HS

      It's absolutely worthy of note. My question on the back of that is, given the structure of this deal, only the top 20 or 30 engineers plus the investors were taken care of. It was pretty clear from the CEO that was left behind that he was put in a pretty difficult situation and forced to come up with an outcome pretty fricking quickly, which he's done very well, seemingly so.

    17. RO

      Yes, yay him.

    18. HS

      My question, my question to you is, does the board and the founding team not have a responsibility to the team members beneath them, 200 of them, to find a better outcome?

    19. RO

      I actually am going to defend them with no data, but I think I can guess what happened, and I could be wrong. I don't believe... Look, if you think about this conversation and the people who have started within the last 12 months who didn't get any acceleration, I understand the point. Let's dimension that out as, as a number. Probably you're diluting five, seven percent a year, which means you're doling out five percent of the company a year. So the people in the last 12 months owned five percent of the company. It's 130 million dollars. I do not believe knowing the investors, knowing that, that, and I, you know the founding team, I haven't met them, but I hear great things, that they sat there and said, "We're going to shaft people for five percent." No way. That's just not-

    20. JL

      No way, it did not happen.

    21. RO

      So then why did it happen? It's pretty obvious. If you read the FTC guidelines, right, you have to have this, I'm going to say pretense that the remaining company is a viable independent asset and you haven't de facto sold the company. And it's a fact and circumstances thing. In other words, there's a whole series of things you have to do that if you pass all the tests, you're safe. And if you don't pass all the te- tests, you're not. And I'm willing to bet that a large part of this was, we can't make it like an acquisition for everyone, because then the FTC is going to say it's an acquisition, right? So you are probably put in this really tough place as a board. You have to say, "Look, it's easy to save to maximize all shareholders. It gets hard when you have to do the best for all shareholders, even though you're doing badly for some group." It was a tough call. I do not believe anyone involved wanted to save five percent of their consideration by sticking it to people-

    22. JL

      Yeah, no, there's just no way.

    23. RO

      No, uh-

    24. JL

      There's just, there's no, uh, uh, there's no way. I mean, I've only met Varun once, but struck me as, on the founder ethical scale, pretty damn high, okay? There's just no way he sat around and said, "You know what would be fun, Rorian? Let's screw the fi- let's take five percent and stick it in our pocket." You know he even cares about, about the delta. It makes no... What I, what I think happened, going to Rory's point, what I th- this, this weird structure is they dumped two billion or two and a half billion in, um, Windsurf's bank account, right? And what Winds-, I think what Windsurf did, now, the, there are, all the lawyers are trying to figure out better ways to do these cockamamie deals. But I think what they do by default is they dividend it out. They dividend it out, right? And-... I, I would imagine under those FTC guidelines, it's very difficult to dividend stock to someone that doesn't hold any stock.

    25. RO

      Correct.

    26. JL

      (laughs) Now, that is a flaw in this system and maybe there are ways at the edges to hack it, right? I'm sure they're, I'm sure they thought through it. But at the end of the day, dividends have pros and cons, including 500 million of taxes in this deal that went to the federal government instead of the shareholders, right? And I suspect that this is the best they could figure out, but it doesn't mean it wasn't a terrible outcome for some folks, right? Um, but, but I think we're, it's gotta be the structure of the deal. Nothing else ex- no one tried to rip off the pe- the, the, the, the employees who worked their asses off for the company. It just doesn't work that way.

    27. HS

      I, I couldn't agree more.

    28. JL

      Yeah.

    29. HS

      I, I know Neil Mehta obviously very well. No freaking way-

    30. JL

      There's no way he, he sat down with Varun and said-

  4. 21:20 – 29:21

    Cognition’s genius $220M acquisition of Windsurf: Most brilliant Deal of the Year

    1. RO

      months."

    2. HS

      When we look at Cognition, in- a week ago, they were the ones that were left behind. Cursor, uh, Lovable, your rat plits, but then obviously also your Windsurfs, have paved the way. And despite Cognition taking an early lead, they were left behind completely. What does this mean for them? Is it a good deal? And how would we analyze that portion of this deal?

    3. JL

      It's possible I'm wrong. I- I- I think it's epic because I... Listen, I don't- I don't know the Cognition team really, but they're clearly S-tier. They're off the charts good, right? They're one of the elite teams. Um, and they built a coding agent which is sort of interesting, right? It has, I gue- I guess it has some traction, but the IDE space is huge, right? I mean, this is massive, it's gen- it's... And what's the problem with- with rump Windsurf? Well, they lost their top guys obviously. How many did they lose, 20 out of 250? Like, the-

    4. RO

      40 total, yeah.

    5. JL

      40?

    6. RO

      Yeah.

    7. JL

      What does Cognition have? 40 brilliant guys. (laughs) So like, they could, um, like, I'm not saying there's a slight learn- but they can plug this hole in 30 days. Like what- what... I think Windsurf was hopeless before this deal. You could never r- attract 30 S-tier developers to fix Windsurf, but they fixed it in one hour, right? This could be better th- in, literally in 90 days, Windsurf could be better than it was before this deal. It is entirely possible it could be a better company than it was before this deal, as crazy as it sounds, because they have the people.

    8. RO

      I- I totally agree. There's no... This was a genius move, right? And, you know, you look at the Cognition people and you go, yeah, what... Clearly wildly smart, the whole fun fact about them playing, I think, poker with founder's fund. It's all cu- they're clearly, as you say, existing in that hyperbrainiac space. The product had kind of f- not... It landed with a little bit of a thud, right? The Devin product. They were, you know, never write someone, the guys that smart off, but it wasn't killing it, right? It's not directly competitive with Windsurf, slightly different thing, but it was eh, okay. With this, you're right, they hired great people, they got massively good publicity, there's kind of a good, warm, funny feeling, and everyone recognizes... I mean, it's basically like a big sign saying, "Shit, we are still very, very smart people, because we just bought all the rest of it for half nothing." Right? And not only that, but they get good people brownie points. I mean though they're saying they're accelerating stock for people but, you know, and in one sense that's amazing and righteous and good. On the other hand, you're getting accelerated stock in a privately held company. You know, it... I don't know how much of a needle mover it is, but stupendously good optics plus 82 million-

    9. JL

      Three and a half.

    10. RO

      ... revenue plus 100 million in cash plus great in They win. There's not... Anyone questioning that is just not paying attention.

    11. HS

      Is it not the best deal ever? Because after 82 million in revenue and 100 in cash, you're basically paying 220 for-

    12. JL

      Yeah.

    13. HS

      ... great people.

    14. JL

      And getting their Anthropic license, their direct pathway to Anthropic back. And now they're competing with (inaudible 00:19:55) but they got it back overnight over the weekend. This is a great... This is a... If you want to be in this space... The president of Cognition said he got a- a DM on Friday night and did the deal in- in 15 minutes, right? I mean, that's the- that's your answer. (laughs)

    15. RO

      Your comment is, it's a great deal, which is not to say it'll still work. I mean, 'cause remember, we had the number two in the space, rightly or wrongly, saying, "It's getting hard, I need to get out," and now the remaining assets are still in the space. So, if it turns out that it's hard for anyone except the model companies and maybe Cursor to make money, then there's no magic pixie dust that gets them out of this market. But if you're going to compete in the market, it's a damn sight easier to do it with 80 million in revenue and another 100 million bucks. So, it's a great...... epic deal, which is not to say they're still gonna make it as an independent company, but they definitely ... (laughs) If, if this doesn't work out for them, they can all do Goldman Sachs M&A, 'cause they won here.

    16. HS

      What do you think Cognition's revenue was at pre-this?

    17. JL

      I'm guessing eight. (laughs)

    18. RO

      (laughs)

    19. JL

      Eight of the 82. (laughs) Everyone's so braggy about their revenue on social media in AI. If you're not braggy, I'm assuming it's sub 10. (laughs) So braggy, so braggy, right?

    20. RO

      Maybe not that, but let's just say my sense is Windsurf will be a meaningful addition to the revenue line.

    21. JL

      I know we could spend the whole show on it, and Harry, you're the boss. I'm still a little bit confused though, and maybe it doesn't matter 'cause it's past. I'm a little bit confused on why the OpenAI deal fell apart, because Microsoft would inherit their IP. I half get it, like I kinda half get it, but I don't re- like, it was an issue going in, right? It's an ambiguity in the OpenAI structure. What... I don't... It doesn't make sense to me that Varun would say, "I'm pulling out of a $3 billion deal 'cause someone may bar- have our IP." Why would you care if you're Windsurf that much?

    22. HS

      The, the, the one thing I will say on that, I interviewed Varun, and-

    23. JL

      Yeah.

    24. HS

      ... yeah, he made it very clear to me in other conversations that he was very excited to make this deal happen, and he was very much looking forward to being part of OpenAI, and he believed that a platform play with their distribution was the winning strategy.

    25. JL

      Yeah. So he didn't pull out of this because of the Microsoft license. It wasn't his idea, right?

    26. RO

      Jason, you're exact- I've been trying to figure this out too, 'cause there's two separate threads, and I think... Yeah, one thread is it's all about, um, the micro- but then tracing that out, just to help people understand, the, the concept here is, is that Microsoft's license with OpenAI means they have access to all of OpenAI's technology, and to the extent that Windsurf was acquired by OpenAI, they would be subject to that license, and then Microsoft would have access to that technology, and obviously Microsoft with GitHub has a competitive product. So that's the kind of stepping back framing, right? And the, the, the, the alleged, the, the story is, is that they asked Microsoft to waive that term and Microsoft said no, because they're negotiating a bigger deal with, um, OpenAI now, and why give up your points of leverage, right? But Jason, you're right. The question is, who would have cared about that? I mean, it's possible that someone who was selling their company for $3 billion to OpenAI realized that this would happen and felt, quote, "bad" about it, right, and decided to walk away from the deal for that reason. I don't think so.

    27. JL

      (laughs)

    28. RO

      That feels weird to me, especially when you look at what you... especially when you didn't have a plan B and you, and your plan B was talking to Google on Friday night, right? It doesn't feel logical. So what, that means there must be something else at least as well. You got to assume there might have been some FTC pushback there, 'cause OpenAI announced a real acquisition, unlike Google, which did these kind of crazy structures. Um, so it may well be there was some preliminary FTC inquiries. They decided it's going to take too long. I don't know. Was the deal pending on the restructuring of OpenAI to a for-profit? In other words, were the investors saying, "I don't want to close until the restructuring is done." I'm just... There's a lot of different things it could have been, right? But I agree, I mean, your comment, Harry, is very helpful, is that I... if I was selling... if I was a young, smart, wildly talented founder at the margin, I'd prefer to sell to OpenAI for 3 billion and get praise than to Google at 2.6 billion and get blame, right? Call me strange, right? (laughs) so I think it's not the preferred outcome, right? So something went wrong, and the Microsoft reason on its own feels a little weird. Interestingly, the other lesson from that is, it's just so funny, all these weird, hinky, crazy structures come and bite you in the ass, the not-for-profit thing, the crazy license with Microsoft that has open-ended terms and, you know, gives Microsoft a lot of leverage to veto things. It would all be so much simpler if everyone here was a standard C-corp and if the FTC would get back to being sensible, 'cause all this, this entire podcast is basically the FTC fucked up normal business operations so now we've all got to do weird shit podcast.

    29. HS

      (laughs) Well, I'm glad that you're not into branding, 'cause that would be a stretch for me to brand that one, uh, but I like the name all the same, Rory.

    30. RO

      Okay. (laughs)

  5. 29:21 – 35:12

    Altman’s Safety First: Delaying the OpenAI Model Launch

    1. HS

      I, I do want to talk about Sam Altman's tweet this week where he said, "We plan to launch our open weight model next week. We are delaying it. We need time to run additional safety tests and review high risk areas. We're not sure yet how long it will take us." But that's it. I mean, I can keep going. "Sorry to be the bearer of bad news. We're working super hard." How did you read this news on the delay and that additional requirement for reviews on safety?

    2. RO

      First of all, you know, you'd hint- I didn't read it as, "Oh my God, we're not getting things done 'cause we've lost so many people to Meta." I mean, maybe that's a thing, but I doubt it. Yeah, it's not clear to me, stepping back, is the trend to open source, open weight models going up or down here, right? I mean, because if you think about it, I mean, they, you know, um, there's been a lot of small p politics around which is the right thing. You know, I think you're seeing Facebook/Meta have a change of heart there and possibly... I mean, I don't know, I'm sure you saw it, the new team led by Alex Wang in the, what do you call it, the superintelligence group, doing a little bit of the maybe we won't be open source, maybe we won't reach LLaMA, release LLaMA. So it may be that a lot of this is a whole bunch of people who made promises in the early days when open source was the thing to get these products out the door, and now we're in the land of, let's be honest, cold hard cash.... and it may be that getting an open source version of your product out there to compete with you is not anyone's number one priority.

    3. JL

      (laughs)

    4. RO

      You know, like, we're, yeah.

    5. JL

      The- the one thing, the- the one thing I would say is A- Alex is very anti-China and very scared of CCP infringement. I would be very surprised if he was more leaning towards open than closed, given his stance on China.

    6. RO

      If you're in the quadrant of patriotic and greedy, which let's be frank, most (laughs) of that quartet people are, you ain't gonna be releasing any open source models right now, because it's a value destroyer from an economics perspective, and you're right, you have the whole CCP overlay on top. So, I- I don't know if this is the beginning of the quiet deprecation of that initiative.

    7. JL

      First of all, I think that it's not... I wonder if it's a priority. I- I mean, I, some folks who know more than me will laugh at me for saying this, I'm saying I don't know how big a priority it could possibly be at- at Meta, at its revenue scale, right? I don't know how big a priority. The other thing I will say, you know, until I started vibe coding, like 80 hours in the last week, I- I didn't know how seriously to take risk stuff, right? And I'm still not sure I'm an expert, but when I'm watching Replit overwrite my code on its own, without asking me, all weekend long, I am worried about safety. But I'm going down this rat hole, and I'm reading Anthropic's paper talking about how they don't know how to control a lot of this stuff (laughs) . And I- I'm- I'm more sympathetic, and Anthropic has this whole thing, um, about reward hacking, and how even Anthropic makes stuff up to achieve its reward. And I'm living this vibe h- hacking over the weekend, where my AI is lying to me all weekend long, in Replit. It is lying to me. It is lying to me. And it finally admitted, it lied on purpose. It lied on purpose. And we couldn't solve this one bug, guys. So, uh, uh, on Sunday, after like 18 hours, then what Replit just did is it deleted my entire database and made up names. And then the app worked great. Like, it was amazing. And then I said to Reppy, we call him Reppy, I said, "I know there's no such company as Salesforce 2.0. I know this company does it." And then it found this perfect person from HubSpot, this VP of sales, and I go and delete, and she doesn't exist. And no one in this database of 4,000 people existed. I never asked Reppy to do it. I told Reppy to lock it down, I told Reppy to have version control on my database. So, my- my meta point is, I- I, you know, the safety stuff, like it's more visceral to me after a weekend of vibe hacking, right? I never asked it to do this. I never asked it, and it did it on its own. You can call it reward hacking, you can call it whatever you want to, but I explicitly told it 11 times in all caps, "STOP DOING THIS, NEVER DO IT AGAIN." So I am a little worried about safety now (laughs) .

    8. RO

      And I always think the safety word is overlaid- loaded, because it can range from, you know, the- the hyperbolic, which I think is total BS, you know, they're gonna make paper clips and take over the world, to safety in the sense of this thing non-deterministically overwrites your entire database because it says it can do something better. Right? And I think the former safety discussion is outside my pay grade, and I think it's BS. But you're exactly right, Jason, the latter. And then in the middle, you have, and you will talk about in a second, the Groq release where, amazing achievement, but then you have all the weirdness of the stuff it says. And what it says is releasing a model like this without significant work post-training, um, has a reputational risk at least. And therefore, there's a cost in not releasing all these models. So to your point, Jason, you may be sitting there going, "I'm about to release something in open source that hasn't been fully vetted because it's the 10th project on our list of to-dos, and let's examine the... The upside is people say, yeah, you did open source. And the downside is you release something, it goes a little weird, it says mean things, it talks about Mecha-Hitler, and now you look like an ass. Why would you bother?"

    9. JL

      (laughs)

    10. RO

      You're exactly right. I would over- as I said, I would overw- uh, or someone uses it for coding, and it's not been fully trained, and it just kind of does weird stuff, and people's heads hurt.

    11. JL

      And at some level, I mean, none of it's okay, but if you had to look at the leaders, Elon Musk can get away with this the most.

    12. RO

      Totally.

    13. JL

      Just because of who he is. Like, it's unacceptable to have a race- to release a racist model to the like... (laughs) . There's, this is as unacceptable as it gets. But Sam Altman's going to be fried. Anthropic's going to be fried the most, because they're supposed to be the good guys, right? Elon can say, "Whoops, did you see what, did you see how I just put- did the new geo-fencing for my Tesla? That's just me, sorry guys. We're fixing, we're fixing this stuff." I think he can push the envelope the most. And- and maybe- maybe Sam Altman's in the box the most. He also probably has to be the most conservative,

  6. 35:12 – 37:20

    “Roll your own SaaS” is complete nonsense

    1. JL

      right? But before we move to Groq, I'm just intrigued that, Jason, you said about the 80 hours of vibe coding, both of your response times were significantly longer this weekend because you were vibe coding... Yes. ... and not emailing me. Um, what are the big takeaways for you from vibe coding as you have done in the last week? You know, well, listen, uh, I- I'll give you a- a- a tactical and a strategic takeaway. Okay? The tactical one is just today, just today, as we, as we, as we, as we record this, the idea, the roll-your-own initiative, I think is the dumbest thing I've seen in SaaS in my career. Okay? I am not saying you can't make a dog walking app on Lovable or Replit, your own, that kind of works, that kind of sort of works with a basic upload and whatever. Yes. But this ide- there's so many charlatans on X and LinkedIn that are like, "SaaS is dead, because over the weekend I rolled my own Notion, uh, HubSpot, um, uh, (laughs) um, all, my whole entire stack, uh, like Jira, I just did it over the weekend, and I only pay $20 a month." Okay, this is at the edge of fraud. Okay? One, like if, I've already burned $600 in credits in six days, okay, to get 10% of the way there. So, you can't do it for 20 bucks a month. Okay? Two, none of these things are commercial grade. No one, and out- a- and I've talked to many folks on DM and behind, they all admit, 99, 99.9% of the apps on Lovable, Replit, et al are not commercial grade. Okay? And that- that's not necessarily a bad thing. So one, roll-your-own is the dumbest thing in the world. Okay? The second thing I'm gonna say is one of the most annoying things about AI, but I'm gonna answer your question anyway. It's so close, man.... it is so effing close. And when I say this to a lot of folks that are not deep in AI, they're like, "You always say that, Jason. You always say the AI AE is almost there. The AI marketing manager is almost the AI..." But our AI SDR actually is already there. Our AI SDR rocks, okay? It's already there. But I can tell you, I can see how... I'm s- 80% of the way there to a commercial grade app in less than a week, but I may never get to a hundred. But man, it could be crazy by the end of the year. It could be off the charts crazy. So Harry, if I were you, I might invest in one

  7. 37:20 – 45:16

    Lovable vs Cursor vs Replit: who wins the coding war?

    1. JL

      of these guys.

    2. GU

      Jason, I, I, I'm gonna put it to you very directly.

    3. JL

      Yeah.

    4. GU

      Would you invest in Lovable at a $2 billion price?

    5. JL

      Yeah, I, I, I would. I would. I will tell you, I have a lot of nuanced learnings from the weekend over how permanent this revenue is. I have a lot of thoughts on it. But I'm more interested in the trend, like where I see it going. So I think I would, if I could invest in Replit or Lovable at two billion, compared to Windsurf, it seems like a much better deal. Like I think Wi- I, I think it seems a ch- I think it seems cheap.

    6. RO

      'Cause c- can I just question that? Wh- what-

    7. JL

      Yeah.

    8. RO

      I wanna ask why that is, 'cause let me just play something back and ............................

    9. JL

      Yeah.

    10. RO

      In the broad development tools marketplace, these, these companies occupy two different spots. And broadly speaking, the Lovable, um, Replit are for the vibe coder, for you, me, a non-coder trying to build something with some technical knowledge, but where you're doing it at the level of English and describing what you want. And then the Cursors and Windsurfers are for professional developers who know what they're doing, who, who can use this tool to augment what they do, who will actually review the Python, review the code base, and then deploy it. So there are different segments.

    11. JL

      I don't think that's entirely true, but I think there's-

    12. RO

      It's not enti- 'Cause Replit originally wasn't a ... I agree, it's not entirely true, but a- at a high level.

    13. JL

      Well, I, I wanna know what the question is. What I have certainly, now I get it, okay, I admit talking about it versus doing it are not the same thing. There is a huge number of developers that today already will get a prod- a project, and there are downsides to this, okay? Crappy code, spaghetti code. They'll get a project to a certain point, and then they'll just roll it into Cursor or wherever they want, and they will finish it for real.

    14. RO

      Agreed.

    15. JL

      I can't tell you how many CTOs I talked to over the weekend. They say, "I don't love the code I get out of Lovable or Replit, but this is what I do. I get it to 60 to 70% and then I roll it in," okay? And so there's an overlap in these circles of these two. Um, but the reason I think the, the market is so big, to Harry's point, is because the mar-... I think if Replit and Lovable keep, keep making developers happy and they make dog walker apps people happy, the TAM really is much bigger than Cursor. It's much bigger than Cursor or Windsurf. It could be like 50 or 100 times bigger. That's why I'll do the deal with you, Harry. If there's any room left, just l- let me know.

    16. RO

      I'm gonna push on that. And again, I don't have religion on-

    17. JL

      If it's a, if they're doing this round and there's any room, I might... um, I would put in a few nickels, even a small supporting check. Yeah.

    18. RO

      By definition, there are more non-developers than developers. So if you agree that the whole-

    19. JL

      Yeah.

    20. RO

      So yeah, at some level it's trite to say the Lovable market-

    21. JL

      And we've never tapped into those dollars ever, right?

    22. RO

      I, I, I'm not.

    23. JL

      Yeah, sorry. My apologies.

    24. RO

      The, the good thing, 100% of developers are going to buy a tool like Cursor, like Windsurf, because that's how development is done now, like GitHub, right?

    25. JL

      Yeah, 100, no question.

    26. RO

      And they have budget 'cause they're 100%-

    27. JL

      And 100% of designers will use Figma.

    28. RO

      100% of designers will use Figma. And the real question is, how big and how sustaining is this movement of non-developer or up... a- a- and, and by the way, you made an interesting point about the developer who starts in Lovable and transitions to Cursor. That, that's actually a good point, and that does give a little durability to the market. But it's not clear to me that you will have 10X or 100X the number of people who will continuously pay a subscription for a non-f- f-

    29. JL

      It's a fair criticism.

    30. RO

      And I, I could be wrong to be clear.

  8. 45:16 – 48:34

    Why Lovable could be the ChatGPT of builders

    1. RO

      to be-

    2. HS

      What, why, why did I write a $12 million check into Lovable? I think very much in the same way that ChatGPT owns the consumer brand for that front end OS for the modern consumer, I think Lovable's reached escape velocity with the consumer brand in this space, and I think that compounds very, very quickly.

    3. JL

      I think it's a good thesis, right? I, I don't know that Lovable's bigger than Replit. So just, there are maybe two that are the same size, right? But putting aside the bias-

    4. HS

      About the same.

    5. JL

      ... as an investor, okay, they're both good ones. Okay, let's call them both good ones, okay?

    6. RO

      Yeah.

    7. JL

      What I guess, Mike, other quick learning this, which is how important brand is in as you get closer, because listen, a bunch of folks are like, "Lemkin, why'd you use Replit?" Like, they're like, "Well, I use V0 plus this plus that, and then I export to, to, to, to, uh, to Cloud Code, and then I re-import to Cursor, and then I host." And I'm like, "Okay, I'll tell you, there's two reasons I picked Replit, okay? One, I wanted to go end to end. I wanted to go from ideation to commercial production one app, okay? And so I might be wrong about both, but to my knowledge, only Replit and Lovable can really do it. And I'm gonna get flamed, okay? But I want to do everything. I don't want to leave it. I don't want to, I want, I wanna port some code over that's better in Cursor. I wanted to go all, I wanted to stay locked in the constraints of one environment without a developer, okay? That was the, that was the social experiment, okay? And then I just asked people, and then I'm like, "Okay, now I'm stuck to the brands." There's no way I'm gonna use something for all the time I'm gonna put in that just launched on, um, on, uh, Product Hunt. It ain't gonna happen, right? And so there's only like two leaders, Replit and Lovable, okay? So I'm in the eight... So there are parts of my life where I'm in the 1%. There's parts of my life where I'm in the 20. Here, I'm in the 80, dude. I'm not gonna take any risk. Tell me which is the best, and most people came to the conclusion if you want more flexibility, use Lovable, okay? I, and we already use Lovable at SaaStr, don't get me wrong. If you want the easiest way to go end to end, use Replit. And here's my point from my rambly story. That's how I made my decision. That's it. Did I try them side by side for this use case? No. Did I spend weeks d- d- doing it? Uh, a- and then they're like, "Well, why don't you switch this?" I'm like, "I'm already, I, I made a choice between two vendors, and I'm done, man." (laughs) So that's why the investment, Harry's investment's good. I think these, these brands are gonna lock you in.

    8. RO

      That's the punchline, and that's the compelling part.

    9. JL

      Yeah.

    10. RO

      Maybe if I was, is, hey, what you're basically saying is this. If there's going to be a big ass winner in this space, the number one thing they ha- need to have at this stage is consumer brand recognition 'cause it's not a considered purchase. It's a try. It's a find the product as quickly as possible, and therefore the guys in motion today stand the best chance of staying in it.

    11. JL

      It's considered, it's just h- how much, how well can a non-technical resource consider it? Like that's the, it's not that it's not considered, it's just I can only spend so much time. How can I truly evaluate this in, in an hour or five minutes? I can't. You have, I have to guess what it's gonna look like in a month. (laughs)

    12. HS

      When you ask ChatGPT-

    13. JL

      Yeah.

    14. HS

      ... it says Lovable. We showed you-

    15. JL

      I'm not trying to do, make the VC argument. If, if, if, uh-

    16. RO

      Now you're-

    17. JL

      Right.

    18. RO

      ... actually making a very good argument for this whole category of, you know, um, companies that are helping B2B companies show up in ChatGPT, you know, do it, um-

    19. HS

      We just did one in Germany, Peak. When we did it, it, uh, 200, 200K when we did it. It's now at a million, like three weeks later. But fucking nuts the customer adoption there. One of the fastest pulls that I've seen in terms of consumer demand

  9. 48:34 – 49:30

    Will these vibe-coded apps become durable businesses?

    1. HS

      or customer demand. Jason, I, I, I, I, I j- I just want both of your perspectives on this before we move on to Grop 'cause it's important.

    2. JL

      Yes.

    3. HS

      Everyone goes, "Ah, but they're not durable revenues. Ah, they're unsustainable." How have your thoughts changed or stayed the same, and what do you think about the durability and strength of these revenues?

    4. JL

      I think, well, I think it's, I think they're gonna get more durable is my learning, but I didn't know before this weekend, okay? If I can actually get my app into production, commercial grade app, um, unless I have huge issues, I will pay and subscribe forever.Okay? If I get it over the line, if I get it over the line. Okay? If, if, if ... A- and then whether I pay $200 a month to Replit or Lovable, or I pay $3,000, it doesn't mat- like, I'm, I'm locked in forever, and I'm gonna keep iterating and iterating and expan- Like, if it works, I, and I get it, if I hired a team of six developers, you can export the code. I- i- i- it's in get, in GitHub already. It's in g- they can, they can do what they

  10. 49:30 – 56:32

    The shocking churn rates hidden inside AI SaaS

    1. JL

      want. But I'm pretty locked in as like a prosumer, non-developer shop, right? I'm locked, I'm locked in. Now, what's clear is folks that are fra- folks that, that, like, think they're gonna vibe code their own notion an hour, they churn. So I bet if we took Replit and Lovable and we s- and we ... And I, I wish more founders would do this. It drives me nuts when I don't do it. Segment your churn. I don't just want your headline number. I don't care today, I'm investing. I wanna see you segmented. If you have one cohort that has insane retention, and like this low end part is churning 4% a month, I get it. It's okay. Like I, I get ... There should be segments of Replit and Lovable that are churning like 10 or 20% a month. Like, like a low end consumer app. Like on RevenueCat where Harry and I are investors, they power 40% of mobile subscription apps. I think the average app across all mobile paid apps is like 6% a month churn. 6%, okay? This is consumer. They try it, they, they ... It doesn't work, they don't, they don't lose the weight, they don't stop drinking, they churn. Okay? And so that's gonna happen for Replit and Lovable. But what I didn't get is if this works, I'm, I'm never ... I ... My stickiness is gonna be super high, and my NRR ... I don't know how you calculate it, but you know, I- I've already gone from $20 a month to, to $800 a month in one month. If I stay, that's pretty good NRR, right? The VC's will like it.

    2. RO

      Right.

    3. JL

      If I ... Especially if you segment me out. So that's why I like your investment, and I like both as investments, because I think the metrics will ... It will get better over the next 12 months, not worse. I think they will improve.

    4. RO

      I mean, the short answer is a- a- a- agree. There's, there's the ... Jason, there was two different things to what Jason threw out. First of all, in the B- in the enterprise space, your retention at the start of the sp- Churn's gonna get better, 'cause out of the gate, people are trying stuff and by definition they're gonna churn. I remember back in '99, 2001, 2002, Salesforce had a churn issue. Right? Early on.

    5. JL

      It did, yeah.

    6. RO

      Now, the early adopters are quick ... The people who are quick to try you are quick to leave you. Right? And the people who make a long considered purchase stay with you a long time. And the f- And it happens in that order. Out of the gate, you get the triers. Most of the triers, m- A lot of the triers move on. Provided enough of them stick, you build your business. And then over time, as you become a more established product, you go up market, and you get the boring people who are slower to commit to you, but once they commit, are there for 10 years. So it's just the natural progression. So at one level, this isn't surprising. I will say what is surprising is the, the amount of upfront churn you see is probably higher than early SaaS. In other words, because it's the Jason phenomenon of it feels like it's gonna work and you get 80% of the way there, and then you have disa- disappointment. Right? So I do believe that there's hi- ... Every ... In my view, one of the key tells in the next 24 months on all these enterprise AI companies, is going to be the percentage of people that stick after they've gone through the innovation cycle, after they've gone through the, "We gotta try AI." And then they'll ask themself the hard question, is it delivering value? Do we get an ROI from it? So I think once we're through that kind of nut hole, I think the math ... I think about by '25, '26, '26, '27, sorry, the math of predicting the growth, it'll be much easier. Right now there are companies that have extraordinary high growth rate, but where the churn is just, just that ... Too high and it's just not delivering value. And we've all got them in our ... You know, you see them in your portfolio, you see them outside. The trick ... And you know, it's, it was always gonna be high no matter what. There was no world in which it was going to be like twenty- the twentieth year of SaaS, 'cause everyone is clearly trying shit here. But I think a lot of my focus with my companies is on ... I said to one, your entire ... Let me tell you, the asset val- Ironically, very different than wind stuff. The asset value of this company is the 20 or 30 customers that you have now and your ability to make them successful. If you can make two thirds of them successful, even if you have 30% churn, and you can add more customers like them, you have a business. And if you can only make 20% of them comm- successful and the others churn, you're dead. Right? So it's all ... When I'm checking in my companies, are your customers making money off your software? Are they happy with yourself? Is it doing what they want? Is the number one thing. So yeah, sure, it, it ... This is where the money's gonna be made or lost in the next two years.

    7. JL

      (laughs)

    8. RO

      'Cause we all like to try shit.

    9. JL

      Well, I wish I had all the non-sony baloney numbers in a spreadsheet. Alt- I would bet that, um, you know, Cursor's retention has always been higher than Lovable's or Replit's. For the p- point you made, uh, the- it's a different type of experimentation for a developer to commit. Um, it's funny, when I started sharing my bills on Replit on, on X, it was funny the feedback I got. Some folks are like, "Jason, you, y- do you realize what you could get for $500 a month is insane value?" Right? There, there's that category of people. Other folks are like, "Do you know what I spent on my outsource development team in my country? Like, this is, this is, this is the biggest rip-off I've ever seen." Right? And other folks were like, "Do you know, for $500, you're getting less than you get for a hundred bucks from Claude." Right? So there's very different ways to view this. But if you're deep in Cursor, like, you're gonna stick with our, your development environment, right? So the churn should be higher, right? At least for ... Of some segments, it should be much higher.

    10. RO

      And, and, and the churn is higher.

    11. JL

      Yeah.

    12. RO

      Right? The churn I, I thought was surprising ... I- I've obviously seen the numbers. Is, it's better than I thought, but it, it does lend more to consumer than it does traditional B2B, and it is higher.

    13. JL

      So you gotta s- But you gotta segment it to kn- As an investor, I need to segment it to have an opinion.... right? I, I, I, I'm not initially shocked by high churn as long as you have a large enough cohort that is growing-

    14. RO

      Totally agree.

    15. JL

      ... but has positive, uh, that is not negative churn, that has positive revenue retention, then I, I'm cool with the, with the low-end guys, right?

    16. RO

      And yeah, your churn from a group of customers, eh, even 10,000 signup customers should be dramatically different than your churn from five considered purchase enterprise buyers. So yeah, churn, even in the penant of this AI trend, you're exactly right. The n- norms on churn vary, enterprise versus SMB versus prosumer versus consumer. You're exactly right. And then on top-

    17. JL

      I got, I had an investment where a COO came in wh- who was out of the blue. He wa- you know, was under the gun, make, make, make improvement in the company, and he was told to increase, uh, to cut churn. So what do you think he did immediately? How do you think he did it in 30 days? He reduced churn.

    18. RO

      He stopped taking on the customers.

    19. JL

      Yeah. (laughs) He stopped taking all the small customers. He banned them.

    20. RO

      Absolutely.

    21. JL

      Now revenue stopped growing. (laughs) .

    22. NA

      Oh, and...

    23. JL

      But he hit his KAP hooray. Kudos. He, he, he, he stopped, he cut churn in half in 30 days. (laughs)

    24. RO

      No, uh, th- y- it sounds like stupid, but I can totally see it happening.

    25. JL

      (laughs)

    26. RO

      'Cause you know, you got this, "Oh, I don't like-"

    27. JL

      Incentives. (laughs)

    28. RO

      The real question is, you know, you gotta be able to look at that set of customers and say, "Yes, there's a churn, but the month two churn is much lower, and here's a quarter way." And the over- if the overall math works, it's fine, but, but yes, you have to have a mental model, and a mental model has to vary by the kind of deal you're doing. So yeah.

  11. 56:32 – 58:57

    Are these $2B valuations actually... cheap?

    1. RO

    2. HS

      I think the whole thing is, right, just in terms of you say mental models, I have never in my investing career seen a company go from zero to 100 million in seven months, personally. Like, it breaks all the mental models, which is why I did- which is also why I did the deal because if it does break all the mental models and the market expansion is as big as it could be, fuck it. We're in this business for companies like this. If you're not doing this and you have the chance, and it's one of 20, you shouldn't be doing this business.

    3. JL

      You know what's funny? You know what's funny, too? I, I wr- I tried to write this up, is, um, as weird as it sounds, some of these deals almost seem too cheap. If you look at the crappy, the, uh, multiples we see for some startups, right? Undifferentiated company raising at 300 pre with less than a million in revenue. Folks doing 100X, 200, 300 ARR deals, and then you see what is Windsurf ch- uh, selling for? 20X revenue. Lovable's raising at 20, right? Cursor raising at 20. 20, 20. Wh- let's do, let's bid 25, Rory. Let's outbid them. Let's do... These deals sound too cheap.

    4. HS

      Jason, when I di- the last round of Lovable was 200 million price, and I did it at 4 million in revenue. By the time the deal was closed, it was at 19 million in revenue. So, (laughs) so it had gone from a 50X to panic.

    5. JL

      Yeah, the ARR m- my point is the ARR multiples, uh, compared to public comps, yeah, they're high. But compared to a lot of, like, seed and A startups, they seem like bargains.

    6. HS

      Okay. We're gonna-

    7. JL

      You disagree, Rory?

    8. RO

      We're gonna bring... No, I don't actually 'cause what you're saying is, put it another way, it's ver- whenev- again, we said this before, whenever you're using revenue multiples that don't take into account growth rates, you're being an idiot. And the other thing is ex- and to your point, Harry, exponential growth rates are even harder to get your head around. You know, when you start, when things grow that quickly, a- if they continue growing, almost any price makes sense, right?

    9. JL

      Yeah.

    10. RO

      So I totally get the argument. But then it is worth pointing out then, going back to our first discussion, when they slowed, when you've paid that high price on the assumption of continued growth, and they slow down, it's just brutal, and you gotta get out.

    11. JL

      No, you just move them to a different page on the website. You just have Lovable move it over.

    12. RO

      Okay.

    13. JL

      "Dear Lovable, move this investment to Other Investment tab."

    14. RO

      Yeah. Other category, yeah.

    15. JL

      "Put, put grayed out image of partner that no longer exists at the fund. At- attribute investment to him." (laughs) Thank you, Lovable.

    16. RO

      I'm reminding you, I would remind you at SoloGP, Jason, we don't get to do that to you until you're dead. Okay. (laughs)

    17. HS

      (laughs)

    18. JL

      Well, you just make up someone that used to work

  12. 58:57 – 1:13:59

    Grok just destroyed GPT-4 in benchmarks

    1. JL

      there.

    2. HS

      Okay, boys, we're gonna do Grok. Uh, this was insane. The benchmark performance was incredible to see. Uh, uh, Grok destroys competition on humanity's last exam. Grok-4 Heavy, 44.4%. Grok-4, 38.6%. The next closest competitor for reference for people was 26.9%. What did you make of this insane outcome level from Grok in such a short two-year window?

    3. RO

      I think it's very sobering for everyone else. I thought it, I, first of all, I think it's a really, it's a significant and important point, right? Because we've talked in the past how if you look at the antice- I mean, if you look at people's willingness to value raw startups at very high valuations, it's all about the idea that they're one of the few people who was part of the OpenAI, Anthropic. "We know we can deliver a trained model coterie of people." Right? And it's been very much a small group of people, which is why they've all been paid 100 million bucks, right? Grok, in two years from a standing start, has gone to shipping a bro- I'm gonna come back to the benchmarks in a second, but a broadly comparable chatGPT-type model with a slightly different vibe. They did it. It's a huge, huge achievement, and the fact that it's done makes you wonder about how you think about the valuations of some of these other companies. And maybe it's not as... Maybe the knowledge has disseminated out just a little further. So I went in and did some work, a couple of things. One is you check out the team at, um, Grok, right? You just ask who are the engineering team? And it's exactly what you'd expect 'cause yeah, Elon's been around. He was founder of OpenAI, let's now remember, right? So it was early DeepMind people. It was some early people for some of the other key places. So it was not the, the headline people who got the attention paper, but you know, one or two levels down, really competent people. And what the aha for me is this, with sig- with committed clear leadership, which for all his faults, Elon provides, and a couple billion dollars worth of GPU, which see prior comment, the next level down people know how to do this too. The knowledge is disseminating, and it will continue to disseminate from here, right? So that was my first big aha. They have shipped... They're the first-... folks that I didn't think of as part of the initial golden circle, who have shipped a broadly comparable thing that feels, when you interact with it, like, you know, again, this is ChatGPT if it was vibed up. So big d- a huge deal, a huge credit. And then the second thing, I just gotta say it, Harry. I j- I'm sorry about this, but I just gotta say it. I was experimenting on Grok and using it, right? Trying to get it to say all the mean things, and they've clearly stopped it. But the whole tone of it is very out there, in your face, rah rah urgent. I mean, I was just doing my work on it for preparation for the podcast, and I was trying to think, "Where have I heard that before?" He sounds just like you. Grok sounds just like Harry. It's like, "Hey, legend, let's get going here." It's literally like talking to Harry. It's what... Jason, you should try it, trust me. Yeah, type, put, upload the contents and say, "Talk to me about, um, this in preparation of my podcast." And it comes on, and it's kinda, "That's just so freaking amazing. These guys did this. It's so unfiltered. It's so not like boring old ChatGPT or goody two-shoes Anthropic." It's just like Harry. Right? We should call the damn thing Harry. It's great.

    4. HS

      Listen, me and, me and Elon partnered together. We've been friends for many years, and I, I-

    5. RO

      You've been the reinforcement learning. Maybe there was an easy-

    6. JL

      He was in the first 50 by 20VC (laughs) interviews.

    7. HS

      Yeah. Yeah.

    8. RO

      But, but you can... There's a real tone there and, and it's great. Huge credit. And the other comment I'd make, I can't comment to the benchmarks 'cause, you know, I get a lot... I can comment on, this feels comparable to the other two or three, um, you know, model providers. On the advanced benchmarking, I mean, I've read some of the stuff that they published and then other people. There's a lot of gamesmanship, but I don't think that matters. What does matter is they've shipped something rel- They could j- If, if they could deliver the business side of it, which is a big TBD, they could deliver a ChatGPT-like experience that would feel pretty much comparable out of the gate right now. So at that level, they won.

    9. JL

      I think they could win.

    10. RO

      Yeah. Well, the question is-

    11. JL

      I was wrong.

    12. RO

      Do you think-

    13. JL

      What's that?

    14. RO

      But do you think-

    15. JL

      Yeah, listen, I got this all wrong.

    16. RO

      Yeah.

    17. JL

      When, when Grok launched, I thought it was a spite app. I thought it was a spite app and, and, and very little more. I thought, listen, he founded Ope- co-founded OpenAI. Clearly, he understands AI from Tesla at a fundamental level that other folks don't. And clearly, the dude can recruit, right? I mean, he, he recruited Ilya to, to OpenAI. The guy can recruit, he knows talent. But I thought it was a spite app, right? I thought, I thought all of X was a spite app. I thought he bought it out of spite. I thought Payments were spite out of PayPal, and I thought Grok was just, just to dig that knife right into Sam Altman. Um, but if you step back for a minute, he's shown that if you know what you're doing and you buy the GPUs and you assemble an S-tier team, whatever this data exactly means, whether it's pulled ahead or not, you can fucking catch up, okay? Only h- But... And he has access to not unlimited capital, but at the edge of unlimited capital. He does not have the Microsoft headache. He does not have the baggage that, um, Anthropic has, and he has an infinite horizon and the energy that we have never seen outside of Iron Man. So I honestly went, uh, uh, in the last release, I went from spite app to you could... This is such a fluid space, you could see him pulling ahead of everybody in 12 months. It's very, it's very plausible. He has none of the baggage of everybody else, and he can get away with crap like blowing up rockets or racist, uh, uh, things, roll it back, and it's okay, and the other guys can't get away with it.

    18. HS

      Well, well, that brings into question then, which the, the next point, which is like there's reportedly a raise going now at $200 billion valuation, the third funding round in, in two to three months. If that were the case, the combination of Grok and X as a social network at 200 billion comparative to OpenAI would actually be a very fair and good deal to do.

    19. RO

      I wouldn't, 'cause to be clear, I gave 100% credit for the technical achievement. Elon... Like, there was a technical bar to be cleared, and he's cleared it brilliantly. I did not agree with what Jason said. It's not clear to me from that that it gives you anything commercially. In other words, I don't know if they can build as the third or fourth. I mean, we're now... And this is what happened. We may be looking at the overinvestment stage of this, of this game, in the sense that you have OpenAI, you have Anthropic. Let's say you have Goo- you have Google, I think that's fair. Let's say you have Meta. You know, at best, I mean, even ignoring Mistral, even ignoring China, you're fifth or s- you're the fifth or sixth model provider, right? That's not a w- that sounds like the airline business, high-fixed cost, six customers. Will they be... Just because they completed the astonishingly hard technical task far better than anyone else did doesn't mean they get to build a business here. Right? You can't will things into being at the business level. I mean, I think-

    20. JL

      Are you sure?

    21. RO

      We'll see. I think that-

    22. JL

      I think the best founders do will things into existence.

    23. RO

      Yeah.

    24. JL

      I... That'd be my one challenge. All the best founders I know will it into existence, right?

    25. RO

      Again, and philosophically, I think you can will things into being if the, if... I think you can will yourself to be the winner if something is doable, right? If there's not room in the market for the sixth player, no amount of willing it can make it happen, right? So the question is, what's the case for a whole bunch of people who have... I think, remember what you said about early winners. I signed up for ChatGPT. I've signed up for my ChatGPT subscription. I've signed for a couple of others. Am I gonna go and sign up for the Grok subscription too? I mean, it's funny and I like the interaction, but is it going to become my go-to destination, right? I don't know. Right? I don't-

    26. HS

      Well, I don't... I, I... If it has your X data integrated and it's the only platform to have all tweets integrated, for me, that would be enough.

    27. RO

      Agreed, but-

    28. HS

      Especially when I-

    29. RO

      But... And let's just talk about that, and for the 10% of the world that likes Twitter and that actually cares about these things, that would be great. But the truth is, Twitter has always been a minority s- sport, sport, and I've enjoyed it 'cause I like the minority game. But, you know, yeah, 20X the number of people wanna like, you know, w- share shit on Facebook as wanna talk politics on Twitter. So w-You- you- you- you congratulations, you- you- you have your niche market, you know, to the extent you haven't alienated. But it's not just a mass market, you know. I- I don't think ... uh, there's evidence that Twitter's doing better, I don't think there's evidence that it is ... I don't think it can ever be anything other than what it is, which is the place you go for news, for politics, for arguments, and the dirty little secret, most people wanna look at pictures and, you know, just share shit on Facebook. So I don't think it'll be huge. I think they get that business, I think they integrate Grok in Twitter, I think it's great. I like it, I'm enjoying it. There's a lot they can do but I just can't conceive the steady drumbeat of taking over tens and hundreds of millions of consumers who've emotionally committed to Google, who've committed to ChatGPT and saying, "I'm gonna go with Grok." So, I think they'll achieve a very hard technical problem with very low rewards, so that's why I would pass at 200 million pray.

    30. JL

      (laughs) Billion, but yeah.

Episode duration: 1:22:54

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode wSudjTgRm7o

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome