The Twenty Minute VCWhy Apple Needs a Management Overhaul & Why Google is Catching Up with Hyperscalers
EVERY SPOKEN WORD
90 min read · 18,271 words- 0:00 – 0:50
Intro
- RORory O’Driscoll
Ironically, Google, who we all piss on, has executed the best of the four. They have a model that works. Apple doesn't even have a product that works. Microsoft bought someone else's product, but doesn't quite own it, so it's kinda weird. Facebook is desperately trying to buy product, but it's not, like, out of success, it's out of terrible psychological need for a product even though they don't have a business to justify it. I don't think these guys are too powerful. If anything, I think they're a bunch of rich people on the back foot behind the new trend desperately trying to catch up.
- JLJason Lemkin
Ready to go? [upbeat music]
- HSHarry Stebbings
Team, I'm so excited for this. We have a great schedule for today. Uh, as you know, it's always my favorite conversation of the week. So I wanna dive right in and start in the world of venture.
- 0:50 – 4:52
Why Benchmark Is Bleeding Partners (and Why That’s the New Normal)
- HSHarry Stebbings
We had Victor leaving Benchmark. Now there's only three partners remaining with, obviously, Eric, with Chaithany, and with Peter Fenton. Um, it's big news in venture. I'd love to understand how you thought about this and how you responded to it.
- JLJason Lemkin
One, listen, if folks are... if the best AI researchers are jumping from Meta to, to Anthropic to whatever in six months or getting massive packages, there's no stability, and it should be the same in venture, right? That's thought number one. Like, this is not... The, the bucolic, whether it was ever bucolic, but the idea of Benchmark in the eBay days of a bunch of, uh, gentlemen, uh, investors staying together for eighty years, it's possible that's part of the past. [laughs] Right? And then the other tactical thing, I learned this myself, if you are going to leave a fund, and you can, you have a hot hand, better to do it early. Like, staying longer doesn't help. Like, it might seem like it helps the fund. It doesn't help you. You're probably not gonna leave with much carry, right? You're gonna have to start over. So there's definitely reasons to stay at a fund, right, especially a top fund. But you might as well leave the moment you can because, uh, uh, you know, there's not a lot of economic incentive, not, you know, all things being equal, to, to stay, right? I mean, starting over is not that fun. [laughs]
- RORory O’Driscoll
Agreed. And yeah, no, I, I, I think it is interest- very interesting. I mean, I think stepping back, I, you know, I remember Benchmark founding in '95. I remember them pitching us in '95. So I would've said what's happened here, and, and Benchmark will be fine, so let's, um, start with that. They'll find two other people to fill those slots, and they'll be fine 'cause they've been fine for thirty years. I think the interesting thing is, to a rounding error, I would've said being a partner at Benchmark is kinda one of the top gigs in venture. They've been able to hire people from other places, other good firms, get them to join Benchmark. We all know the shtick, equal carry, focused fund, no layers of hierarchy. It's a really compelling story, and it's been a compelling story for thirty years. The stunning fact now is someone can be in the best gig in venture and decide, "No, it's not enough. I need something more." And they can actually probably pull that off. I mean, that to me is the big data point. It's, it's less about Benchmark. It's like a talented individual with obviously prior entrepreneurial success can be in venture two years and decide to leave Benchmark, not to go back to being an entre- a, a kind of a founding startup entrepreneur, but to say, "I can get money and continue this business on my own on equally or more attractive terms." It just speaks to where the market is, you know, rightly or wrongly, and we can talk about that in a second, for that kind of top-tier talent. It's very much a sign of the times, right? And you made a comment, most people who are good investors would prefer to just do it themselves. You stick together with your partners for, you know, a combination of, you know... Because, you know, you hope there's some strength in diversity of, um, uh, some diversification, to use a less loaded term, and because there was a perception for a long time of a bare minimum fund size, team size to get funded. And clearly all those perceptions are gone for at least a small number of very successful, talented people, and that's what we're talking about.
- JLJason Lemkin
The thing, the nuance I've thought about over the years is if it were me, if I were myself at that time, at a similar spot, I wouldn't leave Benchmark. I wouldn't leave Benchmark. Because I think what a lot of these folks going out on their own miss, and Harry and I did it, right? And Harry and I had brands behind us. We actually had brands, quirky brands, SaaStr and 20VC. But we under... I think especially today, we underestimate how powerful it is to meet a founder and say you're from Benchmark. It's frigging powerful, and I think, I would argue, Harry, I wanna get your thoughts. I would argue that you and I are at the e-edge. Anything less than a brand that we have, I wouldn't do. I, I just wouldn't. Everyone's different, but it's not that you don't have the skills of hunting. It's not that you don't have the skills of evaluating a relationship building or schmoozing.
- 4:52 – 8:13
“I Wouldn’t Leave Benchmark… Unless I Had THIS” Jason on Brand vs Autonomy
- JLJason Lemkin
But I, I, you know, there's a lot of the best founders, you know, they just want the same guys that are in Figma, the same fo- funds that are in Figma. They do, right? And t- and 20VC's above the line, but it's not, it doesn't have the storied history of, uh, of Benchmark. What do you think, Harry? That's why I, I'm not sure I would leave Benchmark.
- HSHarry Stebbings
I think people greatly over-exaggerate the impact that this has on Benchmark, and I don't mean that rudely to Victor. But I don't think any firm has played AI in the last eighteen months as well as Benchmark have done, despite the team churn that they've had. Let's just go through this. McCaw over a hundred million in ARR. Hagen hitting a hundred million in ARR. Fireworks, a hundred and forty million in ARR. Sierra with Brett Taylor, over a hundred million in ARR. Manus AI, the best Chinese AI team there is. Lagora, the best European AI team there is in many respects, all with double digits ownership in one single fund. Hell, how could I be an LP in that fund?
- RORory O’Driscoll
I'm gonna call stop here, only because we didn't a- you actually didn't answer the question. We agree. I mean, let's stipulate Benchmark will be fine 'cause they're awesome investors. That's just not an interesting discussion, right? I actually think until you cut him off, Jason was starting down the far more interesting discussion, and it's super interesting with this group 'cause I've actually never had this discussion with two people who've done the thing. There's very few peopleWho've done the thing that Victor did, which is raise a solo fund. And the odd thing is I'm on a podcast with two of the people who've done it, both of you, as brand names, right? So I actually think talking about is Benchmark awesome is fun because they are fucking awesome, but it's just not that interesting.
- HSHarry Stebbings
Would I do it if I was Victor? Yeah, of course I would. I would. Uh, and you know why, you know why I would? Because he showed in his investing that he likes to move across scale in a way that I don't think Benchmark likes or allows in that disciplined manner. In a similar way almost to Miles Grimshaw, who's another fantastic investor. But I think they are multi-stage investors and just wanna intersect with founders, whether it's at the pre-seed, like Victor did with Brex, or whether it's riding a $55 million check into HeyGen, which is really outside of what a Benchmark deal would normally be, being that size check on entry. And I think I look at it and go, "Hey, I just want to engage with the best founders. I don't want the, the constraints of a partnership. I can get a lot of money s- privately from great individuals, and I'll own 100% of the carry." A, freaking men, thank you very much. I'll do that all day.
- JLJason Lemkin
I hear you. I just have a more nuanced view now that time has gone on, right? I remember when I left, I had, uh, two, two of the, two of the two top brand name funds reach out to me that I did not know well. Uh, b- just it wouldn't have worked out, right? But I had a pretty hot hand when I, when I did. Harry will remember, right? And I didn't even take the meetings, right, for real, 'cause I'm like, it's just not g- it doesn't ma- I can't get the math to work in my head to, to 100% of the carry, right? I can't get the la- the lack of autonomy to work in my... And I'm like, I've been a founder with a modest level of suc- success. I ain't gonna go work for somebody. Like it was a n- I didn't take any of these meetings, like. But time has gone on, and I, I do think that y- just if you're gonna leave, understand the value of the brand you're leaving behind, right? It just, it just, it's, it's more, it's b- If you have no brand and here's my, my only p- and Harry challenged me. I just think in today's world, if you have no brand at all, it's tough. I think it's
- 8:13 – 28:38
The Unstoppable Force of Elad Gil & The Myth of LP Discipline
- JLJason Lemkin
tough. You disagree?
- HSHarry Stebbings
100%. 100%. You're so right. It's tough. But I do know Victor, and he is in the most exclusive inner circle of Silicon Valley that exists. He's in the Sarah Guo, Elad Gil, Pat Brady royalty-
- JLJason Lemkin
Yeah, and so he has his own micro brand-
- HSHarry Stebbings
Uh, he has his-
- JLJason Lemkin
... and, and relationships that will make it work for him, right?
- HSHarry Stebbings
Unbeli- And I would actually say the thing that's more important and attractive than brand is scale of cash. The only thing that I find attractive about joining another firm, which I never would do, ever, obviously, but is the fact that I could deploy a billion dollar check into a great company. And actually, I think scale of cash is almost a more attractive magnet than stellar brand like-
- JLJason Lemkin
That's the one way the world's changed, right? We wouldn't even be having this conversation-
- HSHarry Stebbings
Yes
- JLJason Lemkin
... even four years ago. Even four years ago, we wouldn't have the conversation, right?
- RORory O’Driscoll
Yeah. What you find attract... I'm just thinking here. What you find attractive about being on your own, to your point, Harry, it, it, it, the, it's the ability to deploy large... I'm just trying to understand it. The ability to deploy large amounts of cash. What, what Jason is missing is making the point is, so you give up on the brand if the existing brand is strong, so you probably have to come to the table with an equivalent brand, and you probably do because if you don't have an equivalent brand, you won't be able to get the big cash. So what you're basically saying is, and I agree, is that if I have, let's just say a functionally equivalent brand as an individual to the brand of the firm I'm at, in other words, Hai Hai, right? So I'm not losing a ton on deal flow from transitioning, and I get all the money myself, then I should just go do that 'cause I can do what I want with the income earned, right?
- HSHarry Stebbings
100%.
- RORory O’Driscoll
What it s- what's interesting is the people who need to think about this a lot are the LPs, and I'll tell you why. Because what it says is, we will now finance you to do all the things that we spent 20 years telling every other venture firm not to do, right? And I think that's a big insight, right? For 20 y- if you think about the kinda, like very much an LP focus on discipline, stick to your lane, do the stuff you do well. We value teams, we value cohesion, we value long term. And what they're now saying, and maybe correctly, you know, this is not a diss, this is a reflection on where the world is and the v- sometimes the value of listening to advice and sometimes the value of not listening to advice. What they're showing by their actions, and we'll talk about Elad in a second, is they also like a product that says, "Thank you for your input on being focused. I'm ignoring it. Thank but your input on being a value add in terms of taking board seats. In Elad's case, I'm ignoring it. Thank you for your input on small funds. As I'm ignoring it. Oh, and by the way, after giving me all this input, I'm ignoring it all, and then you're gonna give me the money." Note to self, you just... What are you doing here, people? It's super interesting, right? What it says is, as I say, "Thank you for your advice. Now thank you for your money." And that's, that's why I think it's, it's an interesting time.
- HSHarry Stebbings
What does that tell you then? That Elad has such a strong brand that he is able to do that, say, "Thank you for your advice. I'm not gonna take it," over and over again. And I don't mean that badly against him. He's incredible.
- RORory O’Driscoll
I, I think what it says is exactly that. I think it probably says two things and they're nuanced 'cause they go one way. One thing it says is, in the end, idiosyncratic success triumphs bland, mediocre, standard advice. You know, winners win. If you see something working, you're just gonna lean into it, right?
- HSHarry Stebbings
Mm-hmm.
- RORory O’Driscoll
You know, 'cause I think LPs correctly are saying, "I had this pattern matching thing and these are the things that I said we shouldn't do." And then I look at this, moving on to Elad here 'cause we had him on to talk about Nest. Oh my God, this guy was a seed investor at Airbnb and Stripe. He's ridden the last 15 years. He's not put a foot wrong. He has access to the best deals. Screw my rules. This guy's gonna make money, right? So on the one hand, it says, you know, capitalism works and there's a... Th- so part of what's going on is there's a big signal coming here that says, the things that we believe for 20 years mightn't be as true today. We should lean into the new. That's the positive side.And I believe that, to be clear. I actually look back and I go... And I'm sure you'll ask about this when I say it in a second, so we'll put a pin in this. That you look back and you go, "Hmm, there are things that I thought one had to do," that you look back and go, maybe you don't. Right? But let's leave that for a second. The other side of the table is this. Sometimes, sometimes rules of thumb or heuristics are in place because across cycles they've been proven to be correct. And it w- my guess is some part of this trend makes total sense, and some part of it you'll look back after the next down and turn and say, "Oh, yeah, that's why we had that rule. That's why we had the stick to the knitting rule. That's why we had the more than one partner rule." And I think you'll find some parts of this, like totally free and easy, will prove to be challenging as you move across an entire cycle. And zooming out, never forget, there's no data on anything through a cycle yet for anything that started from about 2010 on. Right? So again, I think that this-- So I think the LPs, I think they're discarding some of their rules of thumb, and in some cases they'll be right. But my guess is there was an embedded piece of wisdom in that that will come back when you discard it to bite you in the butt as you go to the next downturn. So that's a long answer, but it's a super-- I, frankly, I think it's a super interesting subject.
- HSHarry Stebbings
I, I think Elad is the best embodiment of the concentration of value. And to your point of, uh, you know, picking your space, you said very wisely, Rory, and I always remember this, Josh Kushner is a master property developer in the way that he picks the block and buys the best house on the block. Stripe in fintech, Databricks in data, OpenAI in AI. And I think Elad is very similar in terms of making sure that whatever that category leader is, whether it, it's Helsing or whether it's Abridge or whether it's Stripe, he-- or whether it's Harvey, he is in that leader, and he's in it big. I think that's really impressive with him. Um, and then he's also an amazing picker. I remember he sent me Vanta at pre-seed. The man has access like no one else. He's able to cover spectrum like I've never seen before. Do you know what I, do you know what I worry about when I see this news? Speed of deployment equals relevance. And I think about this because we're very disciplined on three-year fund cycles, temporal diversification. I know everything that we've said that maybe Jason doesn't agree with on temporal diversification. But just, like, doing what we said we'd do on how we invest and how slow we are. Um, and I, I see actually cadence of deployment can lead to relevance in a way that really benefits that manager.
- RORory O’Driscoll
Yes, it does. And I think we have-- We talked about this before. You internalize that in the short term that's true. And again, and the only guardrail on that is... Now, the only people who mind money are the people in charge of minding the money. And the people who are in charge of the mining money here are the LPs, and they're trying to figure, on the one hand, they want... 'Cause people who are deploying capital quickly and getting capital are, by definition, getting short-term positive feedback. No one deploys a billion dollars, values it at eight hundred, and then gets more money, right? So if you're deploying capital quickly, you can assume you're getting markups, you're getting success, you're in the hot deals, right? And what, you know, you're left with is the LPs trying to figure out, is this a hot hand that I have to follow and/or is this a flash in the pan signal and it could blow up in my face? And it's not one way or the other. It's a, it's a tricky thing to figure out. 'Cause, you know, by following... You, you made a comment on consistency. By fo- by doing what you said you'd do, by being consistent, you know, you, you lower the chance of just drifting off and screwing it up, but you probably pass up on the upside of being as aggressive as someone who's not honoring the temporal diversification and who is using that velocity, right? So a-a-a-all other things being equal, the person who's doing more deals has more relevance. So therefore, the only quest- as long as they're deploying it well, it, it's a double win. You have relevance in the short term and returns in the long term. If you deploy it badly, you have relevance in the short term and failure in the long term. And you know, I've seen, frankly, I, I've seen examples of both, right? I mean, I think Elad Gil is, for example, the example of the latter. I think wildly successful, seems to be doing everything right. I, I'm, I'm willing to stipulate one of the most talented investors of the generation and will do it really well, right? And continue on doing it well. That's one example. I've also seen the examples of people who've been relevant for two years 'cause they're putting the money out, and then the money blows up and they're gone. I mean, Tiger's the obvious example. Now, they're still around, but a shadow of their former selves. Their, quote-unquote, relevance in two thousand and twenty-one meant that they saw every deal that year, but they picked badly, so they got sent home. Same thing with SoftBank. You can be relevant... Put it this way, any early-stage VC worth a dime in Silicon Valley knows who the dumb late-stage money is, right? And you can tell. I remember watching, I won't even name names. I was like: Oh, that late-stage investor is attending that early-stage investor meeting. I know why they're there. That's money. That's, that's dumb money that they need. So, and, you know... A-and, and those investors obviously blow up. So in the end, it-- relevance is a short-term thing. In, in the end-- As I can say about this job, the only thing that's true about this job over the medium term is you have to be right. And the hot hand who's right across an extended period of time, and as I say, Elad, I think, is one of the best at that. You know, you deserve to get the capital, you deserve to get the runway because you've proven over 10 or 15 years your ability to pick and, as you say, stunningly pick both early and late, right? I don't think... Yeah, we can come back to what is wrong with the single person model, but I'm willing to say the following. If it doesn't work out, it won't be because the person in charge of it isn't brilliant. He clearly is. It'll be because structurally there are problems with that model or that stage.
- HSHarry Stebbings
Well, the interesting thing is when you look at, broadly, the best performing managers or the breakout managers of the last few years, I think you'd put probably Thrive, Greenoaks, Founders Fund, and Elad Gil definitely up there.
- RORory O’Driscoll
Yes.
- HSHarry Stebbings
And all four of them areOne person-led. I'm not saying there's not great teams beneath them, but they are one person-led very dominantly. That is interesting, and that's a very significant change from the venture partnership of old that was so predicated amongst the industry
- RORory O’Driscoll
It is. And it's easier, I think, perhaps wrongly, to be one person-led the later you go. And the reason I say that is this. If you are-- Let's just play it out. If you wanna be an early-stage investor, right, and you're putting up, doing series As... J-j- I'm gonna say just doing early stage 'cause I recognize that all the names you cited have some amazing early-stage investments, and I wanna come back to that 'cause it's important, right? 'Cause there was an insight in that, that the pure late stage guys didn't have. But if you wanna do ma-- if you wanna put most of your money in early-stage investments, right? What it means is, every partner can only do so much. And if one partner is called the dominant partner, then all the others are, you know, to some extent, not able to speak for the firm. And you don't wanna have a person doing a ten or a twenty percent ownership position who can't speak for the firm, right? So I think in early stage, if m- if all your dollars are going early stage, you, even as a partner in the firm, I want the guy in the next-- who's running ten more deals beside me to be roughly a peer of mine, because I don't want some weak person who's not that good, who's a, frankly, a bit of a beta, who's willing to work for half nothing to be running my money, right? 'Cause you want... Which is why Benchmark, as the quintessential early-stage firm internalized, you need to have, um, you know, you, you need to have equality so that anyone can be a premier partner. It's totally different late stage. When you ma-- when you're not trying to be on the board, you're not trying to be involved day to day, you're not s- you're not-- your time is not consumed by the deal. You're just making decisions, right? Making decisions arguably can be best done as a smaller group. So in the limit, you can have, um, you know, one person making the calls on the late, on late stage deals, and it's much more scale. Because you ma-- 'cause remember, if you're trying to put out a billion dollars and you wanna write in twenty million dollar checks, someone has to make fifty decisions. But if you put it out in hundred million dollar checks, someone only has to make ten. And it's hard to make fifty decisions. So I think these-- what you see about these, as you say, mainly late stage funds, is that someone is capable of making the big deci- they have structure to do other things, but then someone is ca- is, is calling the big shots, and they've been getting them right. And that's true for Greeno, that's clearly true for Founders, and probably be true for... I mean, my guess is fast-forward three years, there'll be three other people in the firm in junior roles, but that guy will be calling the shots, and he's earned the right to.
- HSHarry Stebbings
Rory, can I be so blunt? Does it make you question your model when you see these kind of dictatorial-led firms, not in a horrible way, but dictatorial-led firms with mass capital sources, and then you look at yours more discipline-led sized and with partnership structure?
- RORory O’Driscoll
No. Um, I think no because... And I think we're all children of the era we grew up in, and, you know, we've survived two downturns, and typically in a downturn, the high price late stage guys go bust, right? So, and I'm not saying all these guys will, but it's what happens in every kind. Because remember, you're only taking one risk as a late stage investor, which is valuation risk. And when it goes wrong, it goes wrong in a hundred percent correlated fashion. Everything blows up in your face. So over-
- JLJason Lemkin
My liquidation preference always bails me out, doesn't it, Rory?
- RORory O’Driscoll
Yeah, you mean-- And, and by the way, and just for, for our listeners, you're being sarcastic there. You're exactly right. It doesn't. So I think to bui- you know, to build over an extended period of time, I think you have to build much earlier, and that's been our primary focus. And I think to do that, you have to have much more of a partnership structure. So I think y- that's the place you end up in. Right? You are-- If you wanted to build a late stage firm, you can be much more hierarchical. By definition, almost, you should be because there's a whole bunch of economic reasons we can talk about that. That's a fun thing to do. And look, cynical comment, not cynical, realistic comment. If I, I said this to our LP. If I'd known in 2009 that we were heading into the best equity decade ever with zero recessions, coming out of the worst equity decade in a long time with two recessions, I might have played a different hand, right? If you know that the sto- you know, if you know that stocks are gonna go up pretty unstoppable. We're at an all-time high today. They've gone up for fifteen years. If you knew that, the correct risk-adjusted, work-adjusted, money-adjusted play was to just do Tiger and not quite shank it in 2021. I don't know if that answers your question, is, you know, if, if you knew then what you knew now, you might make that play and make more money, but you, you, you, you design to survive the cycles you've seen.
- HSHarry Stebbings
Well, we, we, we can move on, but I actually would go that Tiger will do better than people think given their position, given their positions in OpenAI, Scale. There's many others that actually are not, are not bad companies. And y-y- you said that they picked wrongly. I would argue they didn't pick wrongly, they just picked everything. And in that, there's some good and some bad. Um-
- RORory O’Driscoll
I totally agree. And it's not, it's not a, oh my God, they're... Look, they're stunningly successful investors. It's just, you know, as an, as an LP, you probably would prefer to have skipped that experience because one of the things you have to look at is, you know, you'll have this phenomenon of the average IRR is great, but if you only-- but if most of the funds went in at the top, you know, the LPs who just participated in that, not so happy. I agree. It's not that they're not gonna make money. I'm sure they have many mentions, but it's not investment excellence as you'd want to live it.
- JLJason Lemkin
I think... Listen, I know you wanna move. I think maybe it's a discussion for another time. To me, the interesting about Tiger, they played the game on the field, right? Like many are doing today. They played the game on the field. What they did though, at least in my limited direct experience, is they, they played it broader than most, right? So today, Harry, we're all talking about piling in. What should we do, should we do Anthropic at a hundred, right? Uh, Tiger rode the vibe of 2021, which I don't even think is barely worth talking about. It's such a strange world where ev- every, every B2B company was a winnerThat was triple, triple, double, double or better. And so what I think Tiger did, my-- and SoftBank, but really Tiger and B2B is they lowered the bar because everyone won. The average public S- uh, SaaS company in 2021 was growing 70%. Public.
- RORory O’Driscoll
Yeah.
- 28:38 – 38:20
Is Vibe Coding the New SaaS? Jason’s $10K/Month Spend Reveal
- RORory O’Driscoll
winner.
- JLJason Lemkin
I'll tell you for what it's worth, what I, you know, you also asked about capping Claude code, right? For, for, for overage. Uh, the one, I, this is, I know this is one, one part of Anthropic, but my experience in Vibe Coding, my, I mean, my, my Captain Obvious learning is we're vastly underestimating the revenue potential per developer, per user of, of Claude and Anthropic. We're vastly underestimating it. We're vastly underestimating it.
- HSHarry Stebbings
Why, Jason? Just educate me.
- JLJason Lemkin
Because what was clear to me, I was on a path to spend $8,000 a month Vibe Coding, okay? And even with that, Harry, what's happening now, and now the context windows have gotten longer. So now on Replit, and I think Lovable just did in the new release, you can have up to a 15-minute long context window while it's thinking through debugging or complex stuff, okay? You know what I would love to do now? Run four of them at the same time, or maybe even 15, okay? So that 8K bill, all things being equal, could easily be 10K. And, and I remember, and I didn't get it, you know, maybe a month or two ago, Farhan, C- one of the CTOs at Shop- I think he's CTO of Shopify, let all their developers use the most AI they wanted, no caps. And the top guys were using 10K a month on credits because they were running multiple things in parallel, okay? And if we're talking about 15-minute context windows and this, you know, whatever, open, open, uh, I mean, um, ChatGPT-5 has a, you know, whatever, a million tokens or a billion, a billion token windows, right? Debug my entire code base. Now do this other thing. And now, now I'm doing 10 things in action. I think everyone... My, my learning is I think every developer at a top tech company growing is gonna give them $10,000 a month of AI credits. Everyone's gonna get 10,000, not 200, which is what we're thinking. We thought that was a lot a couple months ago. They're all gonna get $10,000 a month, every leading tech company. It's cheaper than hiring any human.And you can't find humans. So if that means a 50X growth per developer spend, it's gonna be a CFO's nightmare. But putting that aside, that's 50X growth from where we are today for, for, for Claude Code and Anthropic. Or, or if they use it through Cursor, it really doesn't matter where you're buying the tokens, 50X per human. That's a lot of growth, isn't it? I'm convinced, like, it's not even-- it's, like, blindingly obvious to me now, and Shopify already got there. Like, 10K is fine, not 200, 10K. And we're gonna consume... You know, and the other thing that the CEO of Replit said, I think yesterday about this, was, "You can't cap it because any great developer will consume almost unlimited tokens no matter how cheap they are." So I think we're all gonna land at this 10K, eight to 10K per month. If you're a deve- good developer, you're gonna be on 24 hours a day. You're gonna be on AI, not a c- a couple minutes here or there.
- RORory O’Driscoll
And first of all, I, directionally, I totally agree with what you're saying, uh, which is zooming out a million miles, I think what you're saying is these gu- and put- and bringing it back to Anthropic, these guys have something, a model that makes developers who are expensive assets extraordinarily productive. Somehow the market's gonna find a way to massively reward them. I agree. And you made a comment on capping, and it's been interesting. Separately, Anthropic did some capping for their plans, and I think Cursor did some capping and some token limiting for their plans. And you, you'd asked the question, Harry, is that good or bad? I think it's amazingly good, right? I think it's-- what it says is there's, as Jason said, there's infinite demand for this product. Right now, at the price point of 200 bucks and the cost of generating those tokens, the economics doesn't work for the model provider. And, you know, you could look at that and go, "Oh my God, I'm gonna worry about the fact this is the economic model wrong." And at a point in time, it probably is. My guess is some of these gross margins are troubling. But when you're selling something with a couple of facts, one is the long-term trends are in your fa- favor 'cause all these costs just go down over time. Thank you very much, Mr. Nvidia. And you've got almost infinite demand. You just price it to the point where you just get these people hooked, and over time, your costs are gonna go down, and their hookedness, as Jason said, is gonna go up. And maybe today you're selling them 200 bucks worth of tokens. So you're selling them, you're getting 200 bucks, and maybe you are spending 250 bucks in tokens, so you cap them now at 170, and you think your margins are crap. But next year, that 170 will be 120, even if the token count goes up. And as Jason says, you raise that 200 to 400. What it says is there's a, not infinite, but an astonishingly strong demand pull for Claude Code and related products, either directly using it via Claude Code or indirectly using the Anthropic models with o- one of the other providers like, um, Cursor. And that's gonna reflect in the Anthropic numbers, which is reflecting in the valuation, and it's gonna reflect in a whole bunch of pushing and shoving around gross margins for a year. But in the end, you'll stabilize out to something that works for everybody. So I think, yeah, it's astonishingly bullish on market size and growth.
- HSHarry Stebbings
Moore's Law shows the reduction in cost over time has pretty much always been an immutable law. When you look at that immutable law and the impact that that will have on token cost, surely that impacts what we said there about per spend, per developer being 10K per month. Because in three years' time, it'll just be so much cheaper on a per token basis.
- JLJason Lemkin
Yeah, but the products, the re- whole reason y- your lovable investment works, Harry, is these products are so much better than six months ago. I, I, y- y- y- yeah.
- HSHarry Stebbings
Thank you, Jason.
- JLJason Lemkin
Maybe, maybe. But I, I... And you're right. Maybe it all plateaus. I think a lot of, that's shooting from the hip. I just think these products are getting exponentially better constantly, and we'll consume a, a, a million tokens and a billion and a tri- I mean, or the token value will be more. We just, we have inexhaustible demand to run multiple processes in parallel 24 hours a day for development. Like, the, our w- our agents don't sleep, Harry. This is the thing I didn't get until I ran, like, four different AIs for SaaStr. Our AI SDR brings in... We were, we were DM-ing yesterday, right? Our AI SDR brought in a deal, you know, uh, like a 70K deal, and it doesn't sleep. And you're, and you're-- and if you're an engineer or developer, your AI isn't gonna sleep anymore. Now, you don't want it touching production code, okay? I have learned that lesson, okay? But if it's debugging or, or working on complex issues or building out an algorithm or, or doing s- refactoring part of your code base, you will, you can consume, it can work for you 24 hours a day, right? So, and we will use all of the to- there may be a, an end limit, right? Just, and so maybe this investment will peter out, but I don't see it today, right? I think, good God, we're gonna be coding 20-- we're all gonna be coding 24/7 with our agents. Everyone's saying that, like, it's so hard to get a job in software today if you look at The Wall Street Journal. It, it's not remotely true. If you have the skill set today, if you're super smart in math or CS or physics, and you're in c- and you're in any school today, like, my son's a freshman, and he's off the charts in math. He already has offers, okay? And he's not even looking, okay? So we have this w- you know, this is what we know, but I can see it in the real world. If, if you're top-tier developer, like, you just wanna arm them up with every tool you possibly can because no one wants the, the, the folks that went to, uh, Devry programming school like they wanted in 2021. That market is dead, right? So we're gonna arm them up with 10K a month.
- HSHarry Stebbings
Listening to everything that you say then, Jason, Anthropic is at least a $2 trillion company.
- JLJason Lemkin
Maybe, right? I mean, uh, the, you know, the, it's just, there's just so much investment here, right? Now, are there issues? If, you know, if, um, you know, I mean, I mean, obviously, OpenAI is coming very hard at them right now, right? Um-
- RORory O’Driscoll
I'm not signed up for 2 trillion, but I have, you know, and I don't have to be to say that, you know, this is gonna be an extraordinarily valuable company. I, I, I, I don't, you know, there was a fo- there's only four or five companies that are worth more than a trillion, and it's worth pointing out that only one of them is, you know, really selling just kinda devel- uh, N- Nvidia, obviously, which is pretty much a monopoly selling, you know, 100% market share in the most valuable commodity you have. A- all the others touch every human being on the planet who has IT. AppleYou know, Google, Facebook, and Microsoft. So I don't think you-- for the record, just to say it, I don't think you get to a more than a trillion dollars if you're just selling stuff to developers and influ- which isn't to say I don't think you do extraordinarily well
- JLJason Lemkin
If Anthropic ends the year at, what, seven billion in revenue?
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
Right? Uh, uh, velocity. And, and let's say the average Anthropic developer customer is paying $100 a month now, and that goes to five thousand, okay?
- RORory O’Driscoll
Yeah, if you multiply-
- JLJason Lemkin
Now that's fift- so what's fifty- what's fifty times eight?
- RORory O’Driscoll
Jason, if you multiply by fifty, everything's amazing. I agree.
- JLJason Lemkin
Yeah, but, but, but in all-- but I'm not being, I'm not being facetious. This is my, this is my captain obvious epiphany that this is bigger than the four dollars a month we spend on Jira at Atlassian. This is be- this is... All the VCs like to talk about how AI's gonna tap into the human budget, w- and some of it is just Sony baloney. The... I've learned this is completely true for developers. Ten thousand dollars is nothing to replace a $500,000 fully burdened developer that quits in seven months, Rory. They're, they're q- these kids are quitting, okay? They're quitting for the $100 million off. I'm not being facetious when I think it's 50X. I'm not being facetious, right? It may not, it may not add up to two trillion, but I think the TTAM, the true TAM, is 50X what it is today w-without question because, because we-- this is the greatest case, not of making developers just more efficient, right? But of, of radically, of ra- of truly tapping into that non-existent human budget that people have, right? There's no one to build this software out there, Ha- uh, Rory. There's no one to build it. There's-- Outside of Cursor, no one can hire anybody. And even in Harry's-- even at Cursor, it's hard to hire, isn't it, Harry?
- HSHarry Stebbings
It is. They have a big war chest, and so they're gonna have-
- JLJason Lemkin
Even at Lovable, it's hard to hire, right? Now, Lovable doesn't win every developer they want on planet Earth,
- 38:20 – 40:29
Will Microsoft, Google, or Amazon Win the AI Infra War?
- JLJason Lemkin
do they?
- HSHarry Stebbings
Li-listen, you mentioned Lovable. We mentioned Replit.
- JLJason Lemkin
Yeah.
- HSHarry Stebbings
Um, we saw Microsoft, we saw Google roll out Lovable/Replit competitors. Question being there, is it too late? These are actually very established brands in Replit and Lovable in this case. Is it too late, or is incumbent distribution so overwhelmingly strong that they will be the victors with these?
- JLJason Lemkin
It's a funny one. Um, the, the, the Microsoft competitor to Lovable and Replit, right, in Bolt, um, now this will change. It's beta, right? It's limited. Um, everyone shares one database. So you may have seen some issues I had around my database, but at least I had my own one. Everyone shares the same database. Don't-- So Microsoft has the clear warning, be careful what you put in the database. Uh, may-maybe you don't want to build [chuckles] what, what's the, what's the, uh, what's the app that just had the terrible leaks today that we were just talking about before we started?
- HSHarry Stebbings
Tea.
- JLJason Lemkin
Yeah, don't build the tea [laughs] because everyone's sharing the same database. What that says to me is, look, they're, they... Listen, Microsoft can catch up. That's not taking this seriously enough. That's not take... Or it's being so panicked that you have to put something out there, right? Which is probably more the case. They're so panicked that it is a big deal, right? That Replit and Lovable are two hundred million in six months, that we'll put something out there that no one in their right mind would use.
- RORory O’Driscoll
Yeah. I, I, I agree. I mean-
- JLJason Lemkin
One database. They only bought one in... They could only afford one instance at Microsoft. They only could afford one, one, one database at a time. [chuckles]
- RORory O’Driscoll
They can only code one at a time. But yes, I mean, I, I... So the answer to the direct, is it too late? In other words, are the Lovables already too established for Microsoft to matter? I actually think Jason's is the right answer. It's more a question of are they gonna focus on this thoroughly and comprehensively enough to win? 'Cause if they di- if they put all their eggs in winning this basket, right, they probably could. But you know, it seems odd, for example, to angst as Microsoft about, "Oh my God, I'm losing $100 million revenue to Lovable," when wake up, I'm losing nine hundred million dollars to Cursor in a product I already have been shipping for five freaking years
- 40:29 – 50:07
Is GitHub Copilot the Biggest Miss in Microsoft’s History?
- RORory O’Driscoll
in GitHub Copilot. In other words, if the thing you delivered first can't beat the newcomer, how are you gonna beat the even more trivial newcomer where you're later than them, right? So I think, you know, brutally, I think-
- HSHarry Stebbings
Rory, is Gi- is GitHub just a massive fuck up for Microsoft?
- RORory O’Driscoll
No, no, I think-
- HSHarry Stebbings
If you're sitting in that-
- RORory O’Driscoll
It's like a-
- HSHarry Stebbings
If you're sitting in that meeting, you're... If you're sitting in that meeting, you're going, "Oh my God, we had the la- start, we had the runway, and we've let someone come up from behind and seemingly trounce us." Is that the mindset, you think?
- RORory O’Driscoll
I mean, they probably would say, "Oh, we have five hundred million dollars of revenue we're doing." They probably have some corporate story why they're better. But the truth is, yeah, it's a mistake because whenever you sta... I mean, the thing about in any market when you're Microsoft, if you start with a monopoly and you fast-forward three years and someone exists, by definition, you screwed up. Your job as a monopoly was don't lose the monopoly, right? They had massive market share with GitHub and, you know, they added the Copilot, uh, you know, when that was CEO. It was an... That one said obvi- it was an excellent product and, you know, th- fast-forward four or five years, they seem to have slipped up. I mean, you can't not say it's a mistake. I mean, because you shouldn't let that happen. So yes, it is a mi... And that's my point. If you're chasing shiny object system-- symptom on the next thing, which is now I have to do a Lovable thing, are you really putting enough effort into making sure that your, you know, your Cursor competitor is as good as it should be? And you know, I think for all, and this is kind of, and I'm not fully formed on this yet, so I'm kinda going out on a limb. For all the Microsoft is amazing in AI story, they're better than AWS as of the three hyperscalers, but, you know, they don't have their own model. The relationship with OpenAI is a bit tortured. Cursor's now nine hundred million, and their developers, which is their core kind of raison d'être, uh, in a world where they're at five hundred with a five-year start, I don't think they're playing a perfect game. And I don't know if trying to do still more things is the way to win here versus actually doing some things well.
- JLJason Lemkin
And that's the trap, right? Trying to do even m- do even more things usually doesn't work, no matter who you are, right? Startup or Microsoft-
- RORory O’Driscoll
Yeah
- JLJason Lemkin
... it doesn't usually work.
- RORory O’Driscoll
And, you know, you know, Satya made the ever so funny comment that, you know, he wanted to make Google dance. Well, you know, fast-forward a year or two, they appear to be dancing pretty fucking well.
- HSHarry Stebbings
Microsoft own OpenAI in many respects. Amazon own Anthropic in many respects. And then Google own distribution and Gemini in a way that is very, very powerful and important. You have to buy one and you have to short one. Which one do you do?
- RORory O’Driscoll
Which of the three stocks are you talking about?
- HSHarry Stebbings
Yeah.
- RORory O’Driscoll
Buy Google, short AW- Amazon. They're just-- I mean, and I'm not fully developed on that yet. I haven't spent time think- but Google has the advantage of underestimation, right? It, it could lose its search business. That gnaws at me. You know, there's a, there's a lot in this, 'cause remember, I was just talking about the Hyperscalers. If you looked at the three Hyperscaler execution, then it's easy. Let's just do the Hyperscaler first, 'cause otherwise we'd have to think about retail and search, which are big things, I admit. On the three Hyperscalers, five years ago, Amazon built the category and was the dominant one. Microsoft was number two because they could lever over into the Azure, so they were kind of limping onto the field with a bunch of relationships. And Google was nowhere, to the point where there were credible stories about them getting out of cloud. Fast-forward to this quarter, Google's nailed it. Great growth in Q2. As of the recording, we don't know where Am- where Amazon and Microsoft is, but funnily enough, by the time you produce this podcast, people will know, because I think they're reporting tomorrow. But my gut is, just looking at the trends, that Google is pulling ahead, uh, uh, and is the smallest, but by far the fastest-growing. Amazon's growing-- Sorry. Microsoft is grow- is roughly equal in scale to AWS, admitting that there's always some weird accounting of Azure, and has a slightly better growth rate. So by definition, of the three, AWS is the Hyperscaler underperformer, and I'm 100% certain I could make that case. That was a lot.
- JLJason Lemkin
I, uh, listen, I don't-
- HSHarry Stebbings
That was great
- JLJason Lemkin
... obvious- I would, I would, I-- If I had to make the bet today, I would make Rory's, but I wouldn't, I won't make it for the captain obvious reason that one of the reasons is there's unlimited demand, and it has its own TPU, right? As Amazon licks its wounds and recovers from, uh, which it's already working code red on, right? When it, when it gets back in the sweet spot of what its customers need, demand will be unlimited.
- RORory O’Driscoll
Agreed. I actually-- I think you're right, Jason, which is the unlimited-
- JLJason Lemkin
Look at Oracle. [laughs]
- RORory O’Driscoll
No, agree. Oracle sell-- I mean, if you have, if you have-- I mean, and I look at this a lot 'cause there's a little part of the contrarian in me that says, "Could you hit the point where that stops?" 'Cause if it does stop, the correction's gonna be super interesting, right? So I look at it, and then I have-- And I, and I'm kinda trying to be the cynic, and I'm trying to find some evidence of overshooting on demand. And to your point, Jason, you're exactly right. I can find none. You kinda go, uh, I can't remember, was it Amazon or I think Am- uh, Microsoft said, you know, biggest problem, didn't have capacity. Um, I think one of the other Hyperscalers said, "We could sell more shit if we had more shit to sell," right? That is, uh, and you are right, Jason. That is the best thing in the world, the simplest thing in the world, is to be selling something that everyone wants to buy, and it's just a beautiful place to be in right now. And all those guys are there, right?
- JLJason Lemkin
Yeah, and The Wall Street Journal had an article yesterday of what the number one KPI for public company CEOs is. It's reducing headcount and growing revenue. Reducing headcount, that's the number one. From Bank of America on down, it went through all these big-- not just the little tech startups. Everyone wants less headcount. And, and the answer, meta, is AI. There's unlimited demand to, to, to do that, to drive headcount down and revenue up. So i- it, it's go- I guess it's gonna end, but when we're all just living in the matrix in, in-
- RORory O’Driscoll
I don't think it ends that way, to be-- I'm just gonna go out another limb. I don't know. I've had two coffees today. Is I don't think it ends when, you know, everyone's automated and the demand stops. I think what really... And I'm keeping an eye on this 'cause there's a trade to be done when it happens. I think what's interesting is the creation of AI kind of is happening at hyper pace. There's, you know, $600 billion of CapEx, whatever, just a huge amount going on. Enterprise's ability to adjust it is still relatively slow. Even though demand is growing, you're still at the 10-
- JLJason Lemkin
Super early
- RORory O’Driscoll
... 20%... And I think one of the weird things would be if a- and I'm gonna call it, if-- How long can you continue investing 600 billion a year against a $30 billion growing revenue line? I mean, at some point-- Like for example, even Microsoft, it was interesting. They did a bunch of layoffs and, uh, you know, the-- And, you know, uh, Satya did his comment on the enigma of growth. I didn't quite get-- I think what he's saying is really, he's trying to say, "Yeah, it's pretty cold-blooded to fire a bunch of people when you have hyper growth and 40% operating margins." But what he didn't say is part of the dirty little secret is we're gonna have to start depreciating, you know, half our free cash flow and CapEx through the, through the income statement over the next three years. So we're gonna have to take people out just to be the same level of efficiency. So there's a lot going o- you know, someone-- At some point, someone's gonna depreciate all these NVIDIA chips, and someone's gonna have to pay for them, right? So I do, to your-- And I'm sorry to go, I do think there might be a slowdown at some point in time, but right now there's no evidence of it whatsoever. And, you know, everyone did the, "Oh my God, NVIDIA's in trouble at 140, and now it's at 170." You know, the only trend worth a damn in the last year has been l- lean into the hyper growth. Lean into and limit the demand.
- HSHarry Stebbings
Just so I understand, Rory, you're saying depreciation is the primary reason why you'd see that dampening?
- RORory O’Driscoll
No, I'm sa- Yes, I'm saying-- Not just depreciation. No, never g- No, I'm saying if you continue to spend and run through your income statement large CapEx without an associated revenue line, eventually the world says, "Dude, maybe you shouldn't be doing this," right? And the truth is, right now, you're getting the depreciation. You're not getting the revenue commensurate with the spend. You're getting revenue growth. Jason is right. There is-Large corporate demand, but even with the best will in the world, they can only digest so much software in a year, right? I mean, that's-
- HSHarry Stebbings
Yeah.
- JLJason Lemkin
David Cohn from Sequoia is-
- RORory O’Driscoll
And he's got-
- 50:07 – 53:23
Are Big Tech Incumbents Now Too Powerful to Fail?
- HSHarry Stebbings
Can I ask you, do you think, and I, I sound like a communist here, but we have never seen incumbents at this size and scale before with this dominance, where they can drop $15 billion on a team and it's 40 days of free cash flow. Bluntly, who gives a shit about the money? It's worth the chips on the table. And when you look at Google's asset base, it's insane.
- RORory O’Driscoll
Yeah.
- HSHarry Stebbings
Are these companies too powerful?
- RORory O’Driscoll
No. I'm just gonna go straight out to the no, 'cause whenever you've been... Whenever people have been equivocal on that, one of my bigger hars, whenever people have been equivocal on these kind of questions, you end up with people coming up with dumbass things like the whole AI regulation, like the whole, you know, the whole FTC process. So there are some things about these powerful corporations I don't quite like, right? And I wish we could do nuance and, like, fix some of the problems without going crazy. But if you have to pick one of two worlds, either idiot regulation or just let free capitalism run wild, I'm voting for the latter, 'cause I think in the end, Mr. Market will take care of these things.
- HSHarry Stebbings
I don't know, dude. Mr. Market says that, you know, Adam Smith's invisible hand, and you're seeing that now with Zuck hiring with an unlimited budget, making B2B hiring, to Jason's point, the biggest problem in startups. That will just get worse.
- JLJason Lemkin
The interesting thing is we have to decide if oligopolies are okay.
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
Now, monopolies we could d- monopolies you don't have... We can... The case is clearer, right? And maybe monopolies are good. Uh, uh, monopsonies are even more complicated. But we could argue mono- the monopoly dynamic. Here, all of these that we're talking about, the reason they're interesting is they're oligopolies. We have three to four players, right? We have... Oracle's all of a sudden competitive, right? Ma- massively competitive with Google, with Microsoft, with Amazon. We have OpenAI, Anthropic, Groq and friends, like... Uh, and so if you wa... I, I think if you wanna give up on oligopolies, I quit this venture. I quit this business. Let me return capital. I'm gonna do an open view. I'm calling it a day. Here's the rest of the money. I'm gonna write out my winners, because i- if oligopolies aren't okay, I q- I quit. Like, the... And, and the re... You know, if you read a classic textbooks on oligopolies, they compete on features, not price. There's a lot of hints of that, right? That these products kind of are the same price, right, today. And a lot of folks believe that's good for a, an economy. Is, 'cause it, 'cause oligopolies maximize innovation. They don't maximize discounts, um, but they pour all the money back into R&D because they're competing on features, not price, right? If you believe in technology, you might almost want oligopolies, right? Three to four players. Three to four players.
- RORory O’Driscoll
I totally agree with Jason. I just wanna come back to your point, though, on too powerful, 'cause I think they're two actually slightly different comments. All the incumbent older companies, the trillion-dollar companies, are, in their own market, monopolies. Apple, I would argue near Microsoft, Facebook, Google, in an old market. Search, iPhone, you know, social, though they'll, they resist that definition forever, and whatever Microsoft is corporate, right? So in the old, in the old businesses they're in, they're monopolies. But Jason's exactly right. In these new businesses where they're putting their money and where the new startups are coming in, what you're seeing is oligopolies, which is what you'd expect. Now maybe fast-forward 10 years and they go- grind down to monopolies.
- 53:23 – 58:59
Apple’s AI Problem: Is It Time for a Management Overhaul?
- RORory O’Driscoll
But right now, in the markets that matter, for all the power that Apple, Facebook, Microsoft and Google brought to the table, it's OpenAI and Anthropic that are making the run-in. So I would argue that says there's simply no need for these whining, worrying FTC people to sweat it, because the truth is none of those companies executed well. I mean, ironically, Google, whom we all piss on, right, and say they're shit, has executed the best of the four. They have a model that works. Apple doesn't even have a product that works. Microsoft bought someone else's product, but then doesn't quite own it, so it's kind of weird. Um, who else is there? Uh, Facebook is desperately trying to buy product, but it's not like out of success. It's out of terrible psychological need for a product, even though they don't have a business to justify it. So I don't think these guys are too powerful. If anything, I think they're, they're a bunch of rich people on the back foot behind the new trend, desperately trying to catch up.
- HSHarry Stebbings
Rory, what do you do if you're Apple? You're looking at that assessment, which I completely agree with.
- RORory O’Driscoll
I think as a board you say to yourself, do you have a management team that's too old? 'Cause that's the only button you have. And I love Tim Cook. I have made a lot... Apple is my largest single position. I have agonized about selling it for a decade and a half, and I haven't. So I love that man. Thank you very much, my little house. Much gratitude.But you just do wonder, is this the team to grab what's going on in AI, which they absolutely should have a product on? And at some point, do you say you just have a bunch of folks who aren't figuring it out, and do you make a change? And that, you know, that's, at, that was the comment. Rather than some kind of tactical, "Hey, you should buy X, Y, or Z," look in the mirror and say, "Are you getting this done?" Now, I think it's very hard for a board to do. I think it's very hard for a board fu- full of supernumeraries to do. It's very hard when the numbers are still good, but you know, you haven't grown in five years. You've done a magnificent financial engineering, thank you very much again for your dividend, but you're not winning, and at some point the buck stops at the top.
- JLJason Lemkin
Well, do you think AI, I don't know, do you think AI is, do you guys think AI is a legitimate threat on its own to the App Store?
- RORory O’Driscoll
No. Um-
- JLJason Lemkin
Because if I look at Apple today, it's 25% of the revenue is from the App Store, 40% of its profit, right?
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
And 75% of its revenue is iPhone and App Store, so, uh, if, if, if AI just enables more apps to be purchased and Apple keeps its tax, right? And AI and, and App... Maybe, maybe you should stick to your lane in the short term, right? If you couldn't get, if you couldn't get Siri to work after 27 years, um, if App Store is, is a monopoly, to your earlier point, and it's not, not everything's under threat from AI, right? Not ev- not everything. If it's a beneficiary of AI, right, then maybe, maybe they have the last laugh. Maybe they have this high-margin last laugh of all this AI revenue gets, more and more of it gets routed through App Store, and they take their 26%.
- RORory O’Driscoll
Fair comment.
- JLJason Lemkin
Could be worse. Could be worse than taking 26%. [laughs]
- RORory O’Driscoll
I, I, by the way, I agree with that, Jason, 'cause you're right. When you make the hardware and the logistics of making the hardware the thing you're amazing at, and hopefully the design, your whole business doesn't go away. You're exactly right. You don't have ex- Like, you can talk about Google, ironically, and say existential threat to the downside if no search. No matter what stuff we're using for AI, I think we'll be using on our iPhone. So you, I think that's a very fair pushback. It's not as existential, right? But, you know, the brutal thing about capitalism, it's like, "What have you done for me lately? Congratulations on building the best consumer product in the last 50 years. Thank you for the App Store business and the services business, which is fucking awesome, but did I mention, you know, what's the new, new thing?" Right? Do you want to run the risk of another, you know, we, waving my arm furiously here, OpenAI-type personal companion that knows everything about you? Do you want to let anything get between this and you as a user for your one billion user base? So I, I agree, it's not existential, it's not like... But it's a, if you're not moving f- you know, it's the shark thing. If you're not moving forward, you're dying, and they're not moving forward in AI. Google pays them to be the search provider, and maybe they just collect their tax that way. You're exactly right. Then that OpenAI can do the same thing, and maybe Apple says, "We build great hardware. Our hardware has such a dominant place that people will continue just to pay us money, so, you know, we got it from Google, and we're fine even though the courts are trying to stop that. Now we're just gonna get it from OpenAI." You're right. That's the argument that says, "Stick in your lane. You make the 20 billion from other people giving you money." I don't know if I buy it, but yeah, that's the argument.
- JLJason Lemkin
I don't know how much OpenAI pays Apple today. Let's assume it's still zero, right?
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
Which it used to be. I'm, I'm skeptical. But, um, you know, if it's, I mean, Jesus Christ, uh, yeah, 108 million going through per month, uh, I may have the number wrong. You know, if, if, even if there's a sweetheart deal here, if Apple can just keep 25% of all that, it's a good model.
- RORory O’Driscoll
Uh, one of the challenges I would push, therefore, is owning the hardware is a great bu- and it has a long inertia factor. In other words, long after you stop innovating, you can keep monetizing, right? Which is, I think, the stage where they're at now. And the question is, at some point, do you need to do more than that to keep not just surviving, but growing, right? And that, that's, and that, and, and maybe it is stay in your lane and just get paid by everyone else, effectively for product, for placement.
- 58:59 – 1:04:49
Zuck’s Talent Siege: How Much Is Enough?
- HSHarry Stebbings
I loved that debate, by the way. That was a fantastic-
- RORory O’Driscoll
[laughs]
- HSHarry Stebbings
... discussion. For me as an observer, it was brilliant. Uh, but-
- JLJason Lemkin
It's a tough one
- HSHarry Stebbings
... you said, you said there about management team, switch maybe. Zuck hired again, uh, with another stellar hire from OpenAI, chief scientist or whatever it was. I mean, just he is going all out. To what extent are we, like, 80% of the way there in his mind, do you think, in terms of his acquisition on talent spend, versus this is 5% of the way there? Like, when does Zuck stop this siege?
- RORory O’Driscoll
I don't have a clue. He seems to... I mean, his argument, which when you heard him speak, is if I'm spending $30 billion on CapEx a year, or 40 billion, whatever it is now, do I need to sp- uh, should I skimp on the 5 or 10 people who can spend that most correctly? It's Jason's point. The super S-care engineer is spending so much money that, you know, you, you don't care. My, I mean, my assumption is that at some point, at some point the margin return on the next scientist gets negative, so you probably stop spending 100 million bucks a pop. But you know, I get what he's done. He said, "I'm gonna spend $40, $50 billion in CapEx. I better spend $1 billion, $2 billion, $5 billion to make sure I have the smart people to kinda implement that program." It makes sense.
- JLJason Lemkin
You know, and it's not even that big of a deal because if, if, if he wants, which isn't gonna happen, he can always just stop. It's only in the past. He can stop tomorrow and say, "We're just gonna, we're just gonna harvest classic Faceb- Facebook ad revenue. We're gonna become a, the next Yahoo, and we're gonna kinda just be a cash cow." It doesn't ma- What, like, it, it's, it... If you're a public company that's massively profitable, the past doesn't matter, does it?It doesn't matter. Right? You can write off any- you can write off anything, you can fire all the people, and you're still obligated with these massive, uh, contracts, but you can also write them off as one-time expenses or stick them in another bucket. They don't mat- they don't matter. If you, if you have massive cash, you can always do a mulligan at least every eight quarters.
- RORory O’Driscoll
Yes. And obviously that they did a mulligan on virtual reality with the whole Meta-
- JLJason Lemkin
Yeah, it's a mulligan.
- RORory O’Driscoll
But the interesting-- but the, the argument i- well, again, he gets to the, are you buying insurance or is it existential? I mean, i- implicit in that statement is that if you do nothing, if you're unsuccessful in AI, your core business just continues to compound, right? And as long as that's true, you're right, Jason. If the core business is kicking off, what's it, $200 billion a year of revenue, you know, 40% of... If the core business is kicking off $40 or $50 billion and you make a $40 or $50 billion mistake, you just go in the penalty box for a year and you move on. You're exactly right.
- JLJason Lemkin
Yeah.
- RORory O’Driscoll
But of course, part of the reason he says he's doing it is because he says it's existential. The horrible outcome is you misfire on the new innovation and the new innovation eats your core business, right? Like I would argue for Google, if they misfired on AI, they wouldn't have search. Whereas I think your point was right, more right than me. If Apple misfires on op- an OpenAI type AI product, they'll still be making hardware and it won't be existential. And to your point, if Facebook misfires on Meta, they clearly still have Facebook, and if they misfire again on AI, you're-- as long as they still have, you know, $40 billion of free cash flow, nobody cares. So you're right.
- JLJason Lemkin
And I think it's even-- I think the other thing is, like, when I-- intentionally, y- you know, Zuck is telegraphing to the market, "I'm making this bet," like, very clearly, right? And so, like, when I was a VP at Adobe, you're trapped with your 40% net, net margins because you can't invest in anything, because the market is expecting cash to rain to the bottom line, come... And yeah, Chatoon did a good job when they were moving to the cloud of, of getting, getting, getting some credit once for a massive generational shift. But, but Zuck can not only spend all this money, but he can say to the market, "Give me a little bit." And, and Larry's doing the same, right? At Oracle. "Give me a little time," right? "We're gonna get back to those classic margins," and the market won't punish you. You c- because most public companies are trapped in their margins that are profitable.
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
You're stuck. Like, you can't go back. You-- it is almost impossible for most public companies to actually spend the cash they're, they're generating. So Zuck's got a triple down here because he's getting also a hall pass. The market's saying it's okay to spend everything, right?
- RORory O’Driscoll
Look, remember, it's not only that the market won't punish you, the more important point is the market can't punish you because you avoi- in, in the case of Meta, you avoided control. So, you know, yes, the stock-
- JLJason Lemkin
Well, yeah, your market, your, but your s- your stock price could be punished.
- RORory O’Driscoll
I know. I mean, but it, but it's worth pointing out, if you're-- but I-- that's why I think you're, you're exactly right, Jason. If you're a, quote, "normal company," your stock price goes down and the vultures circle. Whereas if you're Meta, your stock price goes down and you're sad and you say sorry and your employees don't do as well and there's dissatisfaction, but you still have control. Same thing with Oracle. So I think that you're right. Those two players can swing a bat secure in the knowledge that no one has a voice other than them, which must be mad.
- JLJason Lemkin
And, and they, and they can actually take, they can actually spend their cash, which is otherwise trapped.
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
It's trapped with these large public companies. It just builds up in weird areas and it's unspendable. All you can do is repurchase your stock, right? Most-- all you can do is buy something, an acquisition, and do some accounting shenanigans, right? Or repurchase your stock. Most, like, steady-state, boring public tech companies are trapped when they're profitable. It's not, it's not the g- it's great because you're, you're not losing money, but, you know, the laypeople are like, "Why don't, why don't you invest it?" It's not that simple. Your mar- your, your, your EPS goes down. No one, no one likes that with a normal company, right? No one likes that with a normal, normal, normal, mature... No one wants to see Zoom's revenue, EPS go down. Like, it's the whole value of the company today, isn't it?
- RORory O’Driscoll
And this is why, Mr. Buffett is right. Once a company becomes ex-growth, the most important thing to assess about management is their ability to rationally allocate capital, including sending that capital back to the shareholders. You-- which is, as we said before, Oracle has done in spades.
- 1:04:49 – 1:15:22
Figma’s IPO: $30B Return, Zero Hype. What Happened?
- RORory O’Driscoll
Anyway, what else, Harry? What else we got?
- HSHarry Stebbings
Fi- final one, and then we'll do a quick fire, is just Figma's obviously gonna IPO this week. Uh, I do think we have to talk about this. I don't think enough people are excited enough about it. Um, I remember when this would've been the most exciting thing in tech this week, and it would've dominated news. This is, you know, it looks like it'll price at, you know, eighteen point eight. I think it prices, you know, uh, 5:00 PM, uh, either today or tomorrow, um, which is yesterday when this is out. Um, are we excited enough about this? How do you expect this to pan out? Is a pop baked in? Help, help us understand it as an audience.
- RORory O’Driscoll
I think it's unfolding. If you, if anyone listened last week and this week, by definition, they were glutton for punishment. But it's unfolding exactly what we said last week, which was the price, the filing range was, I think, twenty-four to twenty-eight or something like that. They drummed up demand, which felt low. They've drummed up demand. Just in the last twenty-four hours, they've refiled to, I wanna say, thirty-two to thirty-five. So it's worked. The bankers are doing a great job. They've whipped up enthusiasm. Now that they know they're multiply oversubscribed, they've raised the filing range. I'm sure the rest of the drama is pricing at the high end or maybe above the range, and then a pop. So the whole movie's unfolding exactly, you know, mission executed. There's been a lot of talk about how they're trying to make people be more specific about their demand to better allocate the shares and avoid that pop. But I think the structure of these approaches means that that's almost unstoppable. So my gut is, and this is horrible 'cause you're predicting something. You're breaking the first rule of predictions. You either predict a number or a time, but never both. And I'm predicting something now that by the time this thing comes out, people will know if we're idiots or not. But my sense is it's hard not to imagine this thing pricing and then popping very well because the structure of how they went at it is such that there will almost certainly be that money, uh, left on the table feeling. But I'm wildly excited about it. I think it's greatI mean, it's a wonderful company. It's got compelling numbers. It's a great story of a guy picking himself up after a difficult acquisition and showing the world it wasn't just gonna be, you know, a acquisition or nothing story. So, and, and by the way, I think everyone gets excited once it prints, 'cause it's all just talk now, but the reality of four VC firms each making a billion dollars, that'll focus the mind.
- JLJason Lemkin
You know, it's funny, I just remember September 22, September 2022, SaaStr Annual was in September then, and the, the Adobe deal acquisition happened, and it-- no one talked, 10,000 people talked about nothing but Figma.
- RORory O’Driscoll
Correct.
- JLJason Lemkin
Nothing. Now, this was after the-- It was a fascinating deal because the 2021 bubble had imploded, it crashed, right? And all of a sudden, Scott Belsky and team are our heroes. They're going out in 2022 and paying 2021 prices for this, uh, scrappy startup, Figma, and it was, uh... People's jaw dropped, right? It's like this was the deal of the century, right? Fast-forward to today and, uh, this is their, uh... In my ecosystem, people aren't talking about it. Listen, four VCs making a billion dollars, great topic for, for the pod, right? Don't get me wrong. But, but it's just boring to people today. And it was n- it was... The 10,000 people, executives, their, the, the crowd was hushed in 2022, September 2022. Hu- no one ever... Nothing to talk about. And now let's, you know, I might, I might burn, I might burn six thousand on my, uh, on my cloud credits this, this month. [laughs]
- RORory O’Driscoll
The world, yeah.
- JLJason Lemkin
Lovable did what again? I mean, it's just... It's not that this, you know... Ho- hopefully, this is a $30 billion market cap company for all of our sakes, but it, it's, it's, the zeitgeist has been lost here in traditional software. Even the best of the best of the best. Figma is, is S tier. It's the, maybe be- It's the... But, but it's, it's not capturing our imagination the way it was. And maybe that's okay. B2B software used to be boring. [sniffs] Maybe it's bor- super boring again.
- RORory O’Driscoll
First of all, I just wanna cue the Scottie Scheffler, you know, the, the interview, the five-minute interview when he says, you know, "You win the Masters, and then, you know, five minutes later, it's like, you know, the press is asking me, 'How are you gonna play next week?'" This feels like that. It's like, you know, venture returns $30 billion, and all Jason's gotta say is, "Well, you know, uh, what's going on at Lovable," right? There is-- But I think you're right. I mean, this is as good-
- JLJason Lemkin
It's the vibe. It's the world vibers now.
- RORory O’Driscoll
There's no doubt... A couple things. One, there's no doubt that, um, you, once you announce a $20 billion acquisition, everyone's kinda monthly re- mentally ran the cash register so you don't get the... You know, it, it's odd to take the victory lap twice, but I totally think they deserve it. But I think the more salient fact is what you're saying is, there's still gonna be a few more pre-AI IPOs at scale because there's a couple other companies like that, like Canva and stuff. But you're right, the hype cycle has moved on to the new story on what's going on in AI, which by the way, doesn't mean that massive amount of money can't be made in other areas, right? I mean, I always think, you know... And even in, you know, I, I think if you look back, semiconductors. Semiconductors became uninvestable by venture in about 2004, and in the subsequent 20 years in the public markets, people made out like bandits. It was the best performing sector ever, even excluding Nvidia, right? So, I think what's happening is startup activity in the space that Figma plays is not gonna be a thing anymore. You're not gonna do a non... Imagine, you're not gonna build a non-AI collaborative design tool. You're not gonna build a non-AI any kind of B2B software. So, as you say, Jason, the excitement, the new, new thing is not that, but that's a function, by the way, of it taking 12 years to go from a startup to an IPO. By the time you get to an IPO, and I think we're gonna see this across the board, by the time companies that are 10 year old get to an IPO, you probably are half a tech cycle or a tech cycle behind. It's terrifying, right? You hold... And well, I said this to an LP, "Our holding period is now longer than the tech cycle." That's a terrifying fact. Figma's holding period is 12 or 13 years, and you're now in a world where they, they talk in the, uh, S one about, you know, we're gonna have to adapt to AI. Now, it, it's, it's gonna be an amazing company. It's gonna be great, and I think they will adapt to AI, but it's just a f- this... We're talking here about intangible things, not money things, and I usually prefer to talk about money things. But you're right, Jason. From a buzz perspective, the caravan's moved on, but on the other hand, it's moved on and deposited $20 to $30 billion. I think everyone involved will get over their, get over their sadness very, very quickly when they look at their stock certs.
- JLJason Lemkin
And when Figma nails vibe coding, man, that's gonna be what we all want.
- RORory O’Driscoll
What?
- HSHarry Stebbings
Well, they, I mean, they're go- they're going after it with Figma Make.
- JLJason Lemkin
I know, but it's, it's what we, it's what we really want. We don't really want... Because if you look at the, the Replitz and the Lovables, like, today, and the world's gonna change, and I'm a super fan of, of both, right? Now I've said too my bumps. But a huge use case today is prototyping, if we're being realistic. It's prototyping apps, and it's very powerful and it's very fun, and the pixel perfectosity is terrible. Like, they look, the apps look... You can smell these apps, and they're not what real apps look like. So as soon as I go into Figma and do this, and it looks like I really want it to look like, like a designer, for... I, I, I, I, I'm out. I'm out of anything that's not pixel perfect. I, I just don't want it. Like, I don't want... It's, it's just like a Squarespace site. I don't really want something that looks like Squarespace, and I can smell a Lovable or a Replit app within six seconds of hitting their website. And so Figma isn't gonna disrupt programmers, right, like the way Lovable and Replit is, but this prototyping, I, I don't really wanna do it in these apps. Listen, you, you're closer to it, Harry, than I am, but I can't imagine this isn't a very interesting war in 2026.
- RORory O’Driscoll
You mean, take care of the war between, um, the vibe coding and, you know, design tools like Figma? Interesting.
- JLJason Lemkin
Right now they're great. You can go into Bolt or Lovable and Replit in different ways, and in some cases, in one click, go from Figma back and forth. It's great. But today they're great partners. It doesn't make sense that they will be great partners for prototyping in 12 months. It makes no sense. And Canva saw this early, right? It's just I don't know whether they'll really get there, but they saw it. That's why they went re- tried to go really early into this. They got it, right?
- HSHarry Stebbings
I, I, I think Canva's product will much more compete with Lovable's-
- RORory O’Driscoll
Agreed
- HSHarry Stebbings
... being the much more consumer-friendly, children doing science project, mothers doing businesses. Um, that is the... I think the designer toReally average consumer, bluntly, is a very different product paradigm to design for, and I, I don't see that competing
- JLJason Lemkin
Ma-makes sense that Canva's a bigger competitor at, at, at low end, right? You know, that's why Wix bot Base44 or whatever. But to the ex- But I see on my social... You know, we gotta be careful of X, right? It's a weird cross, cross-section of the world. But I, notwithstanding all the complaints, a lot of developers are using these products to prototype. The folk, the folks that have helped me as a Vibe Coder, these are CTOs. I had a CTO reach out to me at 80 million. He's like, "I just didn't wanna distract my team. I'm building my own app in Replit. Here are my 10 tips." And so that I would rather, like, I would rather do that in Figma, that use case. So it just... The, the comp- as you s- the competition is at multiple levels, right?
- RORory O’Driscoll
Yeah, 'cause some part of the Figma user base is, you know, more picture perfect, you know, full-on built out apps. But some app- some is, some is sh- rough me up what this thing would look like. And your point is, you might wanna do that now using a tool that actually builds you a vaguely functioning prototype.
- JLJason Lemkin
Well, even more, like, I... Listen, I, I... Mad respect for Figma, but as an end consumer, I hate it when I get a Figma, a, a, a Figma link because it's, the thing doesn't work.
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
I, I don't wanna click through from, from, like... This is, this is why, why working with designers, I wanna, I wanna blow my brains out. I don't want static stuff, okay? I want it to... Like, when you send me my Figma, okay, I want it to... Now, I'm not expecting that it works p- in production, right? I'm-
- RORory O’Driscoll
I'm just gonna say-
- JLJason Lemkin
I want it to work. I don't want, I don't want, I don't want design elements. And so as soon as my Figma works out of the box, it's Vibe Coded, man, that's a, that's what, that, then, then that's all I want, right?
- RORory O’Driscoll
And then the fun thing for those guys, and I mean, look, the very best companies run by founders who see this, who've demonstrated, and in my view, these guys clearly have the ability to bounce back from stuff, are gonna be all over this shit. And you're right, 12 months from now it's gonna be f- You know, if, if, if you don't have a billion dollars at stake in it, it's gonna be fun to watch, you know, how, what kind of products these guys roll out to be competitive with parts of either the coding, the cursor or the, um, prototyping marketplace, and how that design all the way through the code marketplace shapes out. 'Cause look, there's a big pot of money here, and everyone's trying to make sure they get their bit, right?
- 1:15:22 – 1:28:02
Kalshi Quick-Fire Round
- HSHarry Stebbings
We're gonna have some fun here, okay? So we're gonna do a Kalshi quick-fire, but I'm just going free rogue on this one, and I tweeted this last night because I was just super freaking intrigued to see where everyone landed. And so the tweet was, "Let's play a game. End of next year, over under."
- RORory O’Driscoll
Under, but close. Yeah.
- JLJason Lemkin
Well, you think M- Harry, help me multitask. We're, we're coming up on a billion today, right?
- RORory O’Driscoll
Yep.
- JLJason Lemkin
And the bet is $4 billion next year, right? Um, what is, what i- what does just my L4M model say? Just trailing velocity say? That, that gets me to almost, that gets me to almost $3 billion or so, doesn't it?
- RORory O’Driscoll
Yes, you get to $3 billion. Yeah, so.
- JLJason Lemkin
Okay. Now, now use my math of, of at least spending twice as much per developer bare minimum, right?
- RORory O’Driscoll
Yeah, you get there. You're right.
- JLJason Lemkin
Um, I gotta, I gotta, I gotta go yes on this one.
- RORory O’Driscoll
So y- y- yeah. Yeah.
- JLJason Lemkin
$4 billion.
- RORory O’Driscoll
You can get there. You're exactly right. I, I, I buy that. By the way, it does... And that, by the way, is also proof. If they're doing $4 billion, and I don't believe the gross margins are as negative as people say, but if the gross margins are 50%, that's another $2 billion coming on Tropi's way. So it's been good for them, too. Next.
- JLJason Lemkin
Yeah, I don't, I don't think the gross margins are negative at scale.
- RORory O’Driscoll
I agree. So even at 50% gross mar-
- JLJason Lemkin
They're def- I can tell you as, as someone who keeps paying these Replit checks, they ain't negative.
- RORory O’Driscoll
I agree.
- JLJason Lemkin
I guarantee you Replit's, Replit's gross margins are pretty good, okay? [chuckles] They're, they're pretty... I, I bet they're in the 60s even today. I could be wrong, but, uh, they're not losing money. They're not... At least at the gross margin level, there's just no... They're not, they ain't... There's no way they're losing money with what they pay, what they charge, right? It's a, it's a high markup to give you this environment, right? It's a high markup on those tokens.
- HSHarry Stebbings
Lovable hits $400 million ARR by the end of next year.
- JLJason Lemkin
End of next year?
- RORory O’Driscoll
Mm-hmm.
- JLJason Lemkin
First, n- I have no doubts.
- RORory O’Driscoll
Yeah.
- JLJason Lemkin
That's easier to do the math. One to 100 in six, six months? I mean, it, that would be a colossal F-up if it doesn't get to 400, as crazy as it is, right?
- HSHarry Stebbings
Where does it price then, Jason?
- JLJason Lemkin
What's that?
- HSHarry Stebbings
Where does it price then?
- JLJason Lemkin
Unfortunately, Harry, it might be to a down round to, to the recent price unfortunately. [laughs]
- RORory O’Driscoll
He's messing with ya.
- HSHarry Stebbings
Cut!
- JLJason Lemkin
Cut. Cut.
Episode duration: 1:28:12
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode yRJMlvxOJ9o
