Anthropic提交上市申请｜Cognition以260亿美元估值融资10亿美元｜996工作制

I think there's a tangible feeling of “grab it now.”

Jason Lemkin

Yeah, I'm not interested if it can't be a billion-dollar position anymore.

Rory O'Driscoll

Losing money is like sex. You can talk about it all you like, but until you feel it, you don't know what it's like.

Starting off, we have Anthropic raising $65 billion and then filing to go public in the same week. We have Cognition raising $1 billion at a $26 billion valuation. We have public markets coming back to life. Is the SaaS apocalypse over? Was that the best earnings week in 2 years?

And then, finally, Uber and Microsoft are now pessimistic about the productivity gains from AI. Is there a question mark coming? What does that do to token maxing and token spending? We are done with the, “I don't want to do the public markets. Staying private is cool.”

All these businesses have gone from capex-light, cash-flow machines to capex-heavy, cash-consumptive machines.

Jason Lemkin

I would quit as a developer if you told me I couldn't use the model of my choice. I would quit. I really do think that by the end of the year, we're going to choose tokens over humans.

Ready to go. Boys, it is so good to be back. This is my favorite time of the week.

1. Anthropic Files to Go Public

I want to start with something we were just talking about beforehand. In the 58th week of This Week in AI, what can we possibly say that's different or provide different commentary on? The question I'm going to start with, Jason, is one that you just highlighted brilliantly, I think: Anthropic files to go public. Is Anthropic filing to go public, and going public, good for the ecosystem or not?

Jason Lemkin

Listen, we don't need to talk about how ARR increased 28% since the last show. It's pretty good, okay? It's the fastest-growing enterprise software startup of all time, of all the universe, throughout past Alpha Centauri. But now it's also going to be probably—certainly—the fastest to IPO at anything near its scale.

This dwarfs SpaceX. It's going to IPO in 5 years—5 years to a trillion. Cursor acquired for $60 billion in 4 years, assuming the deal closes. Why would you bother with most of the companies in our portfolio? Why would you bother to even meet the founders? Why would you do anything as a VC? No—other than spend the next 24 months hunting these.

And as an employee, here's the really tough question: Why would you work for any of these companies? We have the CEO of Ironclad now as the head of legal at Anthropic, right? Or OpenAI. Sorry, I got it backwards. Jason Boehmig—or he leaves; maybe he's still chairman of Ironclad.

Why would you do anything when you can build a trillion-dollar company? It is not impossible to build a trillion-dollar startup in 5 years. Why would you rationally do anything else? Why would you even try to have a $400 million exit or a $2 billion exit? Isn't that just a waste of our time?

I know Rory will pick at this, don't get me wrong, but I think it will seep into our society. I think we will all start to feel this way when the bar—not just for valuations, but for time—is reset. Why am I going to quit and spend a year at whatever? It's always made sense to join the hottest startups, but I think this is going to make it feel, at least emotionally, like people should just quit tomorrow and work for the hottest startups, because the outcomes are 2 to 3 orders of magnitude larger.

Rory O'Driscoll

Yeah, you are going to pick on it because, in one sense, what you're saying is true. Look, we're in the business of investing in the best startups. The best startup is now worth a trillion dollars, and you didn't invest in it, right? What do you do with that information?

It is the best startup in the last decade. You can fool yourself into thinking there's going to be another one just like it next year. That's one option, right? I think it's stupid, because it is, by definition, a 1-in-10. Don't make your business plan around finding another trillion-dollar, 5-year outcome in the next 5 years. I think that's just foolish, for reasons we can talk about if anyone wants to argue it.

The second thing you can do is say, “I'm psychologically so damaged by missing this that I need to go home and I can't play,” which is credible. There are going to be a lot of people who do that.

Or the third is you can grow up and be a fucking adult and say, “I wish I'd done that deal. I'd give my left arm to have done that deal, but I didn't. Now I've got to go on and do perfectly good deals that will have great outcomes.” That's what normal, balanced people who aren't damaged do.

As I say in almost every human endeavor like this, there's one person who gets the big prize. As humans, you have to adapt and say, even if you didn't get the big prize, it kind of sucks and you mourn it, but then you go on and live your life. Only 1 person gets to be president; not everyone quits politics. Only 1 person gets to be the richest person in the world. Everyone else can still play in business.

In fact, I would argue that one of the reasons business is more psychologically healthy than, for example, politics is that I remember a dear friend of mine, many years ago, explaining that he was interested in board careers. The problem is this: In politics, the 700th-most-successful politician in Britain isn't even a backbench MP. The 700th-most-successful politician in the U.S. isn't even a congressman or congresswoman, right?

The 700th-most-successful businessperson is probably worth, plus or minus, $1 billion. That's an okay consolation prize.

The point is—but genuinely, because I do actually hear what you're saying resonate. I'll admit there are nights when I lie awake and say, “What was I doing in early February or March, when the Series C went down?” I'd been to some of the early stuff. I'd seen the thing. I can tell you what I was doing every day on the calendar, because sadly I've looked, right? And I'll tell you what I wasn't doing: meeting with Anthropic.

You can mourn that information, but you can only mourn for so long, and then you get on with the rest of your life. I do think I hear you, and I also think something else: It will be fucking great when it goes public, because then we can just move on.

It goes from being the singularity to a magnificent outcome for everyone. Money flows back to the system. It sucks if you want to buy a house in San Francisco, but it's great. The mystery goes out of it. It's just the 10th or 12th—or maybe 7th or 8th, depending on how it prices—largest public-market-cap company, and we can all just get on with our lives.

So, yeah, I'm good with that. I think you're right. You can't necessarily kill yourself for not being in the Series B of Anthropic.

See, I should point out that the Series B was the Sam Bankman-Fried round. That's a poison chalice on every dimension. The Series C was the round where Spark brilliantly led and Menlo participated.

From a seed investor's perspective, my version of it is—and I don't mean this facetiously; I'm not exaggerating—I'm not interested if it can't be a billion-dollar position anymore. That's how it's changed my mind. I'm just not.

Jason Lemkin

No position. I literally had this review with my fund management company today. They're asking me why I was doing things. I'm like, listen, I will make small investments with friends, for sure. I will do things to be part of journeys, but I'm going to pass on anything where I can't have a position.

At this point in my career, I'm not saying I would have done this on my first check to Pipedrive, which had a billion-dollar exit, but I just don't want to. It's not worth the 20 years. Hopefully it'll be 5 years, but I need a billion-dollar position to get excited today. I need a billion-dollar position.

2. The "Billion-Dollar Position" Era: VCs Reset Their Expectations

If it's going to be worth a trillion, I can have pretty low ownership, right? But you've got to be worth north of $10 billion for it to even make sense to me, given dilution.

Rory, before you chime in, I actually totally agree with you, Jason, but I've also interviewed 1,000 of the best GPs over the last decade, who've all said that their biggest winners were the ones where they underestimated the market size, the outcome, and the opportunity.

You're assuming that you're able to know Twilio is a $2 billion company, which you probably wouldn't have said it was at the time. That is a billion-dollar position to you as a seed or Series A investor. How do you think about accurately identifying that, given that we continuously accept we can't anticipate outcome size?

Jason Lemkin

I think it's a good question, and I've made many mistakes. But I do think the inverse is this: If you see tangible reasons it can't create a billion-dollar position—for example, the founders are very good but not great, or the CTO is pretty good but not a jaw-dropping CTO, not going to launch 17 simultaneous products—if you see that, fine.

The TAM is small, but there just isn't a sense of how to grow it. There isn't that drive. If they're 7/10, if there are complaints, I just haven't seen a lot of great outcomes from complainers, from A-minus CTOs, or from small TAMs.

It's okay to start with the smallest TAM, but I want to see that at least you're thinking—even if it's insane—about the large TAM. So it's more just, to me, drawing a black marker through things where I might have taken a little bit of risk before.

Just know I'm out. I'm just out, because it's not that I can say for sure how big it will be, but these are blockers to a billion-dollar position—not a billion-dollar outcome, a billion-dollar position.

I think you can have a billion-dollar outcome if things are lucky, and you can have a pretty good CTO and a midsize TAM, as long as you get some tailwinds and a few things break your way. I still think you can be lucky enough to have a billion-dollar outcome, but not a position.

Rory, is that not the same for you? Your fund is a billion. I mean, I was just asking for a fund returner.

Rory O'Driscoll

My fund is actually—our fund is $900 million, not a billion, just to be precise. Obviously, you'd love to make a billion dollars, but I think you have to ask yourself: How many of those realistically exist as a base case?

As is often the case with Jason, I actually agree with him on the things he's talking about in a founder. You do want the drive, and you do want ambitious, driven founders, upside, and no complaining. So, in practical terms, I agree with him.

I don't think if I looked at the same deals he was looking at with excitement, I'd say, "I think that these are going to be a billion-dollar individual position," implicitly a $10 billion total fund outcome, because I'm just too aware of the base rates.

When we looked at it 5 or 6 years ago, my mental model—which isn't the case anymore in enterprise software—was that you probably had 10 to 20 billion-dollar-plus outcomes a year. Best case, you probably had 2 to 3 $10 billion-plus outcomes a year. Then every decade, you had 1 to 3 $100 billion-to-$1 trillion outcomes. Now, I'll have to say, you scale that up probably by 10% to 20%, but really, there's not going to be more than 4 or 5 $10 billion-plus outcomes in a normal year. I just don't know if that's credible or reasonable.

I think in that $1 billion-to-$5 billion range, if you own 10%, you're very happy you did it. You made $500 million; you're very happy. It's half of a fund, and you're very glad, right? Especially if you put $20 million or $30 million into it and it's a strong capital returner.

I think, as we've discussed, as you go later, it's much more about concentrated positions. But I don't think you can make 20 to 30 investments in a Series A fund or a seed fund and credibly believe that each of them will be a billion-dollar outcome to you personally, or a $10 billion outcome in total.

Therefore, I think I tend to mentally have the following model: I want to underwrite to a realistic base-case return, but I do agree—never do a deal with just capped return. If you can't articulate it, the way we say it is: You want to have your base case, but you want to articulate a credible upside story that can have that magic outcome.

That's how, as I say, I end up in the same place with Jason, even though we don't agree on the math. You do want to have uncapped upside, but I don't think you go in saying, "I'll only do it if—"

3. The Trillion-Dollar Cash Grab: Google, SpaceX, and OpenAI Rush the Queue

Jason Lemkin

I think the big statement you said there is "will be" rather than "can be." I'm so much more willing to up the risk on doing things that I would never normally have done, because if they do work, they're going to be so mega, versus the "will be"—like V1 SaaS companies, where I can see it much more realistically, but it's not that needle-moving to have it succeed. Do you know what I mean?

Rory O'Driscoll

Everyone's always brave at the tail end of a 14-year equity boom, right? Again, my biggest disadvantage as an investor was that I was investing in 2001, and I watched the NASDAQ go down by 90% and most of our investments go bankrupt. Less than 40% of them survived.

Everyone's always saying, "I want more risk," because the upside is there when the risk hasn't come home and the upside is still there. So, yes, I do think you have to at least be cognizant of the fact. It's the old cliché. I've said it before on the show, so I'll apologize for repeating myself, but losing money is like sex: You can talk about it all you like, but until you feel it, you don't know what it's like.

It's not that I'm challenging your math. It's more that I think for founders—

Jason Lemkin

Yeah.

Rory O'Driscoll

For a lot of founders, after this IPO—after this Anthropic IPO—it may get even harder to get meetings. That's my point. There are meetings I won't take now that I would have taken in 2024 or 2023. I just won't take them. And it's not because they're not great human beings or building real companies. I'm just not seeing that the bar has gone up so much. I just won't take the meeting.

I'm not sure all founders get this. I do think that is real. I get this question from founders, and they're implicitly saying, "Where should I be on the grandiosity-versus-boring scale? If I'm too grandiose, I might lose them. But if I'm too boring, I might also lose them because I'm not aspirational enough."

What you're saying is correct: The base rate for aspirational has gone up. In other words, the level below which you're perceived as quote-unquote boring has probably increased significantly, because people have seen what quantifiably amazing looks like: 10x growth for 3 years.

You're right. There's no doubt it has an anchoring effect, and it will do for some time. So again, I think you're probably right on how you think about deals, and you just go that—

Jason Lemkin

That's why I'm not sure the Anthropic IPO is all net positive. I think it will make things harder. It's not just housing; that's already happening, right? I think it will make everything harder when there's a general sense of not being good enough, right? That is reinforced across the ecosystem.

Rory O'Driscoll

I understand what you're saying, but I think we all have plenty of insecurity already based on the private markets. I actually think, to some extent—again, I'm not going to continue the prior analogy, though it is tempting—we are a PG program, not R-rated, so not X-rated.

I do think that when the mystery is stripped away and the financials are revealed, some of the mystery tends to go out of the deal. I'm actually just looking for—I mean, say it was SpaceX. We can talk about what's going on in the wider world, but it was just great to see the numbers, deal with the facts, and go, "Got it. That's what I thought it was."

I can differ on how they're valuing it, but that all makes sense now. There wasn't any magic pixie dust. It was a great technical launch business, a wildly exciting Starlink business, and, oh, wow. It'll be the same thing with Anthropic: "Oh, those are the numbers. Good to see. That makes sense." And you'll just get the mystery out.

Jason Lemkin

Anthropic goes out before OpenAI now?

Rory O'Driscoll

Well, it's not clear, but I mean, it's going to come down to sequencing. Anthropic said they filed—they made an announcement in the last day or 2, so June 1st. I think OpenAI had made a statement that they were filing in May, around May 22nd. I meant to go back and look at the statement, and Jason is better than me at looking at things in real time. Did they say they were filing, or had filed? I think it might have been filed, in which case they're roughly on the same track.

I think I was saying to you, Harry, before the meeting started, the most noticeable thing here is that everyone is gradually jumping forward their cash raise in the public markets. Instead of, "We're done with the public markets. Staying private is cool," we're fucking done with that, right?

SpaceX had 20 years of being private; now it's go, go, go—$1.7 trillion. Now OpenAI is philanthropic. If you look at their statements over the last 12 months, it was, "We may go public next year or 2. We may go public in 2027." Now it's, "Oh, we're going public."

Everyone's coming in to grab the capital. The other thing, just to chart today, is that there was a ton of announcement-day activity. Google announced an $80 billion capital raise. The most profitable company on the planet, with the exception of NVIDIA, said, "I'm going to need more capital. Better go get it."

So, I think what you're seeing here is that even though Anthropic just pulled off a wildly oversubscribed private raise, smart people on those boards are all recognizing that the scale of the capital required means we should all jostle to the front of the queue.

It's a little like one of those airline flights in countries where they just don't queue. When they open the door, it's just a mad rush to get on the plane, right? It feels like that here. Google and SpaceX were going first. Google just got ahead of it. Google just grabbed the first $80 billion. Some of it is done now; all of it is $40 billion if it's over time.

SpaceX has just formalized its price at $1.75 trillion for early June. Anthropic and OpenAI both said they're probably going to do roughly the same in October. So, you're probably looking at, across those 3 or 4 names, including Google, $300 billion to $400 billion of equity issuance, all of which is really AI-related, given the SpaceX S-1.

Google hasn't done a raise like this in a significant amount of time. I'm forgetting the exact year. Is this merely them being forced to in the capex race that we're in, amongst the competitive set they're in now in AI?

Yes—is it forced?

Rory O'Driscoll

Forced. I mean, in theory, they could have borrowed more.

Jason Lemkin

Yes, forced to. I mean, in theory, they could have borrowed more. There are lots of things. I think they're smart. I think the stock's high, and I think equity is cheap. There's not a ton of downside to taking a wee bit of dilution at an all-time premium. You're getting the world's best investor. You're getting a reputable investor on your cap table, and you're getting another $80 billion, which maybe you can lever up with debt.

I think, stepping back, what it's really indicative of is that all these businesses have gone from capex-light, cash-flow machines to capex-heavy, cash-consumptive machines. Generally, that's never good for stock prices over the medium term. Just across history, things that have high cash flow spinning out are really good investments, and things that eat money tend to be bad investments. Google is right at riding a tailwind.

Yeah, on a trailing-edge basis, it looks amazing. I think it's damn smart to raise equity. Listen, I don't claim to be—I've only been in the conversations a few times. I don't claim to be an expert. Of course, it's smart. The dilution, as viewed from a venture perspective or startup perspective, is unknowable. I mean, it's sizable. It's unseeable.

But if they really thought that payback was so quick, wouldn't you issue debt and have no dilution? Because $80 billion is still $80 billion of dilution, right? To be neutral, they're ultimately going to have to repurchase $80 billion worth of shares to get those shares back, right? So, wouldn't you do debt if you thought you could pay it back in any reasonable amount of time?

Jason Lemkin

Well, first of all, it may be that they don't want to spend $80 billion. They might want to spend $200 billion, and they might lever the same thing. They might say, "$80 billion of debt, another $80 billion of equity," let's say, and you'll take on $120 billion of debt without endangering a credit rating. So, first of all, they could do both.

Right, that's one comment, right? Yeah. And then, you know, it's not clear what the payback is. I mean, again, it is not clear. We can talk at some point about whether the payback is there in the end at all. But even when it is, the payback on these AI data centers is normally 2 to 3 years.

Now, Elon has massively outperformed that with his storage deal with Anthropic, where he's getting all his money back if the deal lasts a year and a bit. But normally, the payback is not nothing, and 3 or 4 years is a fair amount of time. So, I hear you. Look, Google is the second most profitable company on the planet. They could borrow all they wanted, within reason. I mean, they have some debt—I think about $70 billion of debt—so they could have borrowed all they wanted. I just think it's smart to have a strong balance sheet if you find that you might want to spend $300 billion a year for the next 4 years before it comes back.

Jason Lemkin

Well, certainly, it insulates you from any colds or flu you get in the debt market on any given week or month. At least you don't have to worry about the vagaries of the debt markets, which do have micro-panics. They do have micro-panics.

4. Is the SaaS Apocalypse Over?

I think we deserve a milestone award for the shortest time given to an Anthropic section in a trio show, which is impressive for us. I want to move to—I'm jumping around, so forgive me for it—but we've talked a lot about public companies, and often it's been a tough conversation, with SaaS not being appreciated by public companies.

We saw Snowflake, MongoDB, and Salesforce's best earnings in a significant amount of time. All of them did very well, and we saw stock surges across the board. Is this the end of the SaaS apocalypse? How did you guys analyze this? Is this company-specific stuff, or do you think Jason's rules apply, which is either you reaccelerate or you attach to AI spend? The guys who exploded did both, and the guys like Zscaler who had a messy story went down, right?

Jason Lemkin

Well, look, yeah, for sure. I mean, the media, if you just read it in the last 30 days, it's great, depending on what basket you use. I have a slightly more optimistic basket, but my basket of cloud stocks—software stocks—is up 5% this year after today, when we record this, another great day in this run, right? But the Nasdaq's up 21%, and I think semis are in the triple digits or close.

So, great that the overcorrection is over, but the fundamental concerns are all there, right? I think our only learning is that there was—I never understood the total panic of the SaaS apocalypse. It couldn't all be vibe coding, notwithstanding Harry's show, where they all want to vibe-code their own CRM. It didn't make sense.

But the meta-issues of seat contractions, and the fact that AI software spend, according to Gartner, will be up 60% this year—so that means it's got to be cut somewhere else, right?—those issues haven't changed.

I would say we're no longer in freefall, because freefall leads to panic—panic buying, panic everything. But even with the reacceleration, Atlassian is still at 4× ARR, HubSpot 3.8×, Salesforce 4×. So, I know Rory really disagrees with me here, but I just don't think investing really works without 10× or higher outcomes. And I know they're public and mature, but I think the panic part of the SaaS apocalypse is over.

We overcorrected, as we always do, and we may be over-indexing on semiconductor stocks today; that will be seasonal as well. But the issues haven't gone away. We just over-panicked on the timing of them.

Actually, we are in agreement, because I think we're saying the same thing: This was a rare occasion where literally an entire sector was discounted to the point where it made no sense. A month ago, I was able to identify a bunch of stocks where I could say, "Hey, they're so cheap that it's silly," right?

In SaaS land, now you actually have to say, given that we've repriced, which of these stocks has a genuine reacceleration or AI catch-up story. That's a harder message. I think some of the ones who killed it do, but overall, it's still pretty tough. So, I think we're in sync, Jason.

Jason Lemkin

Yeah. I think the interesting thing is there was, look, a modest reacceleration of multiples for almost everybody, right? Even the hardest-hit, like Monday.com and Atlassian, saw their multiples bounce off the hard deck and just go back to crappy from worthless. But the real—the only real learning of this year to date is that it's been long enough now that the public companies—let's just call them mature—the mature public companies that are benefiting from AI are seeing the boost.

So, yes, semiconductors are up, but when Jeff Lawson was on this show, he said, "I haven't run Twilio in a while, but I'm pretty sure we're going to benefit from AI because agents and AI just need to use more of our voice and other APIs." It took a quarter or 2, but he's right. The stock's up 57% this year. It's gone from, I think, 4% or 5% growth to 20%.

Okta, which was your dad's enterprise SSO system, is up 57% this year—56%, right? Datadog, which everyone uses—every AI leader uses—is up 100% this year. So, the captain-obvious learning is, hey, look, we gave it 6 months, and every software leader where agentic products and agents need more of it is up.

No matter what they say, the ones that only humans use are kind of down, even if they bounced off the hard deck. There's just not an appetite for more human-per-seat licenses, even Salesforce, which reaccelerated growth. It was all through hard work. It was through Agentforce. It was through organic and inorganic purchases. It was through everything.

Marc and the team came on the call, and they said—and they split the business up into 2 verticals for the first time—I think Agentforce, or whatever they're called, and the others. They said the software business will be perpetually in single-digit growth.

Like, that's as good as it's going to get here, but we're double digits in the rest, and that's growing 12 or 13%. So let's not—we bifurcated. We saw what made sense, which is AI-fueled, agentic-focused products, but I want you to know, the classic human-per-seat software really is dying. It's not dead, but no one wants to buy this crap. In fact, they're cutting it because they've got to come up with money for all these goddamn tokens. Everyone wants these damn tokens.

Right, Harry. We just don't need another human seat for folks that don't do any work on our project management software. It just has to be the math, right? You can't grow 60% AI software without cutting some material amount of the rest. So we're seeing it, but we should have known this.

We should have all—forget about the way I made my bets—just made the bet on who fell the least. That was at the nadir, right? I bet on who fell the least. We should have all made the Jeff Lawson bet, which is, okay, who's going to honestly benefit from agents and agentic? It's not that complicated, right? Of course, it's going to be Twilio, Datadog. I never would have thought Okta, but if I'd been a little smarter with Claude, we would have figured it out because it is an obvious one. If everyone doing AI is blowing up, and that's a private company, they should blow up too, right?

I think the real challenge is more a question of what critical mass looks like to go public, to Jason's point, what liquidity looks like, all that. I mean, you guys often talk about Replit. There's an example of a company that was pre-AI, that brilliantly attached to the trend and just got a ton of lift, right? There's stuff you can do to get lift, and if you get lift, there's only one test: are you growing quicker? And if you're growing quicker, you've gotten lift. If you've gotten lift, you're fine.

You can imagine lots of parts of the new LLM-first AI-harness software stack. Isn't it Supabase, the Postgres database, that's just brilliantly co-attached to everything? It's grown quickly too. So there will be instances of that, but it requires deft product management and making sure you're attached to the future. If not, it's not as if you're going away quickly, but you just have a question of how do you create value and how do you realize value, which will segue to a few other discussions.

5. Cognition Raises $1 Billion at a $26 Billion Valuation

Totally get that. Can I ask, we said something about explosive growth and software creation? Cognition raised $1 billion at a $26 billion valuation. Devin, their core product, hit $492 million in ARR. Incredible growth, with some mega customers, some of the largest enterprises in the world. Jason, I thought your statement here was a good one. Was Cursor at $3 billion ARR cheap then if that was priced at 60? And how do you reflect on the growth in this Cognition round? I'd love your thoughts.

Jason Lemkin

I wish I was more of an expert on Cognition today. When we started this pod, I did say that probably the 2 absolute highest-IQ CTOs in my portfolio were using Devon in the early days, when things were still pretty crappy. The idea, I think, is still super compelling—super compelling—and now we have some metrics. The idea, at least, is, hey, it's great that your engineers are 10, 100, or 1,000x more productive. What's far more interesting is if you can have an autonomous AI engineer. That's much more interesting.

I remember that probably the smartest CTO in my portfolio, just in the early days, ran Devon in Slack, and he would just tell Devin to go do these things and come back and make the commit on its own. I'm sure it was pretty mediocre a year ago or something like that, just because of the nature of the underlying models. But the vision to me is actually more compelling than all this Claude Code crap. It's much more compelling to just tell the agent what to build—or not tell the agent, but have these autonomous engineers that do it, right?

Why do we want to empower mediocre sales reps? No, we want to automate them with AI. Same with mediocre engineers. Let's get rid of them, man. Let's have the best ones, but all the rest, let's have Devon. Devon doesn't argue. Devin doesn't only want to work on interesting problems, like most of your best engineers. It's not interesting; I'm not going to join.

I was talking to someone who turned down an offer from Anthropic. He just said it wasn't interesting enough. He didn't want to work on, like, little edges of basic application. It was boring, right? So, well, let's have Devin do it, man.

6. Token Budgeting Panic Hits Corporate America

This is very much a market where the lead changes hands at a furious pace. Every one of the companies that's worth a trillion bucks is going to want to eat your lunch. So it's a high-risk, huge-market return, and good luck to them. I love it. It's what it should be.

I was speaking to one of the best CTOs this morning, and he said, "The analogy is we've just given a company credit card to every employee and said, 'There are no limits. Spend away.'" And that's the token-spend budgeting today. My question to you is: if there is rigor and budget instilled, are we dramatically overestimating market size?

Jason Lemkin

I don't think we're overestimating market size. But let's pray. It's funny—we talked about this last week, and I was going to tweet this because literally, on Tuesday when we recorded, I was like, it may seem theoretical to say that maybe this stuff hasn't got an ROI. By the time the thing came out on Thursday, there had been an explosion of these "Oh my God, ROI" articles. So it was right on the cusp of the zeitgeist, where people finally woke up.

I was thinking about it last night, and it all makes sense. Sometime in November or December, Anthropic produced the magical version of Claude Code that just works. In early 2026, they kind of changed the pricing model, so you have to pay as you go. It's like literally everyone cranked it in Q1, and I can almost imagine that in every CFO's office in the land, someone was doing accrual accounting. By midday, they suddenly realized, "Oh my God, we used to estimate our bill based on this, and suddenly we're 10x wrong on our accrual." Literally, the penny dropped simultaneously across the entire corporate U.S. We told these guys to crank in Q1, and fuck me, they cranked, right? It looks like we spent our entire budget, right? And literally, it was universal.

This is about the time I've discovered it. Now, to your point, I was raising the issue when people weren't. But now I'm going to take the positive side, provided the code that's generated is good and useful, right? I can imagine a scenario where people put a pause on it. I just saw—I think Uber announced today—they're just going to give everyone $1,500 a month, which is about the right amount on what we're seeing in terms of averages. I won't say the right amount; that's a normative statement. It's about a little above what we saw the average spend was. So they're basically going to cap everyone and try to get control of the spend that way.

The thing you have in your favor over the medium term is the frontier models aren't getting cheaper. I want to be clear on that: they're actually getting slightly more expensive. But the model that is frontier today will, in a year, be 5 to 10x cheaper because it won't be the frontier model anymore.

So if you're getting value from this egregious spend today, you might slow down for the next 12 months, but as long as you stay on that pricing curve, what will then be not a frontier model, but an older model, will be available at a cheaper price, and you will be able to continue to get value from it. That's a long-winded way of saying it: I don't think you wake up and go, "Oh my God, we're not going to spend any money on this." I think the pace of adoption might slow markedly as people realize quite how much they spent.

You know, I had 2 of my fastest-growing portfolio companies say they already blew through their budget this year. So it's not just the big guys, right? 2 of my fastest-growing companies. None of my slowest-growing companies have said that—not a single one of my slowest-growing portfolio companies has said, "Guys, we've burnt through all the tokens." Maybe I don't know if it's causation or correlation, but it was interesting that 2 of them said that.

Jason Lemkin

I think that at a practical level, of course, there has to be cost containment, right? We're not all startups that just raised $50 million with 6 people, where it doesn't matter, right? There has to be cost containment, and folks are massively wasting tokens. They're massively— they're massively wasting them. So it has to come that people have to be more thoughtful with what they build.

The vast majority of tokens that are used in coding are in QA anyway. It's not in production. We may have to be more thoughtful about how we do that, or not. But I think it's okay if we don't shoot from the hip as often. I think it's okay if we slow down the number of features that we build. I don't think it's the end of the world, right?

Having said all that, more and more folks are using multiple models at the same time, right? For example, not to talk about Replit too much.

Replit does it automatically now. If you have a complex feature, Replit builds it in Claude Sonnet, not Opus. It builds it in Sonnet to save money, and then it has Codex come in and check the work. It has both of them, and you don't even know this if you don't check, because it's mostly for nontechnical users.

One interesting learning, if you want to learn about cost sensitivity, is to study Replit and Lovable, because they're under huge pressure right now from their customers—massive pressure. So, if you want to see the future, we can look at what the Uber dude said in one of my portfolios, but it's much more interesting to watch how Replit and Lovable are radically evolving their platforms, because a couple of extra dollars there leads to churn at the bottom of their customer base.

They're radically focused on cost containment, and yet Replit still runs—I don't know about Lovable, but Replit still runs—2 models. Not for everything you do, but for anything complicated, the architect agent now is Codex, which it brings in to check Sonnet. It works really well. It's incredibly powerful to check your work in Sonnet, probably Opus, with Codex. It's incredibly useful. Every single time, it finds issues. Every single time, it finds issues. It's so powerful.

So that's going the other way, right? As we go more and more multi-agent and ask them to do more, of course we're going to want to, so we have to have budgets. But the tension is only going to grow. And I don't think we've yet—to Rory's point, you just asked my opinion—I think in the course of this show, what we've learned, and it has changed since the beginning of the show, is that it really doesn't matter—and this could change in 90 days—but so far, it really doesn't matter if older models are cheaper and open-source models are cheaper, because we don't want them.

There are use cases where we want them. There are. But overall, as an ecosystem, we are all in on the best. We are all in on Opus. That's what we want. So we're not benefiting as much from the deflationary benefits of AI, and we're paying into the inflationary side.

Rory O'Driscoll

But we're pro—again, I don't want to sound negative because I'm actually a net positive, so I want to come back to that. But first of all, I think the zoom-out comment is this: at some enormously high-level point, this is validatingly good news for the model providers.

What's happened is somewhere around 3–5% of tech spend, everyone noticed, “Oh my God, we're spending this.” And no one said, “We're going to cut back to 0,” right? What this means is you've just established a category that probably has a market size of $500 billion to $1 trillion, and everyone is now going through the corporate process of saying, “How do I find that money? Where do I find it elsewhere? How do I manage it? How do I cap it?” But what they're not saying is, “Stop it,” right?

So you've built a category that's huge. If I'm Anthropic—I mean, it's obvious to us because we're in the Valley—but the doomers who say it's all just going to go away because it's silly: no. Corporate America has said, “We're spending this kind of money. Fuck, we don't like it. We're spending too much, but we're going to have to have a plan to spend it. It's enormously validating,” right?

I think—I don't know if it would be interesting to see, Jason, as part of managing that spend, will it still be true that we all want the most expensive model for everything? I know I see some of my app companies find a way to use multiple models and use open source for the cheaper stuff.

Jason Lemkin

For their application, but for their development, I was going to say: is it a question that there are things where you want 1, but not 2, foundational models, but there are other things that you offload, right? Because I don't think—look, again, going back to what I said—the cost of non-frontier models keeps going down, but the cost of frontier models keeps going up.

And while I just made a positive statement, I don't think corporate America is saying, “Yay, the cost of the frontier model is going to keep going up, and we're good with that.” So I think we've now been found by finance. They're looking to the CEO and saying, “Dude, on January 1, he said, ‘Use all you like. It's going to be amazing.’ And now it's May 15, and we've got a problem. Let's figure it out.”

I don't think they're going to say, “Keep using the frontier model for everything.” You know, not on a price-per-token basis, but on a price-per-pass, on a price-per-run basis, it's been going up. I think we're now going to have to discover costing and marginal costing.

Jason, do you not think that open source will have a meaningful impact, specifically on development budgets? You clearly identified the difference.

Jason Lemkin

Maybe. I just think we're confusing these narratives of 2 things. And I wish I had the exact number, but we're confusing the models used by applications, right, versus models used for software development.

For applications, everyone—I mean, this is OpenRouter blowing up. This is everything everyone's optimizing. Even if you're not optimizing, you're optimizing, right? Because if everything high-end you do in Opus is like a buck, okay? Or it could be more. It could be dollars, and at the low end it's $0.50. You can't do $0.50 for a chat for every single chat, or $1.

This could change, and there are certainly workflows where you need massive amounts of inference and thought, but I think we're confusing this with developers being under pressure. These poor guys, as Harry has said on the show, have to work 996. These poor guys are not going to use a crappy model on Saturday. I'd quit. I can't. No, literally, I would quit as a developer if you told me I could not use the model of my choice. I would quit. It's not worth my time.

So I'm not saying it won't happen in some use cases, but I would quit. I think what’s more interesting to me—from all of us—here's the thing that's interesting to me to tie it all together, and then maybe we can move on.

I think now that we're hitting budget discussions—just discussions, right?—and now that CIOs are more involved, and now it's not just Uber, even though I think the Uber story was a little blown up and a little apocryphal, it's going to happen everywhere over the course of this year, right? The budget has to come from somewhere.

I really do think by the end of the year, we're going to choose tokens over humans for engineering and product. We're at the margin. You're going to go in and you're going to say—because this is certainly the way it worked for me in the old days when I was at a big tech company—“Your budget this year is $50 million.” Instead of your budget being 400 heads, guys, or 200 heads, your budget this year is $100 million or $200 million or $400 million for EPD—engineering, product development.

When I worked at Adobe, that was all humans. We ignored all the other costs because they didn't matter. EPD, your budget was just headcount, right? Everyone cost the same: $300,000 a year. It didn't matter if they were an office manager or your top engineer, because, just to keep it simple, everyone cost the same.

Okay, now we're going to have much more sophisticated budget discussions going into 2027, where your budget is this much. You decide where you want to spend it, leaders, and I'm going to be sitting around saying, “Do I want to have another 20 mediocre engineers on my team, or do I want to give my best guys unlimited tokens?”

And that may fuel the real—whether they're AI layoffs or just AI backfills—it may fuel another wave of this, which is very distinct from the ClickUp-whatever-excuse-for-getting-fit, right? This may be a rational choice at the end of the year. There is only so much money, and I'll take tokens. I'll take tokens over a B. We've already made that choice at SaaStr. We got rid of all our B's. We'd much rather have tokens.

Therefore, the next question is: you have to come to some kind of opinion about what the percentage mix is. People are going to be pushed into some kind of hard choices, and then it will force a belief. Do you really believe that these lifts give you a 20% lift, in which case, assuming no net change in demand, you need—you know, you have a 25% lift, you have 80—you spend $80 on people and $20 on tokens. Do you feel it gives you a 50% lift? I mean, sorry, 100% lift, at 50/50? I mean, it will force quantification, right? Once you have a dollar budget and a set of deliverables, which I think is the next shoe to drop—

You will.

Jason Lemkin

But even more than that, at more competitive companies, the best people won't tolerate not getting what they want. So you're going to look, and what's going to, I think, happen going into 2027 is that engineering QA departments will get destroyed, because you'll be like, “I've got 6 QA engineers. I've got 10, and they're great, but do I really—I'd rather go to 2 with tokens.”

Whatever's left in my customer success department that I didn't put into FDEs, I'm just going to get rid of them for tokens to manage customer onboarding and customer support. I'm going to get rid of all the marginal roles, the ones at the bottom of that list. I'm going to sacrifice them for tokens. It's so easy. It's really simple. The ones on the bubble that weren't already cut in the first wave will be cut for tokens.

It would be interesting to see if you're right. Or rather, you'd have to have—I mean, you know, are you still going to waive testing? Are you going to waive code review? You have to think—

Jason Lemkin

Shrink it to whatever the minimum you can do is, and have the models do the rest.

You will just inherently make that choice rather than someone who’s just okay, right? I will tell you, we had this CFO, CCO, and CS summit at our SaaStr Annual event. It was a couple hundred leaders. Almost everyone was talking about how they’re getting rid of agents and all the rest in their teams.

That was the topic this year: we’re getting rid of all the people and all the software because it’s more efficient to handle this at the agentic level. This was everyone, old and new.

I’m sorry. That’s in customer support?

Jason Lemkin

Customer success. Yeah. I’m just saying these are all roles on different lines that will just get cut at the end of the year. Whatever is left in these departments, I’m not going to cut my best engineers. I’m not going to cut my actually smaller sales team than it used to be, right? I’m not going to cut my best folks, but I’d rather have tokens.

I’d rather have tokens do the inbound call. I’d rather have tokens handle the $3,000 deals. I’d rather have tokens do QA or CS, right?

Jason, with all due respect, I think I liked what you said earlier, which is a nice way of saying, “I disagree with this.” At the app level, for customer support, I think the cost of tokens is so low relative to the total cost that it’s in the noise. I don’t think token intensity for something like customer support will be a significant factor. You won’t look back and say, “That’s a huge amount of tokens.”

I’m just trying to understand the end-to-end product and development life cycle across initial engineering, code review, testing, all that, right? What do you think the split will be between dollars paid to engineers in total and dollars paid to tokens just in that area, leaving everything else out?

Jason Lemkin

I don’t know for sure. I’m just saying, listen, let me go back in time. Let’s imagine I was back at Adobe. I had 400 people in my little BU, okay? At the end of the year, I had a fixed budget, and we got around the room and decided we wanted to go into next year with 300 people and the equivalent of 100 humans in tokens—another 100 of tokens. I wanted it, right? This is what I would do today if I was fixed.

So now I’m going to get rid of 100 people. Immediately, whoever’s left in support that isn’t great: gone. I would get rid of my entire CS team except my head of CS. I would get rid of most of my functional QA team and just leave the smartest guys. I would get rid of all of those people because I need the tokens, man.

Again, I’m just going to be fact-based, right?

Jason Lemkin

I think it will happen at the end of the year because people will make that choice.

What you’re saying—but listen to the math. Of those 400, what you’re basically saying is that every remaining 300 people—you took 100 heads and replaced them with tokens, which implies roughly a 33% allocation. Every engineer who’s getting $200,000 or $300,000 has roughly $100,000 in tokens. That’s what the math would be, right?

Jason Lemkin

Sure, but salespeople will have sales applications that aren’t cheap. Everyone will have agents.

That’s why I don’t want—I mean, I would, but you keep refusing to—I’d like to just keep it to engineering, because I think that the math will be different in sales and customer support. I think it’ll be more app-based, with less token intensity, but maybe third-party apps.

I think engineering is where it’s more interesting. I can’t remember—was it Uber? Someone announced today they’re going to keep it to, as I say, $1,500 per engineer per month, which is $18,000 a year. Call it roughly 10%, right? That’s probably a first-pass swag that says, on a $200,000 engineer, they’re getting a 10% token budget, right?

You imply that on a $200,000 engineer, you’re getting a 33% token budget, which would be—what’s that—$66,000, right? Yeah. I don’t think Uber is one of the greatest software shops out there. I think it might even have been Uber that said it.

My point is this: this is the question that’s going to get litigated this year. The amazing thing is, you can get to the Anthropic and OpenAI trajectory on even 10%, right? This is why I think it’s—

Jason Lemkin

I think you’re having 2 different discussions. You’re absolutely right: I don’t think even just 10% is necessarily enough to fuel their growth, right? I think it may be higher than 10%. It may be higher.

If it’s 33%, then 2 things are true. A, buy at any price in the IPO—just any price. And B, it’s going to be pretty tough, because 1 in 3, 1 in 4 engineers, per your construct, across the entire engineering and product development stack, get replaced, plus or minus any growth that comes from that.

So, yeah, I’m not convinced it is as high as that, just to be clear. But I also admit I’m not an expert, and I don’t know what absolute state-of-the-art token efficiency in engineering looks like. I do know 1 thing: this is the number that I most want to understand over the next year, and it’s the first question I usually ask all my VPs of engineering: how are you thinking about it? What’s working? What’s not?

Because this is the number that will determine whether $1 trillion is a fully priced company that could slow down a little bit for a year while it digests, or whether it’s, “Oh my God, no one’s even going to pause for breath. We’re just going to keep rolling this shit out. It’s going to eat 1/3 of engineering salaries, and that’s going to get you to $4 trillion 2 years from now.”

I just think that this idea of capping tokens, like this Uber thing, is a transitory thing. It’s not a utility. It’s not just electricity, or our density for desks in our office. I think it’s a great thing to do now, but by the end of 2027 or 2028, you should give department leaders a choice, and they’re going to choose tokens in good companies. They’re going to choose tokens over people.

Agreed. I think what’s going to happen is you’re going to give them a choice. But practically, the thing that is so insidiously clever about the AI products that the CFOs are going to be tearing their hair out over is that it’s a product that allows you, as an individual worker, to be and look way more efficient and have it take away a whole bunch of your grunt work. The cost isn’t borne by you, right?

Which would you prefer: crank for the next 2 hours on a PowerPoint to get it just right, or type it into Claude and say, “Make me a PowerPoint that does this, this, and this”? Especially if there’s no trade-off cost, no one’s going to want the restrictions, but every CFO is going to want the restrictions.

The dynamic around that is going to be huge. Benedict Evans gives the example of whether it ends up like cellphone minutes, where you give people big buckets. I don’t know, because the problem here is that you can only talk for so long on a cellphone. I think the analogy doesn’t work because, as you say, as an engineer, you can spin up agent after agent.

I don’t think, at the level of the VP of engineering, you can have a budget and a trade-off. At the level of the individual engineer, you’re going to have to figure out how to empower your best engineers without letting them bankrupt the company. It’s going to involve something, and it’s going to be a moving dialogue.

If I were building a harness company—I don’t like to call Cognition’s Devin the Cursor of this, but having something that was great for the engineers while giving people some kind of peace of mind on the budget side would be interesting. I imagine there’ll be some movement to that.

Jason Lemkin

All I can say on this is, if I go back—if I put myself back in time, when I was a VP at Adobe—if you came to me and I said, “I could have a choice: I keep 400 people, or I go to 300, and my EPD team would commit to tripling our productivity this year,” it was—

By every measure.

Jason Lemkin

Not only that, I can instantly think of the people I’d get rid of. It takes me about 10 minutes. Pre-AI, I would have kept them because I needed someone to pick up the phone, Harry. I needed someone to go to Meta and keep the customer.

But right now, if my team made this commitment to quadrupling output, I could instantly think of 20 to 40 people. Just goodbye. Like, goodbye. It wouldn’t even take me an hour; it would take me about 10 minutes to get rid of them, because if I had to make the choice, I know which way I’d go.

I feel the need to say this: I have this feeling about you that you always instantly know the people you want to get rid of, and it kind of chills me a little. Which of us is going when it comes?

Jason Lemkin

In reality, I’ve almost never let anyone go, because you always needed people. I’ve always been lean. I was profitable at $6 million in revenue in a B2B company. I ran lean, right?

I was impressed with that. No—

Jason Lemkin

But I enjoy it. At the Adobe scale, I would know how to get rid of 50 or 60 of them. It would take 5 minutes, right? A lot of them I inherited.

Let me ask a question, then. You’re the VP of engineering. Let’s just play that out, right? Let’s make it real. You’re the VP of engineering. You had 400 people in your engineering department, and you said, “I tell you what, guys: I’m going to drop down to 300. I’m going to take 100 salaries, turn them into tokens, and I promise to deliver not even 3x, but 1.5 times what I’ve delivered before.”

Do you know VPs of engineering who will say, hand on heart, today, that they know they can do that end to end—not just lines of code, not just pull requests, but shipped product with the features the rest want—with that level of cut? Just curious.

Jason Lemkin

Yes. Yeah.

Okay.

Jason Lemkin

The faster the startup is growing, the more it's true. The slower it's growing, the more they tell you it can't be done.

Great. I think it's totally true for startups. I have them come and say, “Oh my God, I can't believe we can do this with 5 people.” It will be interesting to see, at a large company, can they do that? Because remember, from a budget perspective, that's where the money is.

I agree: our smart startups, with 10-, 20-, 25-year-olds, are cranking and doing more than you could do with 40 people, right? The interesting thing is, can Uber take 1/3? Can Microsoft take 1/3 of their engineering and do that? You're right, we'll see. It will be fun. There will be engineering VPs on both sides of that trade who lose their jobs.

The emails are going to go out: “I'm sorry, you've been laid off for tokens.” This is the next—you think of this as the next—it's all the stuff we've talked about the last 6 months. It's not even interesting. It's theater. It's theater to get efficient. It's theater to free up budget by December 31st. People are going to get these cruel emails: “It's not you, but we needed the tokens.”

Jason Lemkin

We all know in our portfolio there are plenty of old-school folks who still don't think this stuff works. They don't think it's worth it. There are plenty of folks. It's not all age, because there are young curmudgeons and old curmudgeons, and some of the earliest adopters are the most experienced engineers because they're kids in a candy store. They love it the most, right?

But there are resistors to this day. These are the products, like Marketo, that haven't added a feature in 11 years. Good luck to them. I just released a show with Brendan Foody from Mercor, and he said they now spend more on tokens than they do on engineering salaries.

How many people in engineering does he have?

Jason Lemkin

That's a very good question. I don't know. 40.

I think more like 80. But, yeah—

Jason Lemkin

But that's the conceit in those stories, right? Hooray. How many engineers do you have? 1,200? No, 80. Okay, well, then you know what?

But give credit—BFD. No, I'm sorry. Give credit. That's the future that you're envisioning, right? I'm sitting here, just not trying to be a skeptic. I'm just trying to understand: where does it come in? There's no doubt. But if the new companies starting with a clean slate really can do 50% plus tokens, 50% people, and they are successful and able to ship, then that is the future. Everyone else is just a question of how long it takes until you get to that future.

If that is the case, we are underestimating the size of these markets even now. Remember, I also mentioned the EDA software market, which is the most automated market today in terms of tools relative to engineering spend. It's roughly 13%. So, for every engineer you hire, you allocate 13% for EDA. I'm saying 10%; you're possibly saying it could be 100%. In other words, for every engineer dollar, dollar-for-dollar tokens.

That number is the most important number. It's the implicit number in every one of these models, is my point. I freely admit I don't know. It's the old Einstein quote: “If I had an hour to solve a problem, I'd spend the first 50 minutes thinking about the question.” I've thought about the question. This is the question. I don't know the fucking answer yet, but this is the question on TAM.

It goes back to what we said about Benioff, but spending 3.8% of developer salaries—the $300 million that he spends on Anthropic—and whether that 3.8% goes to 20%, because that's a very different TAM ultimately for the model providers.

No, it's just—there's another trend as we think to 2027, 2028. There's another trend the other way, though, which is a big deal, and this is why I think organizations will rebloat up to a point. Because as we're able to launch far more products, more early, far more quickly, it's not just features; it's products. No matter how good your agents are, you need humans to manage the products. I wish we didn't need PMs and all that, but we do.

And so I've got one company crossing $100 million that literally was going to end this year with 3 times more products than it did last year. The EPD team is going to grow larger than I'd like, because this just needs—they're not related. You just need humans to talk to them. Even if you need fewer humans per product—even if you need half the humans per product—if we have 10 times more products, help me with the math, Harry. It's hard to get super lean.

And so, our startups that we're excited about, I think ultimately they will achieve the same historic level of bloat, maybe half the size, but they will get as bloated as they can because they will have much larger, broader product lines. Everyone will be a Rippling with 22 products the first year, and you've got to have 22 PMs to make that work.

Agreed. And just to spell it out, I think this again gets to the number. I talked to a VP of engineering over the weekend who said exactly that. He said, “Look, we're speeding up. We're using the tools. But the problem then quickly shifts. The problem isn't our ability to ship stuff in engineering; the problem is the ability of the organization to turn that stuff into money,” which means productization, product marketing, sales enablement, blah blah blah.

I totally agree. I think there's an interesting academic paper that was cited, I think, just today on the impact of AI on GDP productivity. You have the 2 schools of thought: it'll be amazing, it will grow at 10%; or, where I stand, the statistic—the average over the last 200 years—has been 2%, so it'll stay at 2%.

And then they say, “Why is it going to stay at 2% if this shit's so amazing?” It's exactly what you said, Jason. They said, “Even when one part of the organization speeds up, it doesn't matter if you can make a gazillion pieces of software if you can't package it, price it, sell it, train it.” So it makes this reference to weak links: the weakest link in the chain is what determines the speed of the convoy, the wagon train—in this case, the company.

So, again, it gets back to: it may well not be the amazing productivity lift you think, and therefore maybe the budget won't be as much because you have to spend more on people than you would have guessed. Right? I think this is a real factor here.

7. Big Law Flex: Kirkland & Ellis Pledges $500 Million to Build In-House AI

If we're okay to move on, I do want to move on to the next kind of segment or function to be heavily impacted, one would say, which is legal. We've spoken at length about Harvey and Legora. There are 2 elements specifically that I want to touch on here.

Number 1 is Kirkland & Ellis spending $500 million on building their own Harvey and Legora, feeling they have proprietary data, proprietary workflows, and that they should build their own: $100 million a year over 5 years. This is a big commitment from one of the world's largest law firms and a big slight on 2 of the biggest players, as they're being told, “We don't need you.”

The other point I'm going to make is Jason from Ironclad last night announcing that he is joining OpenAI, and the impending—or, you know, coming—threat from OpenAI and Anthropic in legal, which we will see in the next 2 to 8 weeks.

Jason Lemkin

I don't think the Kirkland story is as interesting as it looks, for what it's worth. Okay, so you've got—yeah, it's a law firm, but it's a law firm that does $11 billion in revenue, growing 20%. It is committing $100 million a year of it. That's probably coming out of their Windows NT box or some other crappy budget that they don't need.

And this doesn't mean that they won't put $20 million into third-party software as well. It doesn't mean they won't dump it if it doesn't work. It doesn't even mean—I don't know—they could be a Harvey or Legora customer too. I just think this is a reallocation of less than 1% of revenue into AI to maybe build some proprietary stuff.

They should do this. Working in a law firm, I think, is one of the most soul-crushing businesses there is. But it's very profitable if you do it right; segments of it are high-margin, and $100 million is nothing to win the deal, right? How much would that be to Andreessen to win a deal? It'd be like nothing. It'd be like setting up a media company to win a deal—a 24/7 media company. It's like nothing.

So it sounds like if Harvey was doing $200 million in revenue or something, it would be a big deal. But they won't even notice it, right? They won't even notice it. It'll come out of the bleeding edge of their LexisNexis/Thomson Reuters budget, for some old terminals that get dust in the corner or something.

I don't think it's a threat. And if it is, it'll make them better. If Kirkland can build a competitive product, then the single-source vendors—that will force them to be even better and say, “Listen, this is like anything in AI: you can do a lot on your own, so the vendors have to do more.” That's just a good thing.

It's not 2023. It's great for everybody that we're under AI pressure. It's great for everybody. Everyone should spool up and try to build their own CRM and see if it's worth it. More power to you. If you want to get rid of Salesforce or HubSpot, go for it. They should do this. It'll keep everybody on their toes.

I think Kirkland & Ellis have already won, because they said it first, so they got all the publicity and their clients are aware of it. They look great. They look AI-forward. They didn't even do something.

Rory O'Driscoll

They said they might, in the future, spend 1% of revenue a year for 5 years, right? So, good move. If that's all they do, they win, right? And generally, if you're in a transaction business, Kirkland & Ellis definitely comes down on the hard-headed, mean-as-shit side of things, and this is just continuing a lifelong trend, right? So, tough call to announce it.

Is it realistic? And the second comment is: Jason's right. Have a go. Knock yourself out. You can do a lot in AI. I think it is hard, in a partnership structure, to build that kind of technology. It's traditionally not been possible, so we'll see.

But I think the other thing is—and again, I don't know, was it Harvey themselves or people talking about them? Forget even the models for a second. Think about companies trying to become “full-stack law firms,” right? I'm not saying Harvey or Legora wants to do that, though you saw some third-party Twitter comments to that end.

It would be crazy to think about that, because nothing could piss your clients off more. I mean, if I was Harvey or Legora, I'd be like, “No, we will never do this,” because what you cannot be is an AI provider to a vertical industry—a vertical knowledge industry—with even the slightest hint or intent that you intend to compete against them directly by being a full-stack provider yourself, right?

To some extent, this might be K&E being a little defensive and thinking, “If I am giving”—and it boils down to the question: This is always the rule on when a big company buys from a third-party provider versus builds something themselves. You buy from a third-party provider where it's a horizontal product, where there's no unique differentiation. You're not giving up your secret sauce. You're not going to be able to monetize it differently by virtue of having that product.

So law firms buy their case management software, their document storage software, and their deposition software. Even lawyers use Westlaw; everyone has the same shit. It doesn't matter. That's not how they compete, right? Fast-forward 5 years: If AI is just like that—yeah, it's a great lookup tool; it's kind of modern Westlaw, modern case management, modern drafting, and it's all virtually the same—then they should continue to buy it from an outsourced provider and compete as they do on the basis of the ruthlessness and relentlessness of their senior counsel, and the willingness to flog their associates almost to within an inch of their lives. That's how law firms compete.

If, on the other hand, this AI can become some level of encapsulation of your secret sauce, which is a little bit of the magic that people are saying, then I can see why people pause before they give that away. If you really think it's giving away your secret K&E sauce, your Cooley sauce, or your Gunderson sauce, do you really want to let Harvey train on that? Even if they say they're not training on that, this is the “Are you giving away the Crown Jewels?” argument.

My gut is, I don't think you are. But I can totally see why the managing committee at an $11 billion firm said, “Hold on here, guys. If we pay Harvey $10 million for their software, but in return for that they know the K&E way, maybe not.” So, there should be some dynamic there, especially if they're also saying, “And maybe there'll be a law firm soon.”

I think it's really fun and interesting to watch this. And then we didn't even talk about, on top of that, if you thought Harvey and Legora were fast and loose with your IP, Mr. K&E, wait till you see what Claude does with your IP, Mr. K&E. Which is why I don't think big-ass law firms are going to be willing—if you're not willing to outsource it to Harvey or Legora, who at least are focused solely on your thing, I don't see large law saying, “I'm totally fine with doing this on Claude.”

So, long-winded answer. There's a lot of dynamics here, but everyone is looking at everyone else's lunch and saying, “I want that too,” right? This is what typically happens when a new technology comes on. I do remember—I'm doing my old “I remember” thing. I remember in the mid-'90s, the story was Microsoft would be a bank. Microsoft would take over Intuit, and then they'd take over your money. They were going to be a fintech provider.

This is what happens when lines blur, and then over time it becomes obvious what goes where. I think the same thing will happen here. Fast-forward 5 years: To be clear, I think there will be AI-focused service providers to law firms. They'll buy the product, and the drama will be out of the deal. Could be wrong.

How meaningful an entrant do you think the AI services legal entrants will be?

Jason Lemkin

My gut again—and this is, by the way, none of this was on the agenda, folks. None of us had time to prepare, but thanks, Harry. I think the answer is this: I think it can be market-expansionary in the sense that, if I couldn't access a lawyer today, at the individual level, I think this is really great.

I'm getting sued, or I got screwed by some big company. I can't afford to get a lawyer. Now I can get an AI lawyer. I love it. Cheap divorce, cheap wills. Explain the facts, explain the circumstances. I think there's a ton of additional demand for legal services that can't be met by ordinary people that will be met by AI, and that's freaking great, right?

Same thing for small business, right? I'm a 10-person contractor. I get a document. I can get decent legal advice for $100. I can't go to a lawyer for less than $2,000.

I don't think a full-stack AI law firm will replace K&E, right? What you're getting from K&E are—even the nicer ones, the Wilson Sonsinis, the Cooleys, the Gundersons, the guys out on the West Coast—you're not just getting the knowledge; you're getting the whole experience, which I don't think you can encapsulate.

I don't think full stack goes all the way. I think it takes—I think you still need the human. Let me tell you this: When you're doing a $20 billion transaction, at some level, you want a human to, frankly, as a CEO or a CFO, hold your hand and tell you, “These are the last 10 of these I did, and they're going to work, and this is why this is legal.”

So, yeah, I think K&E would be just fine.

Thank God K&E can still sponsor the podcast. It would be a bit awkward if you said they were shit.

Rory O'Driscoll

You did just say they were like the mean guy.

Say something.

Rory O'Driscoll

You'll always pay the premium for that high-level judgment on mission-critical things, which is why these jobs are terrible, right? This is why they're critical, because you just want—if you're a young associate, you just want a 40-hour-a-week job, but everything you work on is goddamn mission-critical to the client.

The $60 billion Cursor acquisition, the SpaceX IPO, whatever, the bankruptcy—the stress. That's why these lawyers can charge up to $10,000 an hour now, because the commodity services we do in Claude—but $10,000 is nothing on a massive, $10 billion, $20 billion, or $100 billion transaction. It's nothing. You need the guy. I want Rory on this deal. I mean, that's who you want, right? You want Rory.

I totally agree with your comment, and it reminded me of way before generative AI, when we looked at some of the AI startups in mid-2018 and 2019—natural-language startups. We were evaluating a really interesting one, and we had this young graduate from Stanford who was an associate at a big law firm. We just said, “Hey, I'll pay you a bunch of money over the weekend to crank and use this for 5 different things.”

She came back—she's really smart—and said, “Look, I tested this. It's 98% accurate. This is really impressive.” She said, “I wouldn't touch it with a 10-foot pole. My boss will sack me if I'm not 100% accurate. I have no interest.” It was exactly what you said, Jason: “I'm getting paid to get something right. That's a $100 million or $500 million transaction. I have no interest in this.”

So, at the high end, I totally agree. I think you're going to have that human in the loop. And besides, I was going to say something else: Given K&E's reputation, I doubt the Anthropic safety committee will allow them to build a model quite as mean as your average K&E bankruptcy attorney. I literally think it will fail the safety test. The ethics are just too mean.

And there we go. That's the partnership gone. K&E is now no longer partnering with 20VC.

Jason Lemkin

They'll forgive you. As far as they were concerned, they're going to put that in their advertising material. That's the product they're selling, dude. They announced earlier this year that they paid every partner an $11 million bonus.

Rory O'Driscoll

And they didn't do that by being pussies when it came to it. I mean, they were famously aggressive in bankruptcy, to the point where they actually had to step back on some stuff.

Jason Lemkin

Also, we all like to talk about 996 and work ethic. Oh my God, these guys work like we don't see.

No.

Jason Lemkin

I mean, maybe you do at 20VC.

Rory O'Driscoll

It's a terrible job. It's the worst job there is relative to the pay balance. I don't know whether it's a 2×2 or 3×3: the worst job for the most money.

Which is at least better than the worst job for the least money.

Jason Lemkin

Yeah, there's plenty of those.

Guest

Bottom left.

Jason Calacanis

It's top left: high money, low happiness. Top left of the 2.

It's top left. It's okay. It's okay. Winning out.

8. Robinhood's AI Move: Automating Financial Planning vs Beating the Market

Guys, I want to open up. Are there any that you think are really important that we hit on? Personally, I think Apollo and PE software returns being disastrous is quite a statement, but I don't want to guide you all. If there's one...

Do whatever you want.

Jason Lemkin

I do think—listen, maybe out of all of our collective skill sets, it is in my interest area. I do think Robinhood letting AI agents invest for you, if it really goes to the nth level, is pretty interesting. I think everybody should be—I mean, it's good that they are exploring the limits of what agents can do, because everybody should be doing this: What can your agents do, right?

And just so I understand, do we not just see the commoditization of trading? Because if everyone wants to make gains and all the agents are going and trading in the same way, how do we think about that?

Jason Lemkin

No, I think—I'm being optimistic—I think Rory will be with me on this. There's a version of this that's like Wealthfront, but what we really want is that you talk with your agent and say, "This is exactly what I want. I want this risk profile for this amount of time. I'm this old. I have these expenses coming up. I want to buy a house in 3 years."

"Okay, I'm willing to lose up to 18% of what I have, but more than that is stressful. I make this much from my job." There may be ways that an agent—I think wealth management is the worst, the lowest quality, of any professional service I've ever worked with. Humans in wealth management are terrible. They all put you in the same crappy models and come up with the same 11 proprietary products they want you to sell.

I think Robinhood's agent has the potential to leverage the best of AI to really do this dream of giving you the right answer, because no one understands finance well enough to answer these questions. I don't. I need a product. I have a certain amount of cash; it's in my bank. I have a certain amount of public stocks that have done pretty well this year. I have carry coming. I have homes. What the fuck? No matter who I talk to, I'm fucking guessing what to do with this crap, right?

I want an answer from AI. I asked Claude, but I would love Robinhood. I'm probably not the right fit for Robinhood. I would love the right answer for every single individual. So many folks will not get ripped off if we can nail this for everybody. Fidelity doesn't do it. Vanguard doesn't do it. None of them do this.

You know, look, I think we actually made an investment in a company, Range, that does this for kind of the low end of the high-net-worth market.

Yeah, we talked about that. Yeah, yeah, I like that.

Jason Lemkin

We talked about that. And I remember you giving me shit for it, but I think—

Rory O’Driscoll

Oh, Harry did. I said I liked it.

The interesting challenge about this business, I would argue, is there. That's why the Robinhood thing's interesting, because I do disagree with one part of the thing you said, right? You made a comment: no one knows how to give that financial advice.

The truth is this: it's pretty widely understood what the correct financial advice is and what the correct portfolio allocation is. A lot of this is actually just getting the information from the client, understanding the specific circumstances and, as wealth managers will tell you, helping the client to stay on the straight and narrow and literally not letting them do crazy stuff.

I actually think knowing what to do in financial management at a macro level is pretty well understood: risk allocation relative to net worth relative to goals, right? Which is different from actually managing money or picking individual stocks, and I would utterly separate those 2. The big-picture asset allocation and financial planning decisions can be automated, should be automated, and are knowable, right?

I think LLMs have a really meaningful role there, right? And I think for sensible people who think in terms of asset allocation, it will be great. That's why we made that investment. I'm sure Robinhood could do a similar version of the same thing.

The thing LLMs have been somewhat unproven as of yet is the ability to actually trade stocks. You'd basically be a pod manager in the hedge fund world and outperform humans. The record on that isn't there yet. To me, that's a less interesting problem, even though it's where all the drama is associated, right?

I think Citadel is going to use LLMs, but I don't think they're going to replace people with LLMs just yet, right? The fun thing is, I just didn't think of the Robinhood demographic as the people focused on long-term planning for retirement. It'll be interesting to see how that meshes with the trading-as-entertainment part of the Robinhood product, and maybe, as those folks grow up, they kind of grow up with them.

I watch my son trade his Robinhood account. I don't think he's focused on where he'll be at 65.

Jason Lemkin

For sure. Let me just say one thing that I really like about—

Rory O’Driscoll

Focus on where he'll be at the close of market, for fuck's sake.

Jason Lemkin

This is what I like about it, though, and this is where Andrew Bialecki, CEO of Klaviyo, came to SaaStr Annual this year and talked about what they're doing in AI. The first one is how they're building software and their harnesses. But the second thing he said—in a quick aside, it went over my head because I didn't know it was on the agenda—was basically, "We have these AI agents. We have them for marketing and support and all the things they do at $1.4 billion in revenue, but the most important agents we have are there so that every single person using Klaviyo now is a true expert, a true expert in marketing," which wasn't possible before AI.

So the idea that I can go into any account, whether it's Robinhood or Morgan Stanley, and be an expert in whatever—I can't tell you anything. I actually think that the YouTube agent is really good. It tells you everything about how your YouTube video performs better than any human could. Try it if you haven't used it. It's amazing, right? It has access to data you can't see and isn't expressed, right?

All applications should make you an expert in their domain and in their product. An expert—and literally, I mean, I have so much money in different places. Every quarter, Morgan Stanley tells me I need more private-equity exposure. That's their insight. Does that align, though, with what Robinhood is doing, which is basically allowing you to not be an expert? It's allowing you to say what you want, and it does the expert work for you.

Rory O’Driscoll

I think they're related, right? Whether it educates me or whether, because I'm not educable, it executes for me, I just think it's a line of autonomy in agents, which is a big discussion. How much is it education versus autonomy?

I think the aspiration is that your product makes all of your 10,000 customers, 1 million users, 100 million users truly experts in your domain. I think this is something, as executives and founders, we should aspire to: that you log in and, on the first day, I'm a fucking expert in sales, marketing, CS, engineering, product, whatever. That day I should be an expert.

To be clear, an expert advising Jason would be someone who looks at the totality of his holdings and says, "Jason, you do not need more private equity. You got a ton of risk here. You need—you need whatever." And to me, that is expertise that's available and should be available to everyone. That's the kind of idea that these things should be doing.

And again, I'm saying it again for completeness: if you then tell the agent, "I want you to outperform the S&P by 200 basis points by trading stocks aggressively. Figure it out," that agent cannot do that, because that task cannot be accomplished by that agent. I just want to be clear.

Financial planning can be done much better with AI and with agents, I think. Actual trading—the record on being able to perform, I mean, is not there yet. And, to state the obvious, given the power of LLMs, given the power—and given its potential, given the way people like Jane Street use this stuff—if there was an edge, they'd be doing it.

Some of that millisecond trading stuff, yes. But is AI going to give you a meaningful opinion on whether you should hold Microsoft or Apple for the next 5 years? It might make you more informed than if you didn't ask it, which is why, to your point, Jason, you'd love to have everyone, as they're going to trade, served that information. That's great. It's not clear yet, based on actual trading performance, whether or not that answer will be better than a random-number generator or the rest of us.

But move on, because if it was, someone would fund one of those companies and wouldn't tell anyone. I think you can create alpha that didn't magically exist for every Robinhood customer, for sure. Maybe I'm indexing on something that's less important than it is to Robinhood, but this idea that you could—listen, maybe we're all just going to trade GameStop and SpaceX up to $5 trillion, but at least the agent can make me clearly understand what I'm doing.

"Okay, here's the risk. Here's why it doesn't work. Here's the historical dispersion of similar things over the last 1 year, 5 years, 10 years. If you want to do it, that's great, but let me tell you about a few things you haven't thought about, Jason." That would be epic.

That would be epic. I agree. And it would help things.

9. Apollo Warns PE Software Returns Are About to Be Disastrous

Jason Lemkin

We can do it today. I might have to get rid of 100 employees on my team to get the tokens there, but we can do it.

Final one before a rage-bait, but real.

I have 2, so we can choose which one you think is more impactful. I do think Apollo saying PE software returns will be disastrous is very impactful, given the percentages of the portfolios of some of the largest allocators in the world. And then, tied to that, Harvard saying that 41% of their book is now in privates. It’s a very high number.

Rory O’Driscoll

And, I mean, yeah, Apollo, as always, was talking their book, but they’re probably correct. If private credit, which is the senior lender to a whole bunch of PE-based deals, is struggling because they were the first, they’re half the consideration, and they were at 5x EBITDA leverage, they’re worried. Then the PE guys, who are from 5 to 10—in other words, the equity is below the debt in the stack—if the debt’s in trouble, the equity is dead. These are all the SaaS companies that we’ve been talking about for a while, and we all agreed that they’re not dying.

They bounced 30% in the last month, but as Jason points out, they’re still trading at 3x, 4x, 5x, 6x. If you bought the thing at 10x 3 years ago, and now it’s grown a little bit, but you’ve had to pay some debt, and now you’re at 6x, it’s just very hard to get out from under that. I want to hedge that: maybe it won’t be a total train wreck, but maybe they’ll have to own them for 10 years, do a whole bunch of bolt-on acquisitions, and grind out a miserable 1.2x or 1.3x.

Yeah, it’s hard, because in companies growing at 100%, you can overpay and get saved. In companies that are growing at 20% and then suddenly slow down to an 8% or 9% growth rate, if you’ve overpaid, it’s kind of like overpaying for a real estate transaction. There’s nothing you can do. There’s no accelerant; there’s no magic that’s going to happen. You just own a mature SaaS company.

If you step back, if you bought Salesforce at 14 times revenue in 2021, congratulations: you own Salesforce. You paid for it—half with equity, half with debt. So now you have 7x on the debt and 7x on the equity. You now own Salesforce. The public market thinks it’s worth roughly 5x or 6x revenue. You’ve got some growth, to be fair, but your equity is challenged. That’s all he’s saying. The math is pretty harsh.

Jason Lemkin

I mean, the LPs are going to be so excited to get their Anthropic distributions that they have to give you a pass on all these. To Rory’s point, the point Apollo is making is: look at the debt struggling; the equity has to be worse. You’re just not seeing it. That has to be true.

But we’ve got to move on and not care anymore. We’ve just got to move on and fight the next battle. I do think these distributions will facilitate us somewhat ignoring maybe some bad funds. Just move on. It’s life, right?

Rory O’Driscoll

I think if you’re the LP and you have a diversified portfolio, you have to move on. But if you’re one of these PE shops, the whole way it should work—and it should work—is you don’t get to just, quote, “move on.” You have to, as part of your management fee, spend the next 5 or 7 years, because there’s a big difference between giving up and getting a 0.5x now and grinding it out and getting a 1.5x.

You’ve got to find an exit somehow that gets you there.

Rory O’Driscoll

You’ve got to make it happen somehow. Do it right. And that’s why I always say to LPs, I think capital commitment really matters, because if your guys don’t have skin in the game—and especially if they don’t want to raise again—they’re like, “Not me,” right? Whereas if they’ve put in, as Peter Thiel put in, in a very different, very positive fund, hundreds of millions of dollars, they’re going to sit and make it happen.

So it will be interesting to see how the PE firms deal with that, or even the venture firms. If we all have 1 difficult fund, how do you respond to that? It will be a function of: are you playing a multiyear game? Does the next thing look good? And what are your economic incentives in that fund? Do you have capital at risk?

You mentioned distributions from Anthropic. I think one thing that will be interesting is just how several firms deal with massive distributions in terms of team retention. Menlo will make $10 billion in carry. Spark will, too. Plus, Founders Fund will make more than that from SpaceX. When you have such huge amounts of cash coming to a team, humans are humans: they go off and do their own things. It does change the structures of firms.

Jason Lemkin

Okay, but so what? I mean, I guess it’s interesting, right? Of the last generation, you’ve got OpenView, which, after its Datadog and other money, called it quits, right? They just didn’t want to do the AI thing. They’re all multimillionaires, right? Especially the guy that founded it, right?

You even have our friends at Emergence. Most of them called it a day after becoming almost billionaires, right? Not all of them, but everyone but Gordon retired or did their own thing. I don’t think every VC firm has to last until the 23rd century. I think it’s okay if some of these glorified institutions—or if some folks at Menlo want to quit. More power to them. Who cares? What does it matter?

I think some of these folks really are in it for the love of the game, right? I mean, what’s Peter Thiel’s point otherwise? He’s too rich, right? And if you’re not, then retire. If you don’t love the game, retire when you make 8 figures. Just leave when you have 8 figures. That’s the simple math, isn’t it?

Rory O’Driscoll

When people make money, the thing it does is allow them to be what they want to be. Some people are like, “I want to go back to work the next day and do another deal.” And some people are like, “I want to teach high school,” right? And go team. It’s wonderful. Everyone has that chance to do that, right?

And look, I think that it just reveals preference. A few firms are going to make an awful lot of money, and more power to them. That’s the way the system’s meant to work. Some people who make a lot of money say, “I’m done,” but there are a lot of people who can keep going and enjoy it. So I don’t think it will be as impactful in the way you said it. I also like what Jason said: you’re right, if everyone decides they don’t want to do it, then don’t do it.

Rory, if I gave you a $10 billion carry pool, would you come in tomorrow?

Rory O’Driscoll

Absolutely. I like the job, because the real truth is, the terrifying fact—less so at my age, but especially at a younger age—is there’s nothing more terrifying than getting that kind of sum and then not having anything to do with your life.

I always tell people to be very careful of large amounts of money and large amounts of free time. They tend to be pretty destructive, especially in your 30s, 40s, and 50s. It’s hard to fill your day with, and there are so many things to do around. So, yeah, I’d like to keep working.

But again, someone else might decide, “No, they really want to save the whales or save the planet,” or run for governor of California. All these things are possible, right? Or even run for governor and fail, and just spend a lot of money trying. I mean, political consultants, thank you. It’s trickle-down in action.

Jason Lemkin

I just think, going back, it doesn’t even matter. I think people got the OpenView story wrong, too, just like the Kirkland story. This was a rational look: “Guys, we’ve made more money than we will ever spend in our lives, and going forward in venture, we’re not excited about what this takes.” And so they’re living their best life, right? I mean, that is a rational decision for most human beings.

They returned a lot of their fund, right? This was not struggling for 10 years to raise a $100 million Fund III. This was, “Guys, we all made 9 figures. Maybe one made 10, and it’s time. It’s enough already of this venture stuff. Like, enough of these unappreciative kids.”

Wow. Tough crowd.

Jason Lemkin

But wherever it came from, it was clearly an intentional choice. Not for the next generation, right, who kind of got kicked to the curb, but for the founding managing partners, it was a very intentional choice.

The real problem if you have a very large distribution at a venture fund is that, for most people, even if you want to keep going, you might be worried that the next distribution just can’t be as large. Is it worth it? When I went into venture, it kind of pissed off the folks I worked with. I said, “I’m only willing to do this if I can make 10 times as much as I made as a founder.”

And I said, “I don’t care about—I really don’t care about money, right? I already made enough. I have my houses and cars and whatever, but I don’t see why I want to do this for the next 20 years if I can’t make—if I’m going to make 0.4 of what I made as a founder.” It’s got to be 10x to be worth it intellectually, and I don’t even care about money.

So if I made a couple billion in carry—which I haven’t done yet—and I’m looking at my next fund and I’m like, “God, for 20 years I might make $20 million from that,” I would quit. I don’t. More power to the young kids. I would give them the keys, the code to the office, and tell them to keep all the fees and have fun. But I ain’t going to do it for a fraction of what I made on the big win, right?

But just as a reminder, I can give you the quickest way to make 5 times more than you made, right? Which is to be the LP as well.

Jason Calacanis

I thought it was an Anthropic SPV, but okay. [laughter]

Well, my point is, yeah, you can just invest more of your capital.

So, there is a solution to your problem, Jason.

Jason Lemkin

Well, that's what Peter Thiel did, right? He's 1/3 of the fund, right? As always, when Peter Thiel does something, you should assume it's the entirely rational, coldblooded, correct solution. If I have so much money that the marginal utility of the next deal is so low, and I'm only getting 1/4 of maybe 1/2 of the carry, then the only way to solve that is to get 1/2 of the carry and 50% of the LP. Now it's suddenly much more interesting.

10. The 996 Work Ethic: Performative Theatre or Startup Reality?

If you like the business, you can put more money into it. Even if you don't like the business, you can go buy a football team or whatever it is you do, or play, or whatever the other things people do.

Okay, final one. Rage bait, but real. You can kill me for this one, guys. We went and did a show with Nico from Gorgias. They work 7 days a week, and they have a 24-hour café. It's a very intense work culture, unlike any I have seen before. To be fair, the company scaled to a $2.5 billion valuation very quickly and has been very successful. I'm not going to pick on Gorgias. I don't want to.

But I'm just asking: in your best-performing companies, are you seeing a different level of intensity and work ethic than you've seen in prior cycles, or is this just rage bait?

Jason Lemkin

Can I simplify it? My learnings—I wrote this on Twitter. My very first startup job—I had never worked at a startup before—and I rolled into the office on Saturday at 9:00 a.m. It was me and the co-founder.

I said, “Well, I've never worked in tech before, but in all of my service jobs, I worked 6 and 1/2 days a week. We just didn't call it 996. I had to work 6 and 1/2 days a week before I worked at a startup.”

And he said, “It's so great to have you here. I haven't seen anybody in the office on a Saturday morning in a long time.” So, he was there, right? The founder was there. As a founder, I worked 7 days a week.

I think with 996, we're getting confused. There was a while in late 2020 and 2021 when no one really worked, but in general, it's just a question of how deep it goes in the organization. How deep does working Saturday and Sunday go?

While many folks think it's toxic, if you're trying to build, I remember what the Cognition guy said. What's the CEO's name, the Cognition guy?

Scott Wu.

Jason Calacanis

Yeah. When they acquired Windsurf, he said, “We're letting a lot of the folks go, and it's because we work 7 days a week.” He didn't say it in a douchey way. He didn't say anything—I thought that was very thoughtful.

If they're worth $26 billion, or wherever we started the show, and they're all going to make $40 million or $50 million, I think for the first 50, the first 100, it may be okay today to have certain expectations. But you better deliver them. Back at Gorgias and Cognition, you better—you better not. A $150 million exit doesn't justify that, right? So, there has to be a quid pro quo.

I do think it's just a question of how deep in the organization and for how long. I don't think it's as toxic as—or, frankly, as new as—the world makes it out to be. I don't think it's as new. I think there's a performative element, even with the Gorgias guy making it sound so new.

Just do it, man. Hire those people, pay up, give them 4 times the equity. Make them all have 20 co-founders—you know, they're not really co-founders, but give them the equity—and tell them, “This is what we want. If you don't like it, go work somewhere else. It's cool. There are a lot of companies.”

Totally agree that it's not new. The truth is, there are different jobs. They pay differently and have different levels of responsibility, risk, and intensity, right? You can pick where in that thing you want to be. I'm not making a judgment on you, right? Different folks are moved by different things.

You're right: startups consistently have an intensity significantly higher than most companies. I was reading *Apple in China* and then the 50-year history of Apple. This is not new. It was brutal there. They talk about heart attacks. They talk about the pressure, right?

Unfortunately, sometimes to do really hard things, you need small numbers of people to concentrate 24/7, pull their resources, pull their minds, and just will it. True. It's not sustainable for 50 years of your life, right? It's just not a way to live.

But I do agree with Jason. We're joking about big law. Everyone at K&E and all these places bills 2,100 or 2,200 hours a year. That's exactly that math. Some people choose to make that trade in return for the success of the company.

So, I agree: don't be so performative about it. Don't be toxic about it. Just be realistic about your expectations. Most founders will do that, and most founders don't even regard it as a punishment. Most founders regard it as the thing they most want to do.

I love work. They're like, “No, I don't want to go to the ball game. I just want to work.” That's why it's their passion. The first 50 people who are doing it all-in will do it that way. When you get to 2,000 people, have a large organization, and have to hire folks with lives and additional interests, you're probably not going to have that same level of intensity across the board.

I'm with you. It's nothing new. It's normal in its unnormalness. In every generation, there are places like that. It's a small percentage of the total workforce, because most folks are doing different jobs at different intensities.

But, yeah, go do it. You're right, Jason: you better deliver. There's nothing more sucky. I had my own startup, but it didn't work out. I look back and I worked 7 by 24 for 3 years and made no money. That sucked, right? As a reminder, that's the modal experience, in the correct sense of statistics. That's the single most likely outcome.

Jason Lemkin

If you're implicitly promising 8 figures to these early employees, then, sorry, you have to think about this every minute. No one who's wildly successful, no matter what they say, is thinking about their company every minute. They have distractions. Maybe they own a sports team or 2, but you've got to be thinking about this every spare moment. All your energy has to go into it.

You can expect that of more people, but you better give them 8 figures. You better give them a shot at 8 figures, right? You better give them a shot at 8 figures.

One thing you have to watch is that you don't over-devolve into weirdness, bad thinking, and losing your judgment. I find that when you're working really intensely, stressed, and caught up in something, the good news is you put in an extra 10 hours of effort. The bad news is you lost your judgment in doing it, especially if part of your job is a judgment job.

You need to step back, go out, cut the grass, take a walk, and make sure that instead of rage baiting or rage working, you're not just performatively working and not achieving. It sounds weird, and I'm not hippie-dippy, but you have to make sure your psychological health and judgment are good.

No, for sure. Listen, it's a marathon, not a sprint. Unfortunately, we've replaced you with tokens, but it is a marathon and not a sprint. Both are true. Unfortunately, there's good news and there's bad news.

We agree culturally: it's a marathon, not a sprint. But we need the token budget for the folks in the office working 996. We just need your tokens.

By the way, to your point here, it is bizarre, and we've grown to accept it. We're all here in the Valley with a plan to automate white-collar work such that there's going to be mass unemployment, according to these folks who are totally wrong, in my opinion, in 3 years. All the work will be done for us by agents. Yet you talk to every single person in this Valley, and they're like, “I've never worked this hard. I'm working 24/7.” The contradiction at the heart of it all is hilarious.

Jason Lemkin

And also, my number one problem is hiring and recruiting the best talent.

Yeah, you can't get people, and you have to work 24/7. But, by the way, we're going to automate all work and it's all going to be fine.

Jason Lemkin

Maybe. It's hard to predict. It's just hard to predict. I think it's actually pretty easy to predict that it won't happen.

Yeah, it won't happen. I think things will be great, but it's all a little delusional. Things will be the same as they have been for the last 200 years. I repeat myself: 2% real GDP.

Jason Lemkin

We'll see. Listen, we could go on forever. I think you can't predict because I just don't know, being objective, what happens with the Bs. The As—everyone can't hire enough, and they're worth more. We absorbed so many Bs in tech, and then we got full of them. I know you think there are going to be plentiful jobs for them. I'm not convinced. I'm not convinced.

I remember many years ago, I had a CFO at one of my companies. She was fun, hard-nosed, and she had her quirks, right? But then she pointed to another ex-member of the staff who had gone on to something else and said, “I look at him,” she said, “and he's got a job. As long as there are people willing to hire idiots like him, I'll be okay.”

It always stuck with me. I look at the world and hear they might—I mean, the truth is, people will get jobs. They might not get the high-status jobs. It sucks, but people will be repurposed. Maybe you won't get another job that pays $400,000 and allows you to work from home 3 days a week. I think you will get a job.

I'm not in the “the Bs will be doomed” camp. The Bs might just have to recognize that there was a moment in time when they got wildly overpaid. Life will go on, and maybe they'll be happy doing other things. I'm benignly okay.