Jack Dorsey Block Tech Layoff Tech Sellof Artificial Intelligence

I'll be very curious to see if we fast forward 10 years, were these billion-dollar inception-stage companies good investments? They might turn out to be good companies, but were they good investments? That's something I always come back to is a good investment is an investment that makes you a really good return relative to the risk that you're taking. There can be a good company that's a bad investment for you.

It's the week of March 2nd. Welcome to The Learning Corner, a Precursor Ventures podcast. Welcome to the Learning Corner by Precursor. I'm Mia Farnam, and I'm joined by our GP and managing partner, Charles Hudson, where we will be talking through the articles and Goodreads that caught our team's attention over the past week. For our first piece this week, an Anthropic researcher named Mironk Sharma published his resignation letter publicly on Substack, where it then spread quickly across different channels. He spent 2 years at the company after finishing his PhD, contributing to foundational safety work, including writing one of Anthropic's first AI safety cases. What drove him out? Wasn't dissatisfaction with the work itself. It was a deeper reckoning with whether the company's values were actually governing its actions in place. Mironkh's departure comes right as Anthropic made a major policy reversal. That's worth layering in here. Time reported last week that Anthropic is walking back the central promise of its responsible scaling policy, a commitment it made in 2023 to never train a new model unless it could guarantee adequate safety measures were in place first. The company is now framing that position as untenable, given that competitors aren't making the same commitments. And federal AI regulation never materialized under the current administration. Their CSO, Jared Kaplan, put it plainly, we didn't really feel with the rapid advance of AI that it made sense for us to make unilateral commitments if competitors are blazing ahead. Mironk didn't name this directly, but he writes about repeatedly watching the organization face pressure to set aside what mattered most and ultimately deciding he wanted to work in a way that felt fully in his integrity. Anthropic's latest valuation sits at $380 billion, and the talent AI war between them and OpenAI seems to make a major headline every week. We don't know Mironk personally, so I can only speak to what he shared in the letter itself, but what it reflects is even amidst the kind of wealth accumulation that joining one of these companies can offer, there are people on the inside who are genuinely uncertain whether the reward is worth the broader unknown risks. Charles, I know you mentioned last week that this was the first tech wave you've experienced where there hasn't been clear consensus about charging forward for the parties involved. So did this piece surprise you or something like this coming out.

I also thought about it in the context of all the stuff around the Department of War and the use of this technology. I read this and I was like, oh, this kind of reminded me of like, don't be evil when Google went into China. And two things can be true. I think it can be very true that in 2023, the people at Anthropic, and I don't work there, so this is all speculation, felt very strongly that like that was a principled position that they could hold. And that I'm sure they were like, there's probably some level of conflict that we can envision between this value and the business directive. But we think it's tenable. And I don't think anybody could have predicted that we'd be in the state of the world that we're in right now with AI, given all the advances. And I think, you know, there's that old saying about, you know, values should cost you something. I'm paraphrasing. And I think they concluded that maintaining this value, as the person said pretty plainly, like it just is not tenable, which doesn't mean that it was never important. It just means that the company doesn't believe they can continue to achieve their business goals. And I think this is one of the challenges you have as a company when you decide to take a position or a stand on something. It can become become more expensive over time. And if you're not prepared to defend it, there will be some people who will question, well, did we ever mean it in the first place? And, you know, I don't think we know enough about the specifics of what OpenAI agreed to do with the Department of War. But there are all these questions about if you're a technology provider. And same thing when there was a whole thing where, hey, should Google work with the military? There are a lot of people at Google who felt like, hey, I didn't sign up for this. And I think on the one hand, it's hard to sell people on joining a company that has a mission and then changing the meaning of that mission Midway through, you are of course going to lose people. Also, I think corporations were in a different place in 2020 and 2021. I think they cared a lot more about how their employees felt about the nature of the work. I think given the layoffs and the cost pressure, I think companies are kind of like, we're just going to do what we plan to do. We're back to the old world. We've been there for a few years. We're back to the old world where there might be some mission orientation to the company when you join, but these value statements are flexible. And if they become inconvenient or untenable, companies will change them.

My one question or thought about this was I was wondering, I assume it's coming internally, where the pressure is coming from to abandon specific policies like the one that Anthropic is leaving behind, because it's very clear that as of today, OpenAI and Anthropic, they are the front runners. I know that investors love to say like, oh, we're definitely pushing those large companies for profitability, but doubtful just given the large raise. It wasn't like that was prohibitive to them, to people becoming interested. So it is fascinating that a lot of this pressure is likely coming internally, even though they are still likely the two leaders in this space.

My suspicion is what you said is true because that's the side of the dot. Like, we don't know what the internal Slack and messaging conversations— I remember when I was at Google, when Google went to China, the amount of internal tension and angst. There were a lot of people who were like, this is a violation of my perception of the company, and now I'm associated with something that I don't believe in. No matter how much you pay them, I'm not saying that these people are not fairly compensated, yeah, but if you're asking people to put in insane hours and make big sacrifices in their personal life, it's natural they're going to identify with their work because you want them to. So when they identify with their work, you have to always be mindful of the fact that the company's identity and beliefs, in some ways, they will see themselves in those. And if they don't see themselves reflected, they might say, why do I want— even with like fabulous compensation, do I want to spend 60 or 80 hours a week? I think it's one of the nice privileges of being well compensated in tech if you've been responsible with your money, no pun intended, you have the ability to say, I don't need to or want to do this. This doesn't reflect— I'm guessing that like the janitor at Anthropic is facing a different calculus and is being asked to do different things. And I don't mean privilege in a judgmental way. It is one of the privileges of the fact that like tech workers are generally well compensated, that if you find companies that are not consistent with your values, I think I'd argue it's pretty hard to stay in commercial AI because I think right now everyone feels like I just want to win. And I'll do almost whatever it takes, although with some limitations apparently based on the, the Department of War stuff, to win. And I'll become a better citizen once we've won. Or, or we're better guys than those guys, so it's better if we win than they win. And you can very quickly spiral on that stuff if you're not careful.

Well, our next piece is a post that Jack Dorsey posted publicly on X last week announcing that Block, his payments and fintech company, is reducing its workforce by nearly half, going from over 10,000 employees down to just under 6,000. That's more than 4,000 people let go. Or entering consultation, and Jack chose to share it publicly in his own words on the same day employees were notified. What made this one circulate so widely is the reasoning. Jack was really explicit that this is not a company in distress. Block's gross profit is growing, they're adding customers, and profitability is improving. The cut is happening because Jack believes that AI paired with smaller, flatter teams represents a fundamentally new way of building and running a company, and he decided it was better to act on that now than to drag out a slower reduction over months or years. He writes that repeated rounds of cuts destroy morale, focus, and trust. So he chose one hard, clear action instead. We've covered a handful of CEO statements about AI being at the forefront of how they're building their companies, but this is one of the largest workforce reductions we've seen as a percentage of total headcount out in the public. Do you think this becomes a catalyst that other companies of similar size start making similar cuts, or does this feel like it's still maybe on its own island for now?

I think it's an excellent question and I have like good friends who work at Block and so I feel really bad for them. That's a massive cut. I think we're also in this zone where when sometimes when there's something macro and I don't, I don't know why they did this. So this is all speculation. I do think we're in an era right now where there is a premium for AI-focused decision-making, showing that you're putting AI at the forefront and you know, Getting rid of 40% of your staff is a pretty strong statement. Now, I read some more cynical takes, which is like, well, the company had just overhired and that this was as much about the sort of Twitter style. We have too many people relative to like what needs to be done. And rather than admit, I don't, I don't know though, but like, I, that's plausible, but to overhire by 40% of the company seems like a lot. And I think it reminds me a lot of I think it was Klarna who got rid of all the customer service people. And then they're like, we might have overdone it. We haven't heard as much about that since then, but the tech's in a different place. And I suspect there are other CEOs in the Valley and in the Fortune 500 who are like, well, I'm glad Jack went first. Now maybe we can do something similar and not be out front and first. With all of these things, I always think it's interesting to see what happens. And tech is now in a like permanent layoff phase. It used to be that layoffs were these like singular events. They happen so often now in big tech, they're just not news anymore. When a big tech company cuts 5 or 10% of its workforce, if you think about how big some of these companies are, that's a lot of people. It's just ceased to be— and in a lot of ways, tech has become like other large corporate entities where they're just constantly tinkering with not just the total workforce but the balance of people across functions and roles. And I think what used to be a career that was safe in an industry that was in kind of permanent growth and hiring mode has really gone in the opposite direction. So I really do hope that we get insights into what happened at Block. And I think the Twitter one's instructive. Like, Twitter, Twitter didn't fall over when all those people left. I'd argue that it's probably not the same asset it was when all those people were there either.

I mean, it was good timing. The Citrini Market Research came— report came out last week, and you saw the markets reflect that reaction. I mean, Block stock soared 24% after hours when that announcement was shared. What this decision assumes is that if this is, you know, done for truly AI purposes and not in a different business decision, like we kind of alluded to, what this decision assumes is that the remaining employees are also prepared and upskilled enough to actually leverage these AI tools and absorb a new way of working without performance taking a hit. And while this is certainly more concrete than the leaked memos about AI efficiency that we've seen from other CEOs, I'd argue that the underlying dynamic is maybe similar, more about signaling or direction than demonstrating clear progress forward.

Two things that, at least for me, are worth thinking about. One is I do keep reading about a lot of CEOs who are focused on profit per employee, and my hunch is, if history is any guide, Block will not tip over tomorrow because all these people are gone. And the impact of people being gone, especially in large numbers in big companies, usually isn't felt for some period of time. And maybe this is the bet, which is usually what ends up happening is like you're not as responsive. You can't build as much stuff. You focus on the core, you miss new opportunities. Your company remains good at the core thing but never gets better at anything else. And maybe the bet is that like AI breaks that paradigm. But I wouldn't be surprised if in the short term Block's financial metrics improve dramatically. You've cut out a significant chunk of cost and whatever the impact is of that on revenue is unlikely to show up in the same time period.

Our last piece is a LinkedIn post from Nick Poulos, founder and GP over at Euclid Ventures. His core argument is that the venture industry has developed what he calls a consensus addiction and shares data to support the statement. Seed startups priced below $20 million are nearly 4 times more likely to produce a 50x return than those priced above. And yet the 95th percentile seed pre-money valuation just hit $80 million. That's up 138% in a single year. Meanwhile, the number of rounds closed hit a decade low, even as total dollars deployed hit a post-2022 high with A16Z just raising $15 billion, roughly 20% of all VC dollars deployed last year. Nick's argument is that capital is flooding into what he calls fundability, hot market signals, brand name co-investors, consensus legibility, and mistaking all of that for investability, meaning actual return potential. And his team traced back every billion-dollar-plus exit in 2025 back to its first check. And the data is consistent. The vast majority started with the $500K to $5 million raises. Platform funds were underrepresented relative to their AUM share. And the key takeaway here is counter to the behavior we see in the market today. If winners were legible at first check, the data would show it, and it doesn't. The 4x return gap between sub-$20 million and above $20 million seed valuations is undeniably large, but my instinct is that this data alone won't change behavior. Funds will keep pricing rounds this way and founders will keep opting in. My question is who actually benefits the most from Nick and the Euclid team putting this out there? And I know that's not what they wrote it for, but my question was like, okay, what does everyone do with this information now other than just kind of have a little public discourse on it via in the comment section?

Well, I mean, we benefit, like I, I obviously agree with Nick. I do think, though, I've talked to a fair number of LPs about this very topic, and there's a couple of arguments I hear from them. Number one is, well, that's old data. The world isn't the same anymore. Yes, once upon a time that was true, but it's not true anymore. Second, you know, a lot of the big funds tell LPs we're good at everything. We're good at seed, we're good at A, we're good at multistage, and we have tremendous access. And there are two things I believe most LPs and GPs at larger funds believe. One is that they see all of the best stuff and that the best stuff is as identifiable at day one as it is at Series A. And I think the latter part is the part where I think mistakes are made. So I do think if you just do the math also, it is a lot easier to get— and this is something I tell LPs all the time— it is a lot easier to get high cash-on-cash multiples with low entry prices. And I will say there are some counterexamples, like if you look at Wiz, that's a company that started life very expensive and whenever the deal closes will end life very expensive. And I think it was $100 pre and $23 billion. But the question is like, is that the new normal or is that an outlier? And so I think right now we're running an experiment which essentially says price doesn't matter anymore. The counterargument is you take the Venkat plus Andreessen argument, which is that the scale of the winners is so big that if you can get companies that are $50 to $100 billion in size, doesn't really matter. You get, you actually can still get 50 times your money. But I'm like, but we just don't have that many of those companies. And I think, you know, it's very possible that we will in 7 years have more regular trillion-dollar private companies. If you just have either some combination of asset price inflation and just the general growth of private markets, but that doesn't mean that that's the average underwriting outcome. We only have one SpaceX. We have one of those. We have like A Stripe. There's like 10 companies that are in this rare air. And yes, the multi-stage people can capture those. And some of those companies started off life very expensive because Elon was involved. So I think part of this is to some extent the answer is both things can work. The question is like, what's the balance? And right now the balance of dollars is going into companies that are beginning life expensive and the balance of dollars available for follow-on. This is the part I think that people need to really internalize is When you start life as an expensive company backed by a multi-stage fund, you're just in a different universe of follow-on fundraising considerations. You're already sort of in the club. It doesn't mean your next round is guaranteed. It doesn't mean they're going to give you more money. You're already in the consideration set. When you start off as, you know, two guys and a dog in a hot tub with $500K, you've got to like work hard to get the attention of the big funds. And sometimes you get their attention à la Clueless by like breaking through the noise floor, and sometimes you get their attention through performance. But the companies that start with their attention are actually just living in a different world. And I'll be very curious to see if we fast forward 10 years, were these billion-dollar research lab, compute-based data center, whatever it might be, were these billion-dollar inception-stage companies good investments? They might turn out to be good companies, but were they good investments? And that's something I always come back to is like, a good investment is an investment that makes you a really good return relative to the risk that you're taking. There can be a good company that's a bad investment for you. And so we won't know, but right now the consensus is that price doesn't matter because outcomes are large and and you shouldn't worry about it. And some, ironically, in some cases, low prices now become a negative signal. It's somehow like, oh, you couldn't raise 5 on 25 right out of the gates. Are you a bad founder? And some of our best companies started life very inexpensively, and I'm really glad that we're on their cap tables.

There's two thoughts that came out of that. One was in the connection to Josh Kopelman's Venture Arrogance Score of like, if you're a mega fund, you need all of these expensive deals in your portfolio. You need to win over, I think when a16z was raising a $15 billion fund, he's like, you need to be in 50 of every single high outcome deal. That's a lot. That's a lot of surface area. So the game is different. Paying up early to create signal or just to have that option is so worth it. So that's also what's happening in the background. And then to your comment about, you know, fundability versus investability, Peter Walker had some good comments in the post after, you know, he said that at Carta, they saw the highest quartile of seed valuations as much more likely to become unicorns.. So kind of like having that head start up for follow-on capital, but the 50x outcomes were spread essentially evenly across 4 valuation quartiles by year.

If you think about it, who makes unicorns?

Right.

The only people who can make unicorns are multistage funds. So everyone's like, look at all the unicorns in this. I'm like, well, they made them all. They put the billion. So if I could put a billion-dollar valuation on every company we invested in, we would make a lot of unicorns at Precursor. And I'm not like making fun of the big funds. But I do think it's just one of these things that's like structurally the people who assign the valuation are the ones who make the decision. And if you're already in a company and the company's familiar to you and your network and the founder's a good fundraiser and a good storyteller, that person has a good chance of getting to a billion-dollar valuation independent of metrics and business progress. And so I am like very curious to see how this all plays out. And I think to your earlier point too, it causes people to do the venture ergo— it causes, you know, Sequoia's in Anthropic now, but, but much later than I would normally think of Sequoia. But if you're them, you kind of have to be in that company given its scale and scope. And look, if you've got— there's a lot of ways to return a big fund. And in a lot of ways, I think the counterargument is, you know, if you're Andreessen and you've got a Databricks, there's sort of no amount of money that feels like too much to put into that company. You're like, oh, you're only going to get 5x extra money at that price. You're like, yeah, but you can put in half a billion dollars. 5x on half a billion dollars takes you a long way towards returning a big fund. And I think that's something I appreciate about Thrive Strategy, which is like, yes, it's a lot of money., but if you can concentrate it into companies that will compound and will ultimately be far outlier valuable, you can put 10 to work and get 30 back.

My last question here is more so maybe thinking about like the non-mega funds who participate largely in this type of round dynamic. Do you think that those GPs, if they're being honest, know the difference when they're writing the check between consensus versus full conviction in the moment? I think in my brain, I mean, like we've got a handful of deals in our portfolio that have been priced around this range and we truly believe we fell convicted in the idea and the, and the reason behind it. But I think that there are funds who are not mega funds that 100%, they're doing only these types of deals. And do you think that they're really thinking they're truly convicted or this is the ride and the game that I wanna play on and I have a right to play in this game as much as the mega funds with less AUM?

Depends on the manager. I know some people who really think this is where value is created and these companies that are born expensive are the good ones and they're expensive because they're good. And we only wanna be in things that are expensive because that are good and finding things that are good and inexpensive is not something I want to work on. I have several friends who subscribe to that and they are very good at investing that way. I think the other thing is on a practical basis, these deals are easy to talk about. If you go to an LP and you're like, hey, I'm raising a fund, I'm in these 3 hot deals, 3 very expensive, but deals you've heard of that Andreessen or Sequoia or Lightspeed or Index is in at an individual security level, the LP might not know anything. But if I told a lot of the people I talked to who evaluate venture funds, you have 2 choices. Someone tells you there's a company you've heard nothing about and Sequoia is in it, and there's a company you've heard nothing about and Sequoia is not in it, most people would be like, well, I want to be in the company that Sequoia is in. And so the thing about fundability is fundability shows up today. Being able to say we, we are co-invested with these people, this person who spun out of this research lab. Yeah, it's expensive, but look at all these other really fancy VC names that are in this deal with us. That's why you should feel good about that. That versus like, hey, this is a company you've never heard of and you will have never heard of for 7 years. And then by the time you've heard about it, I've had to raise 2 more funds. And so I do think the irony is that the people who are being rewarded for having bet on those companies at low valuations, those are decisions that were made a decade ago that are being harvested now. And there's more of that harvesting to come. The question is, 15 years from now, will it look the same?

I don't know. Well, can confirm good and inexpensive is a very hard job. So we're excited to do it here. Thank you so much, Charles, for joining in. We are all out of time today. Thank you to our listeners for hopping in as well. Please like and subscribe, and we will talk with you all next week.

Thank you. Bye.