Why Patience Wins In Venture | Humans Still Beat AI At This | AI Agents Are Your New Workforce
In this episode of Precursor Ventures’ Learning Corner, host Mia Farnam talks with GP and managing partner Charles Hudson about three recent readings that shape their view of the evolving venture landscape. First, they discuss Nick Charles’s essay on “Patience and Owners vs. Tenants,” arguing that long‑term, owner‑mindset investing beats the current rush for fast revenue milestones and that LP‑GP alignment on patience is a key tension. They explore how this pressure filters into founder expectations, the “legibility” bias toward hot names, and the risk of funds dying before their long‑term thesis bears fruit. Second, they review David Deming’s Substack piece on “Efficient Social Learning as the Human Advantage over AI,” which posits that humans retain a comparative advantage in sample‑efficient, relational reasoning and that this edge will remain valuable even as AI commoditises many tasks. Third, they examine George Savulka’s X thread on “You Just Hired a Million Bad Employees,” drawing parallels between managing AI agent workforces and traditional employee management, emphasizing the need for clear evaluation metrics (evals), loops, token‑maxing, and the role of forward‑deployed engineers as transformation guides. Throughout, Charles shares anecdotes about LP expectations, founder dynamics, the challenges of autonomous AI agents, and the importance of human leadership in building shared understanding and product coherence. The conversation ends with reflections on the practical limits of AI autonomy, the necessity of continuous oversight, and the emerging consulting model that helps companies re‑engineer processes to extract AI value.
Chapters
If AI is so ready for prime time, why do we need so many forward deployed engineers? Wildly promising technology requires massive rethink of the way that companies work. This just to me reinforces the idea that this technology is going to take longer to generate business value than people think, because to get the real value, you have to like rip up and redo your processes.
It's the week of July 20th. Welcome to the Learning Corner, a Precursor Ventures podcast. Welcome to the Learning Corner by Precursor Ventures. 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. Our first goodread this week is a blog post from Nick Charles from Asylum Ventures called Patience and on Owners vs. Tenants. Charles argues patience isn't just a virtue investors should aspire to, it's actually a strategy to make more money. He points to venture's growing obsession with speed chasing revenue at all costs, 1 to 2 year fund deployment cycles making a comeback, and metrics like time to $100 million in revenue that ignore whether that revenue is any good. He borrows Jeff Bezos's owners versus tenants framing. Tenants churn through what they hold quickly while owners make decisions built to compound over years. And Nick backs this up with examples like Clay taking 7 years to nail its product before scaling past $100 million in ARR, and OpenAI needing a full decade before ChatGPT became a hit. And his closing point is that patience doesn't mean moving slowly. It means being willing to look worse in the short term in exchange for building something that actually lasts. And Charles, I have a bunch of questions here and want your broader reaction, of course, but I think. The largest point of tension in this piece here is how to align on what patience means between fund managers and their LPs. I think the original version of early-stage venture inherently was tied to patience, and now the conversation has become what is a reasonable level of patience that LPs, and then therefore GPs, should have as it pertains to liquidity. Um, but I'll pause for your broader thoughts first.
I emailed Nick right after I read this. I'm glad he wrote it. It's funny, I went on, when I was on, uh, The Peel, I talked to Turner about this. There's this whole meme about like the game on the field in venture. We talk about this, oh, there's the game on the field. Can you play the game on the field? Should you play the game on the field? And my general view is it doesn't matter whether you wanna play it, but you must be aware of it. And I think about our model, which is to get in super early, usually pre-product market fit. And I would say there have been times in the last 15 years that I've been doing some form of precursor ventures where being super early pre-product market fit and taking time to build things has been very much in fashion. And there've been other moments where everyone's just like, "I want 'just add water' instant growth." And I've been on the same journey that Nick explains in this post. You know, I went to a lot of LPs and said, guys, some of these things that are getting to $100 million in revenue will come right back down. This revenue may or may not be durable. We don't really understand. And I think in a time of great uncertainty around what, you know, what moats are, what's gonna work in AI, things just outta people who said, I'm just gonna chase the things that are the fastest rabbits, the things that get outta the gates quickly. Because to some extent, revenue is a signal of something. Whenever I ask people like, hey, bubble, or some of this stuff, they go, but look at the revenues. I go, look at them, actually look at them. Look at like what underlies them, whether they're going to persist. Whether there's any margin potential there along, and there's almost no appetite for that conversation. And I think we live in this feedback loop where a lot of LPs, venture is a part of their portfolio. It's not the whole thing. And in many cases, I meet limited partners whose primary interest is, are you in the hot companies? And I think in some ways it's because the hot companies are easy to understand if venture is one of 10 things on your plate. You've heard of OpenAI, you've heard of Anthropic, you've heard of SpaceX, you've heard of Anduril. Maybe there's 20 comp— you've heard of Databricks, but maybe there's a great slow growth company that on your statement, it's still flat. And the investor's like, tell me why I should be excited about this company that's been flat for 5 years. Replit. Like there are a bunch of these companies and I think there's very little appetite for that narrative right now, which means if you're gonna pursue that strategy, You need patience yourself. You need fortitude and you need people who will help keep you in business for that good news to turn out to be true. And I think this also has trickle-down impacts for founders because VCs are responding to the institutional incentives, which I think right now are to be in the hot, visible, high-growth things. Some of them will work out, some of them won't, but they're all— it's funny, we talked about what's legible to capital, that whole legibility thing. And I think we're now in the same era for VC managers, which is like, well, what's legible to LPs? And I think legibility is just a huge shortcut to stop you from asking hard questions and thinking, oh, this person went to Stanford and worked at OpenAI, so like they're probably smart. Maybe. I mean, yes, they're smart. It doesn't mean the startup's going to work, but it's easy to underwrite. And I think in some ways what Nick is talking about is the extension of this legibility concept to fund managers, which is like, hey, if you're in a bunch of names I've heard of, whether I have an opinion on those names or not, you're in things I recognize and know, and that's easier to talk about it at a committee than a company that a manager has great belief in, but it's taken 5 to 7 years to show anything. Yeah.
And I think that's, I guess, the even more deeper tension is the conundrum of maybe the arbitrage opportunity is to be patient right now while everyone else is racing ahead to deploy and into things that have $100+ million revenue. That's maybe not defensible, but will you get the opportunity to be patient in the time required for you to kind of end up 6, 7 years later and perhaps you're right. That's what I think about a lot. Like, who's the patience? Yes. I think like if everyone could just take a beat and be like, wait, should we just take stock into what we're actually deploying into? I think most people would say yes, that's a reasonable thing to do. But you have to deploy money. Fund managers need to recognize, or LPs would like to recognize and showcase the great portfolio they have. So do you get that chance and leisure to be patient? I think sometimes is the question in my mind.
Feels like patience is a luxury right now.
Yeah.
Um, but also I think a lot about— I have a handful of my friends who are hedge fund managers or other public markets people. They go, look, the market can remain irrational longer than you can remain liquid sometimes. And I think that's the equivalent of patience in venture is like, if you're really patient and right, you might go out of business as a fund manager before you're proven right. You'll still make money. You just won't have a fund anymore. There are examples of this in my history as a fund manager where funds that were counted out LPs didn't back them. The people sat there and managed out the tail, and the tail ended up being— Bay Partners is, is one example I could think of— the tail ended up being very, very interesting. And people had sort of written, written it off because it wasn't performing on the time cycle that people wanted. Yeah. And so I don't know, I, I think about that a lot in our business, which is like there's this tension between the things you need to explain and show to limited partners to continue to earn their trust and dollars to stay in business, and the things that will actually work and generate returns for you in the future.
I do think the one thing that's very clear is to your earliest point, being able to see the game, understand it, and know when to participate as a fund, knowing all these different dynamics at play.
Yeah.
I do wanna go back to how this impacts founders because I was thinking about this through the lens of investors, of course, and the relationship between LPs and GPs. But patience is obviously something that is inherently important for founders in this age too. And certainly media hasn't done a great job in terms of making people feel like being patient can be rewarded. I also think about it as people who work at startups right now, as well as non-founders, people who see things on LinkedIn or different media outlets and see like, well, my company's not growing at 100x year over year, should I leave? And I think generationally, this is something that a lot of my peers have been talking about is, like the itch to feel like they have been working for a while and not seeing the instant reward. And I continuously come back to the patience piece of like, I think most of the stories that you read are, maybe there's like one in a million that are truly anomalies. And I think that's just what true venture are. The rest is a lot of posturing, um, and kind of just optics to make things seem a different way. But it does trickle down to every facet of our ecosystem.
In the last 10 years, gambling and speculation, I'd argue starting with crypto, but you could go crypto to sports betting to prediction markets. It's just become a bigger part of social life. Yeah. And I think investing in patience is about getting more later by giving up something now. And I think part of this is just a change in social values and social interest and the feeling that long-termism is for suckers. That's the old playbook and it doesn't work anymore. And if you're still following it, You're a sucker. And I think that's like a very damaging social value cuz there are lots of things that take time to, but I also understand the impetus for the people who feel that way given the world that they grew up in. And I think also it's hard for startup founders. I talked to someone in our portfolio, he's like, I tripled this year. I go, nobody cares. And he goes, what do you mean? I go, hey man, 5 years ago tripling would've been amazing. And I don't wanna diminish how hard it is to triple a business. And when I say nobody cares, I mean the people you want to go raise money from, are gonna see your triple and it's not gonna make them jump out of their seats. They're gonna say, well, you tripled. I have a company that's 3x-ing month over month, not year over year, month over month. And so that's what I'm looking for. There is a collective obsession with these like weird— if I see another one of these like 9 charts that all have this like super vertical shape. And of course the startup that's being talked about is the leftmost line for the fastest, you know, to $100 in 30 minutes. We're the fastest startup in the history of Adam, whatever. It was gonna be something silly. That's where people's energy and attention is, and it's very hard to fight.
Yeah.
Unless you have a really good emotional and financial support system.
Well, our next good read is a piece from David Deming's Fort Lightning Substack titled Efficient Social Learning is the Human Advantage Over AI. David makes the case that as AI gets more capable, Humans will keep a real edge in one specific area, learning efficiently from very little data, especially in social and relational tasks, something that we talked about a month ago, I believe. He walks through research showing people can infer someone's preferences and motivations after watching just a handful of interactions, while large language models still struggle to carry what they learn about one person over to a new context. The core economic idea is comparative advantage. Since AI will eventually get cheap and commoditized, the tasks where sample-efficient human judgment matters most— reading a room, building trust, adapting to one specific person— become disproportionately valuable. He's really careful to separate this from the question of whether people will pay for a human in the loop, which he calls a very separate and unresolved debate. His point is really about where human labor still has a structural edge regardless of what customers end up wanting. And David ties it back to evolution, arguing human Humans got good at social reasoning because our survival depended on transmitting knowledge to each other, not because we practiced abstract reasoning, which is actually the newer and harder skill for us. Um, I love this piece. Aaliyah from our team shared it, and I thanked her for doing so. One line from David really stuck out with me. Quote, the context of a single relationship is constantly evolving and requires multimodal learning from accumulated common ground. Social learning doesn't scale easily. Building a deep relationship with one person doesn't let you skip steps with others. And he makes the point that LLMs are also decent at predicting how a rational person should behave, but fall apart the moment someone acts irrationally or out of character. So wanted your, obviously your broader reactions to the piece, but wanted to know, do you think that's the real lasting ceiling for AI in today's world, or is it more that these models just haven't been trained enough on the messy, super non-relational parts of how people actually behave yet.
Uh, I read it and I was like, well, this is a grounded, optimistic take of how this all plays out.
Hopefully.
Yeah. I thought the idea of comparative advantage is a good principle to introduce, which is like, yes, even if the LLMs have an absolute advantage in everything, there'll be some areas where having humans involved still makes sense. And I do think, um, one part of this article that I really agreed with is there are places in the world already where access to humans is a, you know, if you, if you're a high status airline flyer, you're, you don't have to navigate an AI tree or voice prompts if you don't want to, you will get a human on the phone. And oftentimes the human, it's funny, I, you know, I write a lot more code now and there are sometimes I'm just like, that took way too long. If I had explained that task to a human, they wouldn't have gone through the process of breaking down every single step. And then we would have solved it in the process. They would've said, oh, I think I know what you mean. Or, hey, let's go back and forth. And I, I read it and I was like, you know, a part of what I think makes ironically the most successful companies that I work with successful is the ability of leadership to communicate complex ideas up and down the chain through to and through people. Yeah. And those people's ability to absorb a message and figure out how to translate it to the next person in the chain. And I think there's a lot of brute force tasks. That LLMs are just good at because once they have learned how to do them, the variance surface is known. And I have some of those tasks running already at Precursor Ventures Ventures where I'm just like, 99% of the time, the known outcomes for that task are fully mapped. And the 1% that aren't are the things that the team and I spend time on. There are other things I think about, like most of the founder calls I have, I have a vague idea of what we're gonna discuss when we get on the phone. I don't always know where they're gonna go.
Yeah.
What topics they're going to bring up. And I think a lot about, um, I was reading something else that I think that the NextView team had written and I was like, yes, I do the same thing all the time. I talk to 30 founders in our portfolio every week on Zoom or in person. Those, those, those calls have the same structure. They have nothing in common in two dimensions. One, every company has a different set of issues. And more importantly, each founder is their own computer. They're their own emotional computer with their own idiosyncrasies and their own beliefs and their own strengths and weaknesses. And my view is like, the point of that meeting is for me to do the mapping between like, who is this person? What is this issue? And what am I trying to communicate to them? And that is like not scripted.
Yeah.
It's like a little bit of intuition. There is some pattern matching, like, hey, I've had 30 meetings with this person. I know they get triggered by X, Y, and Z thing, or I've had 30 meetings with this person. I know that they can take direct feedback without any kind of cushion on it. But there are other things that come up randomly that are really de novo and you're like, well, what I'm really doing is synthesizing all the information I know about this problem and this person and contextualizing it on the fly. I'm not sure that LLMs are great at that task yet.
Going back to your example of strong leadership, being able to kind of tell people what to focus on and that being such a high, highly valuable skill, is that one in the same to managing people? Because I read this piece, then I was like, oh, are good people managers now going to be like the most valuable people in your organization? Now that's obviously with the belief that like people will continue to work in your organizations and like it won't just be a bunch of agents.
Increasingly when I talk to our portfolio companies, I ask, I'm like, Well, what mistakes are you making with AI? Like, oh, we built the wrong thing. I'm like, what do you mean? Like, well, we built something and then we got 3/4 of the way down the road and we realized like it wasn't right. I'm like, oh, so you had bad product management. Or, oh, we built a bunch of things that customers ask for and now we've lost all product coherence. And I've always felt like part of what good human leadership does is it creates a shared understanding of the job to be done among a group of people. Right now people are devaluing that step because it's so fast and easy to build. But I'm like, you still have to build the right thing.
Yeah.
And I think part of what a good leader does is they spend time with people talking through this is what we're building and why. And they create space for like alternative viewpoints to come up and get debated and maybe integrated or dismissed, or for other people to say, why aren't we talking about this element of it? And people go, oh yeah, you know, you're You're, you're right. As, uh, one of our founders has been helping me with a software development project and he's like, you figured it out. By the time you build the first version, you know what you actually should have built. And I think the what to build, how to build it, and the consensus building, as long as there are people involved, I think that will become an even more important skill because pre-AI, you had to do all of that work before you started because building was expensive. And now because building is cheap, a lot of people are like, well, let's just skip all that and build. And I'm like, well, you're going to end up with the same problem just at the back end instead of the front end.
And all the same workarounds and time required to fix. Well, our last good read this week is a post on X from George Savulka called "You Just Hired a Million Bad Employees." George draws a direct parallel between agent workforces and human workforces, arguing the two break down in exactly the same ways. So we can use decades of management lessons to fix how we run AI agents today. This is a really good tie to what we were just talking about, actually. He walks through a set of specific pairings. Token maxing is just throwing bodies at the problem. Loops are meetings about meetings. Wasted tokens are the new headcount bloat. And context hoarding is the modern version of employees guarding tribal knowledge as job security. He argues the fix in both cases is really the same. Define what good looks like, which is why He thinks evals are becoming the new OKRs. Coding, to what you just said, Charles, is the one function where AI has actually delivered, and that's only because coding has built-in evals. Coding either compiles and works or it doesn't. While most other business functions have no equivalent way to measure success, George also raises a political problem. Employees who hold the most valuable tacit knowledge have the least incentive to teach it to an AI system that might replace them. His bet is that the biggest winners won't be new AI-native startups, but transformation companies. That go deep enough into an existing business to actually build the evals and workflows needed to make agents reliable. My favorite parallel he talked about in his piece was about loops being the equivalent of meetings about meetings and simply people just not knowing what to ask for or what success actually looks like. So you're really just stuck prompting in circles. I didn't know if you had a favorite comparison that George mentioned or one that stood out to you the most?
I think people's like utopian idea, like I tell you, oh, we have this, a self-managed, self-running company. I'm like, but is there anybody checking to decide if it's like heading in the right direction or making good decisions? And everyone's like, full autonomy. I'm like, I run a bunch of agents in semi-autonomous mode because I like to intervene. In many cases I'm like, oh, that's close but not right. That is not, that does not meet my quality shipping bar. Let me shoot in the same way I would help an intern be like, hey, we wouldn't publish that or we wouldn't ship that at Precursor Ventures. It's close, but let me tell you what is required. That's, I think, the state of the art in tech. And I, I keep wrestling me with this like fundamental contradiction that I feel like whenever I mention it to people, they give me this puzzled look. I'm like, if AI is so ready for prime time, why do we need so many forward deployed engineers? They're like, well, they're not really service people. They're not like traditional Cause, and I'm like, fair, but like, what are they doing? Yeah. You're like, well, they're helping these companies with business transformation. I'm like, yeah. So the companies can't adopt this stuff on its own, which is always the case with new technologies. So I'm like, what if we're closer to where I think we are than other people think we are, which is wildly promising technology requires massive rethink of the way that companies work to get full value. And companies don't have the internal Sherpas to do this themselves. At scale. So either the vendor has to become basically the transformation agent, or you need a Deloitte or a McKinsey or a third, or OpenAI, or— Yeah. You need an FTE shop. And I'm just like, this just to me reinforces the idea that this technology is gonna take longer to generate business value than people think, because to get the real value, you have to like rip up and redo your processes. You can't just drop an agent. Into your car. There, again, there's some discrete functions, customer success and a few other things, customer support, and that are discrete where you could drop an agent in and say, we're just not gonna necessarily do a total transformation. And I just think about, like, I mean, I think about what we talked about in a previous episode about Sierra. Yeah. They're in this business and they're very smart people. And even they were like, this is hard. And we're in the business of selling AI-powered customer support to other people. Mm-hmm. And so I'm like, well, what, how much harder would it be for a company that doesn't have the same technical resources and isn't starting basically with a blank slate?
For most companies, I think the issue is they don't know what they don't know. So the forward deployed model works perfectly of just like, I'm gonna have people sit with us day to day and almost like coach me on what good looks like or what it can actually become. But then probably like halfway through the project or the deployment or the transformation project, the folks who are around the table who are actually at, at the actual org are like, oh wait, now I understand what you're building. I have all these edits because I had no idea in context of what was being created. I mean, again, consulting gets to live another bunch of years thinking about it more broadly of this is exactly what that industry was built upon, which was not staff og, but let me carry the burden of figuring out this mess and I'll bring you in at the right times to get your perspective. And I think we're seeing every organization, small or large, now feel like they require that outside skill set to be the sounding board as they navigate what they want to do here. One thing that he mentioned in the beginning that I disagreed with a little bit was his quote that said, managing AI is harder than managing people because AI scales dysfunction instantly. And in my mind, I was thinking, sure, like if what you've built is distributed across an entire set of integrations on a pure task perspective, then it's rather difficult. But I find unwinding that dysfunction to be quite easier. I guess I'm only speaking from my experience with a smaller org. Not talking about working for like a thousands of people org, I, I do find undoing years of dysfunctional teams is nearly impossible to pull off. You're almost better off just firing everyone and then starting from scratch. Didn't know if you had a perspective on that one as well.
I think, um, what I found is that the more autonomous the AI, the more problems and trouble it can create. And, um, it's not always good at telling on itself. And so I think the difference is like, With humans, we've built in all of these checkpoints on work that are generally designed to limit the amount of damage someone can do. And the more trust you have in somebody, the less often you check on their work. But that's usually an earned right. A lot of the AI is like, we're just gonna go off and do this thing. And again, if you don't have eval, if you don't have like, if you don't have software-driven checkpoints, you can look up and be like, oh wow, this thing has like been off the rails for a very long time. And nobody noticed. And sometimes because it's confident and wrong, it's not gonna tell you, I think I screwed up. It'll just be like, just processing, just cranking away over here, sir. And you have to look under the hood. And I think there's this belief that like, oh, these things, once you set them up, they're, they're not set it and forget it. The models change. Um, sometimes like you accidentally inject new stuff, like things break. And I find that like, because AI agents basically will work at whatever frequency you tell them, unlike humans. Sometimes they'll just run and run and run and they can run for weeks before you realize, oh, this thing's broken. It's been broken for weeks. That would almost never happen with a human. Yeah.
Like the thing that makes agents so wonderful on paper, that they'll never burn out and they can run things at scale, like is the limit. They, they don't carry guilt or internal stress in the way that if you're a lower level employee and you have questions or you know something is running poorly or bad, you'll think about it overnight, every night until you finally get it. Like a manager will get it out of you.
Yeah.
What do you actually think is going on? As opposed to just executing, like, here's the whole list of things I have concerns over that are not letting me sleep, and I don't think I can push this project forward if I don't surface them to you. Like an agent, like why would they ever bring that up? It just wants you to continue to use it.
Also, I think I basically have management for my agents now. I'm like, okay, Claude Code, help me write code. But Codex is the VP of engineering and we don't ship anything without Codex and Codex always finds something. Yeah. That it doesn't like. That's usually an improvement. Then I'm like, cool, we're gonna run this for a little while, then we're gonna like check on it. So I'm like, I manage my agents the way I would manage basically a junior employee. Um, the difference being that this junior employee can work 24/7, but it still makes mistakes. And again, it, to your point, it has no guilt or shame. So sometimes it makes mistakes and it's like, oh, I made— I think I made a mistake, but I just like push through it. Yeah. And I'm like, that, that isn't what I wanted you to do. Yeah.
Well, we are all out of time today. Thank you, Charles, for jumping in. Thank you to our listeners for joining as well. Please like and subscribe and we will talk with you all next week. Thank you.
Bye. Bye.
