CEOs Are Now Grading Employees on AI Usage. Is That a Good Idea?

The irony is these two things are not separate. Financial capital and symbolic capital are deeply intertwined. Raising a lot of money so that you can be king-made and be the apparent winner before the battle's even over is a part of the strategy.

It's the week of March 16th. 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 the first topic of the week, we have a piece from Ohm Malik titled Neo-Symbolic Capitalism. He opens the piece with a tweet from Patrick Collison, Stripe's CEO, musing about why certain companies attract so much palace intrigue coverage while others never do. And Ohm uses that as a jumping off point for a much larger argument rooted in the sociologist Pierre Bourdieu's concept of symbolic capital. Ohm argues that in today's hyperconnected, attention-first economy, reputation and perceived authority convert faster and more powerfully than financial capital alone. He uses Elon Musk as the clearest modern case study, someone who's turned symbolic capital into a self-sustaining machine, and then walks through how the Stripe founders have done the same thing for Stripe through craft, patience, and carefully chosen intellectual alliances. The uncomfortable conclusion is that the tech and venture ecosystem has fully internalized this game. Podcasts exist to produce clips. Clips exist to feed social, and speed is now the signal of authority regardless of whether what's moving is true or deep. I'll repeat the tweet because I think it's important for the conversation. Patrick said, there's a lot of unevenness in how much attention internal drama and palace intrigue gets across different organizations. As far as I can tell, this is substantially a matter of path dependency. We know the characters in the sitcom of certain organizations, but not at others, creating self-reinforcing lock-in effects. How much does one hear about the power struggles at Chevron or the Department of Agriculture? There is even significant heterogeneity between ostensibly similar companies within sectors. Ohm frames that tweet as almost like this subliminal confession of symbolic capitalism. So this concept that even pointing out the game itself is a move in the game. When you see founders or investors building a public presence right now, how much of it do you think is deliberate strategy versus just the self-pressure to be visible?

Can the answer be both? I think it's ironic we're talking about this on a podcast, so we should probably acknowledge that. I really have enjoyed Ohm's writing lately. I'm glad he's back to a good cadence. I read this and thought about a lot of things. Just even if I think about the amount of palace intrigue you've read about OpenAI relative to Anthropic, it's not even close. And even Substory's Thinking Machines. Still end up being like downstream of OpenAI. I do think in some cases energy and attention tends to follow companies that are the most visible, interesting, and important companies. OpenAI is a visible, important, and interesting company. So people are curious about what happens there. But to use kind of more of like a reality TV analogy, some companies are just messier than other companies. And so some companies that are messier, like they have more drama, more stuff happens. I think a lot of this for me comes back to like Cluey, as I feel like we talk about it a lot. Cluey has this like very clear symbol to me of the importance of breaking through in a very crowded, noisy, short-term news cycle. And it's interesting, maybe even 7 years ago, a lot of people would be like, oh, that person quote unquote wastes too much time on Twitter. Twitter's part of the job for a lot of people now. It's not some side quest. It is like a big part of the job. And I don't spend nearly as much time on X as I used to. And I know that there are conversations I miss out on because I'm not there.

Mm-hmm.

And something that people used to laugh at people about spending time on has now become an important credibility building and authority building thing. And so I, I, I certainly talk to people who are like, well, I need to invest in this presence. I'm like, well, if you don't really wanna play the game, you're probably gonna get beaten. It's kinda like sports. If you're just like showing up and there's somebody else who's in the gym practicing who loves the game, the person who loves the game or who just has an enormous audience and knows how to feed them, it's probably gonna beat you at social. And so I do have some people who I think walk this line of being both strategic and authentic. And I think for a while there was this notion where you could be either or. You could have like the perfect tweet at the perfect time about the perfect topic with like the right amount of snark, or you could just like be yourself. And now I see more people who are trying to blend the two into a version of themselves online that mostly mirrors who they are. And I think it's become an important place where founders look for signal. I think it's become an important place where LPs look for signal. And I do think part of why I stopped hanging out on X as much as it did start to feel like this is a like attention thought boy game and it's just less interesting. At least on LinkedIn, people are pretty much like, hey, I'm here kind of to promote myself or share what I'm doing professionally. Weirdly for me, like, it's gonna sound cringe, I'm like, LinkedIn is, that's what LinkedIn is.

Professionalism is cringe. I don't know.

Yeah.

In today's age. I mean, does this go along with the long-held belief or widely discussed belief, like all press is good press. At what point does it tip?

I think not all press is good press. I think the most critical thing that made me reflect about Ohm's article, he's like, Stripe is a $165 billion company, and that alone is not enough to maintain relevance. Being probably the second or maybe the third most valuable private company after SpaceX and Databricks, that alone is not enough to guarantee relevance. But I think in some ways, if you go back to the old Andreessen, like go direct. I think what really happened, it wasn't so much that everybody spun up their own media artifice, it's that everybody decided, I'm going to be much more intentional about how my company— kind of like brand management for startups and VC funds. People have decided they're going to be way more intentional about creating artifacts that will influence the way people perceive their brand. And I think Stripe has tried to cultivate this academic intellectual brand around the company. They don't talk a lot. A lot of the things they're associated with tend to be more intellectual pursuits. And even, I remember when they built that big office in SF, they had that huge library. So now I assume everything that anybody with meaningful resources does or does not do is intentional and planned in the same way that if you were a movie star, you'd have a publicist who would try to get you onto certain things or tell you don't do this thing, don't go to this thing, don't sit. And I think as tech is no longer the domain of fringe pirate nerds, image management is a big part of what a lot of people spend their time and energy on. I think it pays dividends. I know GPs that LPs have a ton of respect for. LPs never met that person. Their impression is based largely on that person's online presence.

I also love Ohm's pieces, and I can't help but to think, I guess, do you kind of wish we could go back to the time where leaks as news wasn't the way things go, or going direct wasn't the standard for picking up what's going on on the inside of the company? It really used to be kind of like hungry journalism used to scrape up the news, and if they found it, like, it actually could be reliable.

I think what I miss is when I started in tech, there was a set of journalists who'd been in tech long enough that they— it wasn't even just that they had relationships, they'd seen certain movies before. They didn't fall for press and PR spin, and they had their own opinions about this stuff. And some of those opinions could not be influenced or co-opted by clever brand marketing by firms. And I think it led to a better discourse around tech. There was like the fear if you said or did something utterly ridiculous that somebody would come after you and be like, this is dumb, or this is a bad idea. And the few voices I find doing that are either people who are writing independently for themselves, mostly on Substack, or people who, like Om, who feel like they've been here long enough that they can say what they want to say with little to no consequence.

He makes the point at the end that the network does not reward what is true or what is deep. It rewards what moves. I mean, this feels really relevant today when so much of the narrative gets placed around companies well before substance is there. I guess substance in the meaning of like tangible exit or profitability. Are there specific archetypes of founders that you see where the accumulation of symbolic capital almost always hurts them in the long run? Or is it truly more so case by case?

I mean, think about Adam Neumann, like WeWork didn't stop him from doing the, the real estate thing. Weirdly, the ability to consistently attract attention is very valuable in a world that's changing very quickly and where people don't know what the future's going to hold. And there are certain people who are very, very good at attracting— another person who's in the news, Travis Kalanick, incredible at attracting attention to himself and whatever venture he decides to work on. And I think that's largely due to the success he had in the early days at Uber. But that's somebody who, no matter what he does, people are going to pay attention because it's him. But I also think And many of these people who are good at attracting this attention also understand what are the things to say, what are the way to frame things, how do you talk about what you're doing in a way that will attract attention? And I think that is a skill. And I think the interesting thing is I, I talk to a lot of founders who are still running the old playbook, like, oh, I'm trying to get like this person to cover my launch. And I'm like, you could try that. That's sort of not as effective. But if you're not a person who's like naturally gravitational, it's, it's pretty hard.

The one thought that I had before we move on to the next topic is it is fascinating that Stripe, OpenAI, Anthropic are all playing in this game. I can understand it almost more from like the Cluey lens. When you have nothing, you're brand new on the street, no one's ever heard of you, there is a clear reason to break through this, the noise floor, and it's to attract attention so that you can raise capital and get your legs going. But these are some of the companies that have the most amount of funding that we've seen in, in years, which means to your point that the space continuously evolves. And what's the fear, I guess, from Stripe's perspective, that if they lose out on the narrative?

I think the fear is, you know, if 18 months ago I told you Anthropic owns the enterprise AI narrative, how did they do that? I'm not sure. It wasn't on accident. And imagine you're in a Fortune 500 boardroom and you're like, we're going to spend $1 million with one of these two AI vendors. You might be like, well, these Anthropic guys seem to be— all I read about with them is enterprise, enterprise, enterprise. Shouldn't we just go with them? They stood up to the Department of Defense, if we like, or Department of War, if we like that, shouldn't we have done that? And I think part of it is in a world where all of these products feel the same, and we've talked about this before, brand matters. I'll give you another company that I think is doing a good job. I'm not an investor. I think if I say legal tech, the name that comes to everybody's mind is Harvey. Have most people seen the product? No. Do most people know that much about the business? No. But I think in a world where there's tons and tons of products, being the brand that people have heard of or recognized, whether they know if it's good or not, is in and of itself valuable.

And that founder is out there in podcasts constantly, like all over your LinkedIn. And so if you were to look up legal tech, you would know that name. It's synonymous with what you see on your screen.

And I think part of, part of the symbolic cap— the irony is these two things are not separate. Financial capital and symbolic capital are deeply intertwined. Raising a lot of money so that you can be king-made and be the apparent winner before the battle's even over is a part of the strategy.

The symbolic capital piece, I feel like, just creeps a little too early oftentimes within venture. It's just, that's the issue.

I think people are afraid if they don't fight for it early, some other company will take the mantle. And so I, I don't know. I think in almost every cycle, the companies that end up mattering at the end are not always the ones we thought mattered at the beginning, but you don't know that when you're in the middle of it. It's very fog of war.

Well, our next read is a LinkedIn post from Reid Christian, GP over at CRV, and it surfaces something that seems to be gaining real traction in boardrooms right now. Reid notes that he has sat in multiple recent board meetings where CEOs are actively building token output per month into formal performance reviews as a core KPI. So while engineering is the obvious place your mind might go when you think about who's burning the most tokens, Reid says the most variance is actually showing up in business functions. So like sales teams in particular. Have been some of the heaviest users. They're spinning up demos, they're creating competitive analyses and collateral in ways that simply were not possible before. His prediction is that someone is going to standardize all of this and that public leaderboards and incentives are next. Though my honest question is how long we reward that behavior before we find out whether all the token usage is actually translating into customers or outcomes specifically for the sales example. He closes by framing that these board conversations are exactly the reason he is less worried about entry-level jobs disappearing at startups. Arguing that these roles will go to the most AI-hungry and AI-native people in the room. We've talked a lot about how this concept of token output as performance metrics might start creeping into the workplace, similar how previously lines of code were once rewarded in quantity. In the case of business functions specifically, do you think something like a public leaderboard or a widely adopted token usage KPI will actually push people to prioritize upskilling and then drive positive outcomes? Or do you think we're just going to see a lot of people running around aimlessly just to try and hit a Oh, I think this is just like rife with gaming potential.

So like my most charitable vision would be that this ends up being like GitHub stars, which is a measure of some amount of community velocity, but that's really different. GitHub stars for most people are not tied to your paycheck and are not tied to promotion. And my experience is that in any company, if you have an objective that is tied to promotion or pay, people will figure out how to game it. And I have no reason to believe token leaderboards and token output will be any different. People will figure out, they'll run scripts that just do gigantic data pulls. Like, people will find ways. And I, I think what this tells me is the level of, I can't think of a better word, desperation to connect spend with ROI in a really clean and easy way. And also a deep desire on the part of companies to measure how aggressively people are using AI tools. And so right now we're very much, you have two options, it feels like to me. You can use a crude measure. Tokens, or like if you're on a Teams account in Anthropic, you can sort of see API calls. You can get some sense of like who's using this stuff the most, or you can sort of rely on more of the show and tell, or let me show you something I've built, the more bespoke, but the, the latter doesn't scale. Having everybody show you their project and figuring out like who's, you know, it's like a science pro— it's like a poster fair in middle school or something. And so I think people will push token output as a metric until they realize how it's being gamed, and then they will try to address the game and we will get stuck in this war where eventually I think people will say there must be a better way to measure this. I read that and I'm like, having all of these salespeople spin up all of this stuff, to your point, Mia, is it working? Do customers like these? Like, is it working or are they just making a bunch of junk? But there is, I think, in every cycle, everyone's trying to figure out what's the metric that matters. And I think I go back to mobile and then mobile in the beginning, it was downloads and people figured out you could go buy a bunch of downloads on the cheap and juice your— and I was like, okay, well then it's, then it's active users. And I was like, okay, well, you know, we'll, we'll, we control how we count active user, and then people figured out, well, it's actually the ratio of daily users to monthly users. And eventually we got to things that were harder to fake, not impossible to fake, but harder to fake, and actually more aligned with what we realized actually was evidence of health. And we're just not there with AI yet.

And for the more business functions, I do have questions because the feedback loop on if things are working is longer than with engineering. It feels like you can go push to prod and you can test things out quite quickly and see if things are broken. With business functions like sales, customer support, you might stand up a lot of these tools and allocate a lot of spend to it. And yeah, then like months later realize, did any of this work or was the output worth it?

Also, I think this token leaderboard presupposes something that it took me a while to figure out how to articulate. As powerful as AI is, it is not the right solution for every problem that you're working through at work. And yes, you should consider AI in the same way you would consider, should I do this as a PowerPoint, as a spreadsheet, or as a memo? There are plenty of things I work on at Precursor Ventures where I don't use any AI at all because it's not the right tool for the problem I'm trying to solve. And I think overly incentivizing AI use— I get it. Like maybe if you're a company like Shopify or one of those companies, you're like, look, I'm going to do this because it's not so much I want to find out who's at the top. I want to find the people who are at the bottom. I really want to find that it's not so much about rewarding people who are using the most tokens. It's really about sniffing out the people who are not using this stuff at all and figuring out how to get those people out of the company because the perception is that's not the future. I think companies will regret that, but I have a feeling that's the bigger reward or objective than it is figuring out, well, what are the top people doing?

I agree with that. That was a thought that I had, which was this, in like a positive, more optimistic perspective, could push people to want to upskill themselves and invest in their careers and professionalism, etc. I think the drawback is there's possibly people who just like are slower to take on these tools but have all of the top-tier qualities that make them exceptional in these types of business functions, like sales, like relationship building, that don't come through in token usage. And if you work in a big company and that's how you're using it as KPI, those people end up getting dropped from the bottom. That in the long run, I think will be a mistake to your point. He does end with this point that entry-level jobs are actually safer at startups because those employees are the most AI native. So that feels like a real inversion of the dominant narrative right now. Wanted your thoughts on that one before we moved to the next one.

I think everybody's looking for a silver lining here. I'm not sure there is one. I'm not sure it's doom and gloom. It might just be like, it's too early to say who this is good for. The answer could be we're all overthinking it. The answer could be we're all not worried nearly enough.

The last piece is from Daniel Kimmeling titled Choosing Shoulders. The piece is rooted in Sir Isaac Newton's famous line about standing on the shoulders of giants, but Daniel reframes the real question as not who the giants are, but how intentional you are about choosing them for yourself in the first place. He argues that intellectual and moral lineage shapes the ceiling of what you can build, and that picking the right shoulders is one of the highest leverage decisions a person or firm can make, while picking the wrong ones quietly compounds in the other direction. There's a sharp point in here about alignment as well, that variant outcomes are not just the product of variant people or processes. They are the product of those things being deeply aligned with one another. And he closes with a thought that the greatest giants rarely set out to be learned from. They set out to be great, and then the lessons followed. Warren Buffett obviously has the famous line, tell me who your heroes are and I'll tell you how you'll turn out to be. But in today's world where so much of entrepreneurship and venture is driven by media obsession and flashy moments, like as mentioned in our own media piece earlier on, it feels increasingly easy to point to someone as a hero or role model without actually knowing whether they're quote unquote successful at all. They simply appear to be getting to the visible markers of success faster. And that's really tempting to want to copy. So Charles, what do you think the downstream impact is of that on the ecosystem? If we keep building platforms and rewarding founders and characters who are leading with a narrative over substance?

I think part of the other problem is we live in this world where it's easy to have a parasocial relationship with somebody who's famous or perceived to be successful, who you've never met. You don't know anything about that person's personal life. You don't know whether they're actually happy. You don't actually even know whether the success that they're representing online is real. And I think someone once told me, like, if you're going to copy something, make sure you understand how it works. And I think when you marry this with symbolic capitalism, there's like a lot of things that appear to be working on the surface. And the people who are running those things will tell you why it works. And then you get in there and you're like, no, that's not it at all. Like it actually doesn't work or doesn't work in the way that you said, or It really works because this one person like does a bunch of the crap work and doesn't get— I see fund managers saying, well, I'm going to build the next version of XYZ firm. I'm like, do you, how much do you actually really understand about how XYZ firm works?

Did you work there?

Did you work there? Oh, I've read all their stuff and I look at their portfolio. I'm like, okay, those are artifacts. That doesn't really tell you how or why it works. Oh, we're going to be an equal partnership, just like Benchmark. I'm like, well, have you explored the tension of what that means? Have you actually talked to people who've worked at Benchmark about it? It's kind of like when all the Bridgewater stuff came out and everyone's like, this sounds amazing. ask them what it's like on a daily basis to actually work in that environment. And they were like, oh, what do you mean? I'm like, well, yes, in a book it sounds great. And some of these things for certain people are amazing, but it's probably worth getting some on the ground. And I think the same thing is true. I see people, I'm like, look, you're not Elon Musk. You're not. There are things Elon Musk can do and get away with and pull off that you cannot because simply because you are not him. I like Daniel a lot. Like we've done a bunch with Decisions. Like he's great. I tell almost everyone, And people have been like, oh, what would make Precursor easier? I'm like, oh, it's simple. If I just abandon all the things that we do that LPs don't necessarily agree with and made our fund much more cookie cutter and vanilla, I believe fundraising would be easier. I think I would be far less interested in the work. It would not be authentic to me. And I think we'd make less money. But I do think a lot, you know, I've told Ron Conway this, there's a lot of parts of SV Angel that I've tried to internalize and adopt at Precursor Ventures Ventures, but not the whole thing because I didn't work there. And every time I talk to people who have worked there, there are lots of pieces about that place that I think work only because of a few specific people who work there and the relationships that they have in Silicon Valley. And I don't have those same relationships. And so I think it's, for me, it's more important. There are little bits of lots of other firms that I've seen and taken. I'm a big believer in that old statement, you're either a role model or a warning. And there are lots of things I've seen other venture firms. I'm like, I never want to adopt that policy or be like them. But almost everything I've learned, it's been because somebody else before me ran these experiments. And learn some lessons so I don't have to learn everything from scratch.

The last question that I had here is, is kind of like a more fun one. I mean, the piece draws this clear line between studying the giants in your own domain versus pulling from completely different fields. And Daniel mentions restaurants and record producers as things that have genuinely shaped how his thinking operates. I'm curious to know if there's an unconventional domain or discipline outside of venture that shaped how you think about investing or entrepreneurship more than people who are familiar with Precursor Ventures might think.

Wine makers and restaurateurs are two categories of people. They're both foolish pursuits in some ways. Most restaurants don't make it. They're extremely stressful environments. Every, every great restaurateur I've talked to, or maybe even chef has been better than restaurateur, has had some vision or a dish or an experience they wanted to bring to the world that they felt like wasn't being represented by anybody else. And against all odds, they did it and they built something special that left a mark. And it was hard. And I'm sure a lot of people told them, don't you know how hard it is to start a restaurant? Don't you know how hard it is to like have an impact? And winemakers too, because it's a folly. You're like fighting against nature. It takes many, many years before you know if you're any good. And some of the winemakers I've met who I respect the most have done truly crazy things. They've tried to grow a grape in a region where it has never been grown before, or they've tried to go to a region that didn't have much of a developed wine industry and they've worked for 7, 10, 15 years trying to like understand the intersection of land, weather, technique, farming. A lot of these things are like interdisciplinary, but oftentimes most people are going to tell you to quit before you start and will continue to tell you to quit while you're in the middle of it. And you don't always get a lot of feedback along the way that you're on the right track.

Well, we are all out of time today. Thank you so much for joining in, Charles. Thank you to our listeners for hopping in as well. Please like and subscribe and we'll talk with you all next week. Thank you. Bye.