Stacking the odds: Cate Hall on agency and outlier success

This week, Patrick is joined by Cate Hall, CEO of Astera Institute and author of a forthcoming book on agency, to explore how individuals can systematically develop higher agency in their lives. They discuss the selection effects that draw agentic people to fields like poker and startups, the importance of being comfortable with ignorance and feedback, and practical strategies like asking "Is there a better way to do this?" ten times daily.

[Patrick notes: As always, there are some after-the-fact observations sprinkled into the transcript, set out in this format.]

Complex Systems now produces occasional video episodes like this one. In addition to this site, you can access them directly on YouTube. My kids inform me that I’m supposed to tell you to like and subscribe

Timestamps for video

(00:00) Intro
(02:38) Exploring the concept of agency
(03:08) Selection effects in different fields
(03:45) The changing demographics in tech and startups
(06:52) The rise of systematization in poker and startups
(17:27) The role of mental habits in developing agency
(22:25) The value of feedback and continuous improvement
(25:33) The power of admitting ignorance
(27:29) Growth mindset and personal improvement
(28:18) Using AI for creative pursuits
(31:25) The value of human interaction
(34:22) The unique traits of professional gamblers
(39:40) Understanding and developing agency
(43:41) Triggers for increased agency
(51:02) Wrap

Transcript

Patrick: Hi everybody. My name is Patrick McKenzie, better known as patio11 on the Internet. And I'm here with Cate Hall, who is the CEO of Astera Institute and has a forthcoming book on agency. [Patrick notes: If you feel so inclined, you can sign up for the pre-order list.] 

Cate: Thank you for having me on. I'm excited to have this conversation.

Exploring the concept of agency

Patrick: Thanks very much. So I'm obsessed—well, obsessed is not quite the right word, but quite interested in the topic of agency in the last few years. Both how it plays out in our personal lives and how there are some institutions which seem to be sort of hitting above their weight class in terms of empowering people to have it. And then there are very many other institutions which seem to do an excellent job of killing it as early and thoroughly as possible in people's lives.

[Patrick notes: Why is there such dynamism in Silicon Valley, including among young people, but our leading institutions of higher learning almost aggressively clamp down on actual productivity by young people? Why did my professors model receiving a series of increasingly impressive grants, but not doing anything ex- staying on the happy path for grants, as success when I was in my early twenties?

YC doesn’t do that and that matters.

Why do we have such flaccid listlessness in so many of our societal institutions? Why did it end up being Vaccinate who got the covid vaccine on Google, and not e.g. the CDC? Why will I have to expend heroic efforts to get into the room again if there is a “next time”, while the current constituents of the room where it happens enjoy de facto tenure, irrespective of underperformance?] 

I would love to hear your thoughts on it as someone who's thought about this specific topic a lot.

Selection effects in different fields

Cate: Yeah, so I think it's an interesting observation. My head immediately goes to selection effects. 

Because if you think about something like poker, it tends to attract agentic people. I don't think poker causes people to be agentic. But it does have that selection effect. On the other hand, you have something like law, which is where I was before poker, that definitely selects for low agency people. It probably also suppresses agency to some extent, but I think it's as much a question of what types of fields and institutions people gravitate toward as the effects that those institutions have on the people within them.

The tech industry evolution

Patrick: I think that's certainly the case, at least in places like the tech industry, which for many years, and I think to a lesser degree these days, most people would say—you know, the famous Apple advertisement "we're the creative people, we're the thinkers, we're the rebels," et cetera, et cetera. And tech had much broader paths into it than most employers of large numbers of people in the professional managerial class.

[Patrick notes: I lack a convenient ethnography to point to, but I’d note that system administration, security research, and similar cast a particularly wide net, and this was even more true in the past. Those fields have in the past 20 years semi-professionalized, with sysadmins being rebranded DevOps and now being treated as engineers (which they always were) and security research having its own quirky trajectory which, for historically contingent reasons, ends up being a pipeline into tech from the military (and related careers).]

And descriptively the behavior of people in tech was much more varied than most people who are in high positions in society. We have our fair share of weirdos. We still do have our fair share of weirdos. (I'm certainly not exempt from that.)

[Patrick notes: I say “weirdo” affectionately.]  

Patrick continues: But it's become more of a beaten path. And the entering classes at tech companies these days look more like the entering classes at Stanford or Harvard or an investment bank than they did even 10, 15 years ago, I think.

Cate: Yeah, absolutely. I think tech has gone through that cycle that all things do—you know, the geeks, mops and sociopaths cycle. [Patrick notes: I don’t necessarily endorse this essay, but one reason it resonated so hard with people is it points in the general direction of a real phenomenon. Entryism is a bit subtler than sociopathy and entryism is not the only reason why locations/scenes/cultures change over time.] 

 Y Combinator's changing demographics

Where tech has become so prestigious and the locus of so much power in society today that everybody is drawn to it. And I think that it has led to a certain sameness in a lot of the people where it used to be maybe a more diverse population before. I haven't been around tech for that long, so I might just be making this up. But I think if you compare Y Combinator in 2006 or whenever its first year was to what it is today, it seems very different demographics.

Patrick: In the very recent past, YC had more people who were college dropouts than people with postgraduate degrees. And as of a recent class they're at something like 65% postgraduate degrees. Part of that is that they're wisely responding to a change—many of the companies they're funding are AI-first. And so they're getting an increasing number of people from university AI labs.

But part of it is definitely that Y Combinator was essentially pitched in the beginning, I think they would say in a not unfair paraphrase, as like "do this instead of doing grad school" for people who wanted a relatively institutionalized path up versus throwing themselves into the wide world of capitalism and making their way or decamping to Vegas and playing a poker tournament. 

This appealed to people that the existing life scripts weren't quite right for.

And then YC itself became more institutionalized over the course of the last couple years. There are some people who think: “I've been in high status competitions my entire life. I grinded through high school. I got into a great university. I'll then do a merit-based competition and get into Y Combinator. I will power my way up to that, and then I'll find my next merit-based competition to crush.”

That has become a larger percentage of the pool than previously. And I say that out of immense respect for YC founders who did that. [Patrick notes: A YC partner is a surrogate teacher who has a fairly unique job among all teachers: announcing that pleasing the teacher but losing in the market is a failure, to impressionable founders who may very well have a 20 year unbroken string of pleasing teachers and might be encountering markets as a producer for the first time ever.]

Parallels between poker and startups

Cate: I think it's a really interesting observation. So now that you're describing it, I see a really interesting parallel between poker world and startup world. Which is, so there's this old guard of poker players who were maybe the people who were already poker stars around the time that the poker boom started. And they had a particular way of playing that was maybe less mathematically rigorous on the whole.

When there was this huge wave of popularity for poker starting in the mid-2000s, this new crop of people sort of started gathering. And those people were extremely rigorous, extremely ambitious, very young, because they were willing to sort of catapult themselves into this weird, nascent field in order to—because they thought that they could get a really big advantage basically. And I think in poker, there was a perception for some period of time that youth was an advantage because of that. Because the players who were really successful for some period of time were very young.

I see a similar thing in startups where there was a period of time when everybody was like, "oh, all the best startup founders are in their twenties." [Patrick notes: This was, in my recollection, never a particularly rigorous POV, but was popular as VCs pattern matched on the Google founders and Zuckerberg. It became somewhat less popular not because of social criticism, which I think people misdiagnose as causing the change, and more because Silicon Valley spent a decade doing B2B SaaS apps, with the winning founders mostly being a different (and descriptively older) profile.]

Cate continues: And it seems like the conventional wisdom for both of those things has shifted as that cohort has aged and now you see the best poker players are in their late thirties. And the best startup founders and the ones that people really want to give money to, tend to be in that age cohort as well. Maybe in their forties. So it is this thing where there's an initial maybe selection into a way of doing things that is especially attractive to young people and that gets confused for some period of time with differential ability. And now we're starting to see in both cases, that's not really the case.

Patrick: If I can just have a very minor rant—similar in the way that there was an old guard in poker, which might have had great table reading skills. For the benefit of people who don't play poker—and Cate is a professional, I am an amateur who is barely ahead at a one-two table in Vegas—table reading means you are live reading.

[Patrick notes: Poker is like the capital markets: calling it efficient is a bit of a signifier more than a useful truth claim, but it definitely runs on incentive gradients. Poker professionals are independent contractors in entertainment services, and the amount they can charge is determined by stakes at tables they are ahead of by a sufficient amount to be profitable net of the casino’s rake. This tends to sort the best professionals into either higher-velocity games or higher-stakes tables. A live game in Vegas, particularly at 1 / 2 (low stakes), would let you charge working-at-McDonalds money for being consistently ahead of the time. This implies (correctly) that the regulars at a 1 / 2 game are not particularly skilled relative to most poker professionals and therefore that the tables are generally pretty soft. And so me saying I’m ahead of a 1 / 2 game but probably not ahead of a 3 / 5 game in Vegas is a fairly specific bead on my poker ability: talented amateur with major consistent holes in my play. (I play far too many hands, among other things, and adjust only weakly to the behavior of others at the table.)] 

Patrick continues:You are picking up on some physical signal that a person is giving that can give you information as to the state of their hand. Versus people who are incredibly mathematically oriented and have the optimal poker strategy diagrammed out in tens of thousands of games played by AI before they sit down at the table. This borders on magic.

That explanation out of the way, the way we previously evaluated startups bordered on magic in VC land where it was very connection-based. Very like, "are you good at giving the handshake that the partners like," and very based on the partners having an impressionistic understanding of who are the most successful founders. And so after Facebook was in the news, it was like, "oh, all the successful founders are in their twenties, obviously." When a slightly more rigorous analysis—well, wait a minute, if we look at who actually has the exits here. And granted there's some power law that goes into that too, but yeah.

My general theory, both with respect to poker and with respect to startups at various times in the lifecycle has been when there's some terra nova in society where there's a lot of economic value being created somewhere, the people who get most of that economic value in the first years are sort of pioneering spirits. And they're often pioneering spirits because they took a look at all the other ways that you could plug into society and decided, "eh, not quite for me." And so they tend to be a little bit weirder. They tend to be towards the outer ends of all sorts of spectrums. Many of them are—I won't say dubiously legal, but let's say they are more comfortable with gray zones than people in, say, law and similar.

And that applies for the first couple of years, and then the rest of society gets the memo and the value doesn't all get extracted, but people start to bring other societal advantages to bear—like the capability to get investment, the capability to build rigorous organizations that scale whatever the initial advantage is, et cetera, et cetera. And then after a few years, it's not like the frontier disappears, but it becomes more integrated into the rest of society. And then you get a new frontier and go figure the people there are also a little bit weird, also a little bit young, et cetera, et cetera.

So my general theory—I was around for the poker boom, I was seeing it, but not at a point in my life where I was gambling. At that point. Well, I was gambling by decamping to Japan and building my entire life there, but not with money. 

The Internet poker boom was kind of this new field nobody had experience with. Everyone was busted back to square one and it evolved a different player base and different norms and different practices than live poker had. And interestingly when those crowds mix in real life, they have very different personalities and different spikes in terms of skills.

Cate: Yeah. I think there are some really interesting dynamics at play here. Again, I see this parallel where it's like, with internet poker, that internet poker coincided with this explosion of interest in poker, which was initially triggered by Chris Moneymaker, this accountant from I think North Carolina or something [Patrick notes: Tennessee, but Cate means “not from a physical community of practice which routinely cultivates top-of-the-world performance in a sport which keeps extensive records”] , winning the World Series of Poker and winning like $10 million. And this just ushered in a huge wave of interest in poker. Online poker blew up as a result.

This created this massive opportunity for people who were these prospector types. Where there's a ton of money available, there is not a lot of systematization. And so a lot of it is just first mover advantage. It's like who is willing to improve significantly on the state of the art, but a state of the art that is very primitive, in order to capture a bunch of those gains. And so at first you have just sort of brute force, almost systematization, where you start having people using tools, start having them figure out like, really how do you make money doing this in a way that isn't just magic?

And those people take over, but then you do as it gets more and more competitive, it becomes an AND thing rather than an OR thing. And poker, the live reads thing is a great example where I feel like there was a view among people who started playing poker in say like 2007 to 2017, that live reads were stupid, that they were like folk logic or something like that. And so people were just extremely rigorously focused on what does the model say. Let me run a million simulations of the spot and see what happens.

And I think what happened is the skill level in a quantitative sense sort of asymptoted to game theory optimal to a point where there aren't huge edges to be found in improving your game theoretic play a little bit more. I think this was my big thing when I was playing poker because I was way more reliant on live reads than most people. And my thing was always like, this is actually super under-exploited because nobody has tried to systematize this. They've just sort of done the magic version of it.

But then I think in the last few years, in the time since I left, basically people have started relying more on that as an additional source of value. And so you get competition increasingly in all of these different aspects of the game. And it's no longer sufficient to just be good at the math or good at the engineering or whatever is required to start a startup. You have to be good at the whole suite of things in order to climb your way to the top of the ladder.

Patrick: If you plot the evolution of poker against the evolution of online businesses in the mid to late 2000s—there were, of course we had a tech industry. Of course we had startups. Well, people forget how little the tech industry was relative to its current size. Google IPO'd in 2004. And it was possible for me to not know who they were six months prior to that happening. [Patrick notes: A CS prof did make fun of me in front of class for that. Some parts of my personality are visible quite early in life, and one is that (unlike many in the startup community) I am the opposite of an early adopter, and tend to hold onto tools I like for dear life. Who needed another search engine when you had HotBot?]

And then they went up into the right, obviously, but it was called the "make money online" crowd rather than the startup community.

And it was largely info marketers, SEOs, black hat, spammers, and a lot of folks who operated at the socioeconomic margins, let's say in the 2004-ish era. And you know, then there was something of a laundering of techniques from those more marginal businesses into places like the software industry. [Patrick notes: There continues to be a shadow startup industry with less-household names doing less salubrious things, and this is true fractally, where the very respectable parts of adtech are very respectable and they’re aware of but not on great terms with other parts of the ad supply chain including e.g. toolbar bundlers who are mostly not cybercrime operations.]

Patrick notes: I had my first software company in 2006 and was anomalously good at SEO relative to some very big firms, partly because I was willing to read some forums where the witch kings of the Internet were discussing what was working back in the day.

And it's interesting how those informal sites of cultural diffusion—I think what would've been the Two Plus Two forums in poker—sort of give way into more formal institutions. These days you don't get your advice on SEO from forums. You typically get it from either a consultant that you've paid or a partner at YC who's like, "you know, the things that have worked previously are blah, and maybe you should be prioritizing that or maybe you shouldn't."

It seems to me like one kind of under-appreciated lens with regards to agency is who is willing to go to places that are not on the approved list of places you can learn from? Obviously you could just get a degree at Harvard. Unfortunately, they don't teach—actually, that's probably wrong. They won't give you a degree in poker, and they don't teach SEO. And so if you're willing to scour the Internet and find the forum that seems to be giving out some amount of edge and take advantage of it, then you have some time-limited amount of edge versus the rest of the world. Versus if you are just waiting until it has become a book and there are high status people to teach it to you and it will cause no friction when you discuss it in partner meetings on Sand Hill Road, then you're going to be a couple years behind the curve.

So what are other sort of things that agentic people do that might give them an advantage in the early stages of whatever the new industrial wave is, whether it's startups back in the day or AI now or similar?

Cate: Yeah. So I think, even at a more basic level than specific tips, they're just sort of mental habits that agentic people have. And one of them is just simply asking sort of continuously whether there's a better way to do something. This is actually one of the ways that I sort of first started deliberately developing agency—sounds so simple, but I just left post-it notes for myself everywhere in my apartment asking "is there a better way to do this?" And then I would remind myself to ask that about whatever I was doing throughout the course of the day. Like, is there a better way to do this very trivial task that I'm doing? Is there a better way to attack this large scale problem?

And I think if you can train yourself to earnestly ask that question 10 times a day, you're most of the way there on agency. Other things that I think are part of the mindset—there is this desire to look in inconvenient places for real edges. So there are a lot of things that you can do in many situations to improve your odds of success that people just don't do because they're not part of the standard playbook. They're kind of aversive. And they don't seem necessary for success.

The role of networking and relationships

And I think people—what's an example of this? Personally, anything that looked like excellence in networking, I would say was one of these for me, for most of my life where I just wasn't willing to really pursue relationships with people who I had professional overlap with and interests because I just didn't like it. And I think that that's a relatively common one. I think that if you sort of go out of your way to develop relationships with people in the hopes that it will lead to opportunities to do interesting things in the world, you can become a much more impactful person. And I was for a long time just sort of stuck in the mindset of "I'm effective enough. I'm sort of getting what I want to do done in the world. I don't need to do all this other bullshit, you know?"

Patrick: This resonates quite a bit with me. Yeah, many other people—some people get the memo early in life that reality is a team sport and being able to tap people in your network to take care of things is very helpful. And some people get a very corrupted version of this, which is like, "relying on another person is cheating somehow." And you should succeed entirely on your own merits and similar. 

And I had quite a bit of that.

And so there were—to make a very concrete example when I was originally publishing essays, I would be like, "okay, I will certainly never show this essay to anyone before I publish it, because that could influence them to decide to share it to their audience. And I wouldn't want an unfair advantage versus other people who are publishing essays at the same time."

Cate: Oh, I'm so familiar.

Embracing feedback and continuous improvement

Patrick: And these days it's like—granted, I still don't show most of my essays to anybody prior to publishing them. But I'm at least aware that if I was attempting to play this game to win, I would certainly show it to people both to get their feedback on it. And because, contingent on them giving feedback, they feel some bit of ownership about the essay and are much more likely to share it with their audience than they would if they were seeing it cold from Twitter on the day it is released.

And yeah, the constant learning that everything you thought is wrong or portions of it are wrong, on just asking the question of "how can this be better?" generating surprising amounts of alpha or edge, or however one's community likes to phrase that thing. One of the reasons I love poker is it gives you metaphors you can take into the rest of your life. And descriptively, I pay people in Las Vegas a bit of money every year to continue getting interesting new metaphors. They're much cheaper than consultants.

But I worked at Stripe for a number of years and there were very highly placed people in Stripe that essentially had a button that they would push in almost every meeting. Like, "all this sounds great. Do you think we could do it faster?" And the person who had just announced a nine week project plan is like, "well, sure we could do it in seven weeks if it was really important to do it in seven weeks." And then it's like, "okay, move to button number two. I would prefer seven weeks over nine weeks. Let's do it."

And you know, obviously there is still some prioritization involved, but much of the world simply doesn't go about life asking, "am I at the right margin? Am I making the right decision at the margin for this? Could I go faster in a way which doesn't involve other trade-offs? Could I just choose to be better at the thing?" Right? And so you can choose to be better at the thing. You can just do things like be better at the thing. So what are some other things that you've learned about regarding agency over the years?

Cate: I think that a lot of the things that increase agency look like just learning to have less ego about certain types of things. So I think that fear of exposure in the sense of having your ignorance exposed, your mistakes exposed, being exposed to yourself in terms of like, "here are things that other people see about you that you don't see about yourself that are holding you back"—these are all things that are really, really detrimental if your goal is to learn how to do things quickly, improve quickly.

I think that figuring out how to be comfortable getting real feedback from the world is like the highest leverage thing you can do. And most people have a really strong aversion to doing this. They might welcome feedback in certain limited contexts where they know things can only go so badly.

Patrick: And in socially blessed rituals—the meeting with the boss or the teacher. We all understand that there's only a range of feedback I'll get and the feedback will be delivered in a way which is designed to preserve my ego at this time.

Cate: Totally. Yep. And this seems really unfortunate because I think that for a lot of us, we have a couple things holding us back that other people can see pretty much right away and we just don't know about or are just oblivious to it. Maybe we know on some level, but we might actually just not know it at all. And giving people a way to tell you those things in a way that is not threatening for them is so high leverage for your own improvement on whatever dimension you want to improve on. I'm a huge fan of anonymous tip boxes because of this. But I think that there are a wide variety of ways in which you can sort of invite feedback into your life in helpful ways.

I also think that just learning to be comfortable with not already knowing everything is huge. It's probably the number one thing that held me back in life for a very long time was I wanted to—I felt like it was a sign that something was wrong. I probably wouldn't have phrased it this way, but like that there was something wrong with me if I didn't already know something in a given situation. And so I would avoid being exposed as somebody who didn't know the thing that we were talking about, you know, just kind of nod along in conversation, for instance.

And one, this is probably obviously detrimental. Being comfortable just not knowing things and asking questions has compounding effects. I think this is also in a category of things where the effects that certain behaviors or orientations have on people are actually opposite what people think. So things like saying, "I don't understand what you just said." People love that. People love to explain it. They love the honesty that comes with that. Saying, "you know what, you were right about that I was wrong." Like, no better way to authentically get people onto your side and trusting you more.

And yet our immediate reaction, anytime we sense that something is wrong, is to sort of dig in our heels, resist learning that basically, or resist acknowledging it if we do know it. And there are all of these things that we do that are tacitly designed to save face that I think actually have the opposite effect of what is intended.

Patrick: Yeah. I think this is particularly common among people who are broadly high competence. Particularly because they've been tracked in institutions that have socialized them to believe that you start at a hundred points and you can only go down from there. 

And that if you're bad at something, you need immediate corrective feedback from the people who are good at it. And you know, correcting your moral failure and being bad at something rather than, "no, this is just an opportunity to get good at a new thing."

And this instills in many people a sort of risk aversion and aversion to new things. And I think causes them to, with other factors in society, institutions become even more tracked than the structural reality of the institution. They're ordinarily tracked to be even less willing to do things that are a bit outside the box than the social reality of the institution would accommodate. And then since life is a multiplayer game, when everybody's feeling individually constrained like this, the institution adapts to their preferences by constraining the group even more.

Down this path lies much dysfunction.

Exploring growth mindset

Cate: Yeah, absolutely. I think that I actually think that I got some mileage out of learning about growth mindset as a concept the first time that I did. Because it's really apparent once you sit down and think about it, just how much personal improvement you are leaving on the table if you're inflexible in these ways mentally. And it's like, I think it's actually hard to be extraordinary at anything unless you can learn to overcome some of these perpetual mental tensions that people have. And it just really accelerates growth if you can learn to not tie your self worth up in, "did I look like an idiot in that conversation?"

Patrick: I'm neither here nor there, but a thing that I've been loving in the last couple of years is now, rather than having a conversation with an expert in which one feels one can lose face, one can have conversations at a variety of points up the skill ladder with an AI, and often not feel like the AI is going to remember and judge you harshly for it, although they might both remember and judge you harshly for it. But who cares? It's an AI, at least for the next couple of years.

And so I've been trying to get better at painting the last couple of years, and I was the kid who could never draw even a stick figure successfully. And the last couple of years I'm like, "well, art is a nice hobby to have. Sure, I've said I've sucked at it my entire life, but I've succeeded at other things that I once sucked at. Why not throw myself at this?" And just having a record of three years of photos of finished pieces and, oh, I think I'm actually descriptively getting better at this has been a bit invigorating.

I get fun things to decorate my room with. 

[Patrick notes: A representative example of current skill level compared with three years ago.]

AI and personal improvement

Cate: Have you been using AI in this pursuit in particular?

Patrick: Actually, yes. Okay. So this is one of my favorite, least recommended ways to use an LLM. When I'm halfway through painting a piece, I will—and I'm a very, I would say probably low intermediate painter at this point relative to skill levels in the community—I will take a photo and say, "okay, here's a half painted piece. And my reactions to this are, I think this troll is a little too dark. I want him to be green, but not cartoonishly green. What could I physically do right now to achieve that effect?" And LLMs are surprisingly good at giving you action feedback about what things to do in the physical world in a totally non-judgmental fashion.

So, yeah. "Okay. Why don't you glaze on a few layers of this kind of paint and maybe put on a wash with this," blah, blah, blah. No one cares about the painting mechanics, but the LLM is happy to entertain you about painting mechanics, if that's the thing you want to talk about that day. And then I will do it. I will say, "okay. I think I achieved most of what I was hoping to get out of there." And then ask the next question like, "alright, here's the new photo. How did I perform against the plan? And what can I do to punch up?" I don't know. On the troll, there was a gem. I highlighted the facets of the gem and I thought, "it doesn't look quite gem-like. I've never painted a gem before. What would one do to drive up the gem-ness of this?" And they all had some suggestions and they worked.

And so I have a gem and a not-so-dark, but still green troll and feel happy about this. There was also hopefully some learning. There were a couple of techniques that I wasn't aware of that can apply to other things in the future.

Learning with AI: strategies and cautions

Cate: Yeah, I think I have—so I am a huge fan of LLMs and incorporating them into learning processes. Just because they're so efficient for that, you know, I have, as I'm sure most people do, a few different ways of using them. One of my favorite ones for learning a new subject matter area is start with "explain this to me like I'm five." Then like 10, 15, like, you know, you just escalate the skill level. And I think that you can get up to speed pretty quickly with things like that.

At the same time, this is something that I have started watching out for in my work at Astera—is I think it's easy for people to become really insular because of reliance on LLMs in this way. And so I try to not over-optimize, I guess, my—head of strategic investments, Eli Dourado, absolutely love working with him. One of my favorite things is he's adamant that he always wants to hear stupid questions. So instead of, you know, when I'm reading something of his, going to my LLM and asking about it, I just walk over and ask him all of my dumb questions and he gets to answer them. And I feel like, although I could probably extract the same information from an LLM, there's actually something that is lost in terms of the organizational bandwidth or something from having all of those touch points. And so I'm just trying to be selective about it I guess.

The value of human interaction

Patrick: I would two thumbs up to that. I also think there's—we have some options with regards to rewiring ourselves, but humans are fundamentally social species and there is something about getting the information from another human that is not the same as getting it from text. The people on the podcast or other people who listen to the podcast have sometimes told me that a particular thing that I've written about for many years didn't stick with them until they heard it in their ears at the gym or similar. And I think we should be aware of features of our minds like that and take advantage of them where possible.

I also feel like, you know, there are some happy mediums where if you know that you are going to have a conversation with Eli or with another person who's an expert in an area, you can make that conversation more efficient for yourself and for them by, "Well, there's a couple of obvious icebreaker level questions that I could ask. Maybe I'll get those out the way with the LLM and in place where it's relatively unlikely to hallucinate. And then catch more of the lingo the first time around. And like I had my first question be a decent question, like, 'okay, what I got out of that was you just said the EBITDA goes up. My understanding of that is X. Can you check me on that?'" And then again, people love being asked confirmatory questions.

I almost—not that it's never happened, but there are people all up and down the skill curve and social status curve and similar, who absolutely love, like if you have a 101 or 102 level question about the thing. "Oh yeah. Happy to explain that on a dime to a smart person. Here you go." Or if you ask like, "an implication of the thing you just said—it seems kind of basic, but is it X?" And they're like, "well, yep. That is indeed a basic implication of that, but I'm glad you caught it 'cause so many people don't," yada yada.

Professional gamblers and their unique insights

We had an episode recently with Zvi Mowshowitz, another professional (well, former professional) gambler—about his tips and tricks for prompting successfully. And that caused me to reflect on: there are a few populations which punch way above their statistical weight in terms of the people who are involved there tend to lead interesting lives. And not that society's highest award is being a guest on this podcast, but I interview a grossly disproportionate number of professional gamblers. It seems like I've had at least three or four out of 50 interviews, and I can't imagine that there's—what would that imply in the American population? Let's see, times 6 million. So like 24 million professional gamblers in the United States? No, I would take the under on that one.

[Patrick notes: Selection effects, of course, but why selecting for that particularly? I have an explicit and implicit curation function, and you would extremely expect e.g. someone who previously worked at Google or perhaps a mid-seniority federal employee to get through those functions. Those people are numerous in the world and in my networks, and yet through whatever alchemy I don’t think I’ve yet interviewed any]

Patrick continues: Why is it? And there are a number of people I could point to in Silicon Valley too who are trying to put those days behind them.

[Patrick notes: A number of communications departments have had an unscheduled meeting where a typically low-seniority employee, identified as their employee, has gotten deep on the World Series of Poker. Why does that cause a meeting? Well there is an employee of yours on TV, that’s almost always worth a meeting, and there are professionals who are verbalizing ongoing commentary on e.g. their risk tolerance and repeatedly mentioning how strange it is that an amateur poker player employed at your firm did this well. And then there are some quiet lingering cultural fears that of course tech professionals are well-paid but we don’t want to make it obvious that 22 year olds can trivially afford a $10,000 buy-in.]

 But why is it that people who decided at one point, "yeah, I'm going to gamble for money as the way I put food on the table," end up doing very interesting things in the world?

Cate: It's a great question. I think there's probably multiple factors. So one is just the types of personalities that poker selects for. So these are not always the most organized people. They are, I would say, first principles thinkers. A lot of the time it's people who are maybe slightly hostile to social scripts in a way that makes them rebellious. Not deferential to the way things are currently done. I think it just selects for people who have a lot of degrees of mental freedom.

I also think it obviously attracts people who have a certain orientation to the world that is like thinking in bets. And I think that way of approaching the world is just inherently adaptive in modern society. Sort of thinking systematically about, given the whole range of possible ways I could spend my life, how do I want to decide where to allocate my time and resources? That's the kind of thing that poker players start doing as second nature. I'm sure it's similar for a lot of other professional gamblers. And I think that that tends to serve people well once they get out of poker too. Because it tends to cause people to think deeply about the things that they can do that might be most profitable, most impactful. It's kind of like an optimization of effort and a translation of that into results. And so I think that systematic orientation probably just leads to a lot more outlier successes.

Patrick: Yep. I would also say, and hope this isn't too controversial, that I think that sort of meta optimization and thinking often causes people to select out of being professional poker players after a while on something like the notion that I can see what the distribution of results is in the poker community and I can see what the distribution of results is outside the poker community. And given the variance involved in being in the top 10 of the World Series of Poker every year—modulo some financial engineering that people do these days to make that lower variance than it used to be—maybe I could tone down the disagreeableness a little, get a job somewhere in trading or many of the other paths in life that are more amenable than people would think to folks that come out of the community and then achieve something other than what is ultimately a zero sum game at the poker table or a negative sum game given that there's a rake.

Cate: Yeah, I think that's right. I think that's right about a lot of the subspecies of poker players anyway. There is a certain type of poker player that just loves the gamble, you know? And there's no real substitute that I'm aware of in life for going deep in a poker tournament. And I think it's too much for some people to give up. But lots of people who are a little bit less driven by that dopamine—I love dopamine. Nothing against dopamine. But able to be a little bit more dispassionate, but who have the other sort of traits of drive and intelligence and being rigorous about the way they think. Yeah, I think that there is probably something to that.

Patrick: Yeah. I sometimes feel the same way at a $50 buy-in tournament, which is objectively economically irrational for me to participate in [Patrick notes: A typical first prize might be $2k, and even if I won every time, 6-9+ hours of playing poker mean the effective hourly rate is well south economically rational for me].

This is why I only do $50 tournaments, because if it ever got economically rational, I'd have to make decisions there.

And I'm like, "no, don't let the system that would be making that decision be in a position to make that decision for me. No good would come of it," based on my broader values and what I want to accomplish. But again, do quite enjoy the game.

Understanding agency and its development

Stepping back to the agency topic for a bit, so you have a book coming out? Yes. What is the angle you take in the book?

Cate: It is intended to be a practical guide to agency. So the essential framing is I've experienced sort of intense highs and lows in terms of agency in my life. So I have done some really impressive things, like I've been the top female poker player in the world. I have run this foundation. I've been a Supreme Court lawyer, all of those things. At the same time, I've also had periods of severe addiction, including a three year period of addiction in my thirties that basically completely crashed my life.

And so the frame of the book is like, how can you—how does one go from very low agency, a very low agency state, to a high agency state? I think this is a bit of a response to the general attitude I see. Which is like, agency is an inherent trait. It's like, I think it's like intelligence or is unmalleable, is intelligence. I don't think that that's true at all. I think people have not put that much thought into how to systematically develop agency, but there are ways of doing it. And so it's meant to be a practical guide to like, here are the things that you can do to become more agentic and to have a life that better accords with what you want. And yeah, it's kind of like a blueprint for increasing agency.

Patrick: I would absolutely endorse the "agency is not a fixed quantity over the course of your life." There's—I can't remember in my personal experience of life a feeling like I'm 20% smarter than I was yesterday, or 20% more capable.

But I've definitely had several hundred percent swings in agency on a yearly basis. [Patrick notes: And in the relatively recent past, too!]

And partly it's just deciding to do it. Go figure. Partly it's—some of what are objectively the most agentic decisions I've made (at the point in my life I made them) were to select into an environment that really anti-rewards agency. Like, okay. Be an American, in my situation, immigrate for economic purposes to become a Japanese salaryman. Like one of those is incredibly agentic, though not a great decision. But the other of them—Japanese salarymen are not typically rewarded for doing things that are several standard deviations away from the median of their corporation. And then heavily socialized myself into that for a few years. And then recovered, I guess. [Patrick notes: It’s complicated.] 

And you know, open question. I think we spend a lot more effort on a societal level on constraining agency than we do on inculcating it. Particularly the school system. I loved schools when I was going through them, but oh my, do they want to chop off both ends of the distribution and they are brutally effective at it. And if you as a kid who is on either end of the distribution object to it, or if your parents object to it, you are now—like, the objection is itself a problem to be managed by the school. 

And it's interesting to see people with sort of alternative schooling modalities and similar that are trying to avoid that. It is an open question on whether any of them wins out. It may be that something of what we want to get out of school, and the economies of scale that we want to get out of it, might be almost inherently inimical to having kids just doing what they want to do. And of course, you know, group of seven-year-olds doing what they want to do. Like you want to encourage agency among certain dimensions, but discourage it among the dimension of like, "please don't stick that scissor in someone's eye"—that is not something that will be effective for anyone.

But what are some of those concrete things that people can do to inculcate agency in themselves?

Triggers for increased agency

Cate: So I think even before that, I'm curious. Because one of the things that is most interesting to me is I think that, as you said, there are these step changes in agency that people experience. And I'm really interested in the question of what triggers those changes. I think that I can identify a few things in my life that were proximate causes of increases in agency, but I would love to hear what your take is on that from the perspective of what were the events that sort of precipitated those agentic decisions for you?

Patrick: Something of it is learning from the environment or peers or similar where there's a person socially close to you for whatever reason that is far above the mean and where you can recast agency as something that you actually want. I feel like I grew up in a culture in which "that person is ambitious" is inherently a negative moral judgment with respect to them. And then got into the tech industry where ambition is broadly much more lauded, and more important than simply being in an industry where it's lauded. Being able to see up close to the example of these people are clearly ambitious in a good way. And like just having the notion of you don't have to simply walk the default trajectory given the educational attainment and career position, et cetera that you have in life. And then project that forward by a few years and get the gold watch at age 60, which—that sounds like a joke to younger people. That was a literal thing that was present in American society. And then for some of us who were in the between the time when that was a literal thing and today, it was like—not a thing you could realistically expect, but it was at least a life script for you, right? Find an institution, glomm onto it. Ride it for the rest of your career.

And so meeting those people who are—do we have a word for that? You, ambitious. That's the word. The characteristic of ambition that you are a wild animal that has a bone to chew and you just chew it. Sounds like one of those wolf memes on Reddit. But so meeting people like that tended to trigger more in me.

There's been a few times where it's been response to an emergency situation for good or bad. The VaccinateCA experience—people were dying. What do you have to do? There was a private emergency in my life many years prior to that, that triggered a fundamental reassessment of what I wanted out of life. And I stopped playing World of Warcraft and started getting serious about making a software company after many years of saying I would like to have a software company one day, but there is World of Warcraft and the guild is meeting at eight o'clock. And if I'm not there, we definitely won't kill the dragon in two hours. 

I would advise having role models is an easier way than having personal tragedies. But you don't always get to pick things that happen to your life.

And then I think culture is a real thing in the world. I think simply talking in more places that this is one axis that you can judge the human experience on. And broadly speaking, there are positive things that—you know, being slightly more calibrated to the agentic end of this axis than a society will by default condition you to be is a good and righteous thing. And thus a podcast about agency.

Cate: Okay. That's very interesting that those line up fairly well with I think the things that were triggers for me too. Definitely having the experience of having peers that are higher agency is undefeated as far as I'm concerned in developing people's personal agency. A lot of actually—a fair amount of my beginning to think about this question systematically came out of my experience co-founding Alvea, which was a pandemic era vaccine company with three other people. And those people were significantly higher agency than me. And I think there was some sort of shock to the system that happened from seeing how radically they moved through the world.

We—so one of the things that we did early on in the company was we set a goal of submitting a clinical trial application for our new shelf stable vaccine within three months of when the company was founded. This is a truly insane thing to do, but we did it, and that was—and it was accepted later. This was driven entirely by the agency and personalities of the two co-CEOs at the time, who just simply refused to accept any reason why we weren't gonna do it. And seeing how different their mode of operation was from anything that I'd seen before was again, a real shock to the system. So that seems huge.

I think of a lot of what we try to do at Astera as being about trying to increase the agency of scientists and technologists who come work with us. Because we're often bringing people in for a limited period of time, helping them launch things and then spin out.

Emergencies are good for this too. I think that my closely related thing is addiction and what I've thought of as "the gift of desperation." That's a common phrase that addicts use. And the idea is just that you at some point in addiction become so stuck that you are absolutely desperate for change and it switches you into this different gear in terms of what you are willing to do to get out of that. And I think that I experienced that with addiction. I think I experienced it honestly to some extent with bad relationships that have ended or being a lawyer. There are moments where you come to feel so stuck in whatever life you have that it just lights a fire in you. And so misery can be an adequate nudge.

Patrick: That was definitely part of what was going on. Yeah. A few times in my career, after—sometimes people ask me what I was thinking when I wrote a thing about a particular thing, and I'm like, "well, there's the surface level of what I was thinking about the topic I was writing about, but the meta level was I have no way in the day job out of the present circumstance I find myself in." And so if there's gonna be any advancement in life that's going to come from the thing I'm doing outside of the day job, if that is writing right now, then I'm gonna write the hell out of it.

And yeah, many of those points resonate with me at various points in my life. I would generally suggest that truism is you become the average of the five people you're closest to. But people should be very deliberate with who they spend the majority of their time with. Yeah. That is an excellent way to get their outcomes in always good and ill.

We're at about an hour-ish. Normally now would be about when we start wrapping up. But is there another thing that we didn't talk about that you would love to get on the tape as it were?

Cate: I feel pretty good about it. Yeah. Cool. Yeah.

Patrick: So Kate, thanks very much for coming on the program today. Where can people find out more about you on the Internet?

Cate: So I'm @CateHall on Twitter. I also write weekly on Substack at Useful Fictions on agency and related topics. And I have a book coming out in hopefully nine or 10 months now. (You can get notified when pre-orders are available .)

 And yeah, I think Substack is my preferred platform these days. I'm very pro-Substack.

Patrick: I am extremely pro-Substack as well, even though I don't use them. It's nice to see more long form writing in the world and knock on wood, hope that continues. Be agentic, write more stuff for the Internet, please. Alright, thanks Kate for coming on the program today.

And for the rest of you, thanks very much and we'll see you next week on Complex Systems.