Tyler Cowen & Daniel Gross - Identifying Talent - [Invest Like the Best, EP. 277]
What this covers
Patrick talks with Tyler Cowen, an economist, and Daniel Gross, founder of Pioneer and formerly of Apple and Y Combinator, about their book Talent. The conversation moves from definitions to practice: what to ask in reference calls, why people's downtime matters, and how age and experience should be weighed. An investing frame recurs. The Moneyball example and the idea of a Signal Half-Life are used to argue that any metric loses its edge once competitors copy it.
On intelligence, Cowen cites the median Swedish CEO sitting at the 83rd IQ percentile and a modest earnings gain from moving someone from the 25th to the 75th percentile. Patrick pushes back that a baseline floor still matters. Gross attacks self-report personality tests as gameable and inconsistent, calls the Big Five Aspect Scale a low-resolution compression, and favors Proof of Work Evidence such as open-source code. He still values the frameworks as a shared team vocabulary. Two hiring claims stand out: consensus processes screen out the outliers teammates love or hate, and conscientiousness is overpriced for lower-level jobs because of adverse selection. Gross also describes Apple's way of tying responsibility to the time needed to recover from a bad decision.
The contested turns come later. Cowen argues that ADHD and similar traits fit a High-Variance Model with higher upside. He discusses how hard mentorship is for younger women, and he is uneasy about evidence that where people end up predicts talent better than where they started. Talent is also treated as a bottleneck on progress rather than capital, an argument credited to Antonio Gracias. The discussion closes on happiness and on reversing the search by asking who will find you.
Cowen and Gross argue that talent is the ability to positively manipulate one's environment toward goals, and that it is best spotted through prior work, energy, curiosity, and durability rather than IQ or self-report personality inventories, which are low-resolution and easily gamed. Their central point is that the valuable edge lies in what others underprice or fail to learn, not in what is generally admired.
- Prior body of work and observed behavior predict talent better than IQ tests or self-report personality inventories
- Talent is most valuable where it is undervalued or where comparative advantage in spotting and recruiting exists, since widely used signals get copied
- Outliers and misfits, and better routing of people to roles, drive outsized success more than generic credentials
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
IQ is very often overrated: above a basic threshold required for many jobs it correlates poorly with success, illustrated by the median Swedish CEO sitting at the 83rd percentile of IQ, which is good but not exceptional.
“The median Swedish CEO is at the 83rd percentile of IQ, which is pretty good, smarter than 83% of the rest of the Swedish people on average, but it's not exceptional.”
Sports analytics advantages such as Moneyball's undervalued three-point shot erode once competitors learn the same metrics, so lasting edge requires learning things competitors are not learning, and metrics-only approaches leave one behind the sector.
“the problem is other people learn them. So Moneyball only worked for a few years or in the NBA for a while, three-point shots were underrated. That hasn't been the case now for decades.”
Like quantitative investing, signals in the talent market have a half-life because competitors arbitrage them away, so the market for talent is an arms race in which useful signals decay over time.
“there's an arms race there just like in quantitative investing in public markets, signals have a half life generally speaking”
The social science papers Cowen trusts show that factors such as IQ or conscientiousness are only so important, and these replicate, whereas papers claiming a factor is very important tend to fall apart or are so context-specific they are unreliable.
“The papers that make bigger claims about something being very important tend to fall apart or they're so context specific you have to be very careful with them”
Abundant financial capital is not the major constraint on innovation, so talent is the rate limiter on progress, and the binding need is a layer of talent beneath rare geniuses like Elon Musk that enables them to create leverage, as argued by Antonio Gracias.
“there's abundant financial capital that doesn't seem to be a major rate limiter in innovation, or great new companies, or institutions being formed. So talent seems to be the rate limiter.”
Personality frameworks are valuable mainly as a shared language for teams, since everyone understanding exactly what openness or conscientiousness means enables more precise communication, even if the frameworks should not be used to select or measure people.
“They provide a language and that is quite powerful. I think language was probably the original technology of our species in scare quotes.”
What is labeled a disability, such as ADHD, is often not a handicap in the workplace with division of labor; people with ADHD often hone it into a propulsive mechanism, switching quickly from finished tasks to new ones, and on average it is correlated with success rather than failure, though the variance is high.
“It becomes a positive propulsive mechanism and it's a high variance thing to have. But I think again, on average, it's correlated with success, not failure.”
Where a person ends up, not where they come from, predicts talent more strongly than Cowen previously thought, because talented people concentrate in hubs such as South England, New York, the Bay Area, and parts of India, and talented people make the effort to reach these places, a finding Cowen finds discomforting given possible political implications.
“geographical location predicts talent more than I had thought”
The most significant friction holding back younger women's talent is the difficulty of developing mentors, because men fear harassment accusations and women fear harassment, and current institutions for men mentoring women are weak, even though protecting women from harassment is necessary.
“it is much, much harder for women, especially younger women, to develop mentors in today's world.”
Social science has not applied machine learning methods in depth, such as clustering text and images from people's online behavior, and the field remains far from the rigorous tooling used in industry, leaving a gap Gross calls a missing OpenAI for psychology.
“No-one's really brought the arsenal of tools that machine learning offers into social science with the correct level of depth.”
Although talent is a reflexive domain where nature does not respond like falling balls, some quant funds consistently outperform the market over long periods, showing that reflexive dynamics still allow a sense of real truth, and talent science need not be adversarial like markets.
“there are quant funds, you would know better than me, that outperform the market consistently. Some of them for a very long time. So there are ways, even in reflexive dynamics, to have some sense of real truth.”
Concerns about income inequality are the flip side of talent misallocation, and the economy operates on bureaucratized allocation processes that are good enough to keep GDP from falling most years, though far better allocation is possible.
“that is simply the other side of the coin of talent allocation. I would say talent misallocation.”
Constantly pursuing pleasure produces type-one happiness, but the brain's circuitry is not built to sustain it, so the rollercoaster-style pleasure stops being exciting within days, and striving for constant happiness may not be the right goal.
“The circuitry in the brain is very much triggered not to want that 24/7. So even if you're on a rollercoaster every day, by the 10th day, it's not that exciting.”
Measuring people by the outcomes of their prior work is a far better predictor of talent than psychometric or IQ tests, and in software engineering, quantitatively assessing open-source code can identify the best engineers, making proof of work the most useful quantitative signal available.
“just looking at the code someone's written is going to be far more predictive, I think of any type of IQ test or psychometric tests are going to give them.”
All personality tests, unlike IQ tests, are self-report, so they suffer from gaming, where people describe the personality they want to have, and from inconsistency, where repeated testing on the same day yields very different results.
“they're all self-report at the end of the day. These are people assessing their own personality. Obviously you have gaming issues where someone's assessing the personality they want to be as opposed to where they are.”
Although intelligence is overrated at the margin, there is a baseline of IQ below which top-tier talent becomes very difficult, and people in highly intelligent environments underestimate how much this baseline matters.
“When you leave that world, you realize there is a baseline of IQ that matters quite a bit for the roles that we're talking about.”
Moving someone from the 25th percentile of IQ to the 75th, a 50-percentile-point leap, yields on average only about 10% to 16% higher earnings, a modest return.
“If you move someone from the 25th percentile of IQ up to the 75th, that's a big leap, 50 percentile points. They will on average earn 10% to 16% more.”
Current personality frameworks such as the Big Five Aspect Scale are a very narrow, low-resolution compression of human personality, so they capture only a small fraction of what determines how someone will perform across environments; the social sciences are at roughly day one in this respect.
“we're trying to take an incredibly high resolution image, which is someone's personality and predicting how that actor will work in different environments through a very, very, very narrow constraining lofty compression filter called the Big Five Aspect Scale.”
Machine learning on large behavioral datasets, such as TikTok's, implies there are more than five macro personality types, so the Big Five framework is an overly simple way to think about personality.
“if you look at what machine learning does in other fields, like say with TikTok, you realize there's more than five graphs on TikTok.”
For lower-level jobs, conscientiousness is overpriced in hiring because the organization already knows which of its conscientious workers are good and keeps them, so the people available to recruit are adversely selected as lemons, and one should be wary of chasing conscientiousness for its own sake.
“if you're trying to hire away conscientious people from other organizations, the organization knows they're conscientious. They want to keep the good ones. They let the lemons come to you.”
Consensus-based hiring mechanisms optimize for even-keeled people, filtering out outliers. The most mispriced and important talent is found in people whom teammates vehemently love or hate, because such people are outliers on more than one dimension.
“if you have consensus mechanism for hiring talent, you end up optimizing for very even keeled people. The most important thing to pay attention to, the most kind of mispriced people, so to speak, are ones where people on the team vehemently love them and vehemently hate them”
Apple's framework for responsibility ties each role to the time required for the company to recover from an erroneous decision: about three to four hours for a junior engineer, a week for a senior engineer, two to three weeks for a manager, a month or two for a director, two or three quarters for a vice president, two or three years for a senior vice president, and roughly a decade for a CEO's strategic error. Age should be judged by this cost of error.
“A engineer, ranking file engineers, maybe three to four hours. Slightly more senior engineers, maybe a week. Manager is maybe two to three weeks.”
The high-variance model of disability implies that people with unusual traits deserve closer look because the upside is higher, and that once understood, they may be easier to recruit, since accommodations signal that the organization understands them, so the recruiting chance is actually higher.
“I think high variance is the model. It's not that everyone with a so-called disability is actually a super person, but it's assigned to look more closely”
Flattery from people who want to use you is a serious risk for successful founders, and true compliments can be even more dangerous than false praise, because a source that is accurate can be trusted, making it a more effective channel for manipulation.
“If the compliments are true, it can be worse. The people who BS you and say things nice about you that aren't true, there's actually maybe a greater chance they're the ones who want to use you.”
Talent-seeking should be reversed to ask who will find you, because you want to be neither too hard nor too easy to find, and attracting the right people is at least as important as one's own judgment of candidates.
“I try to focus a bit more not on whom can I find, but who will find me? And it's thinking about the problem in reverse”
Talent discussions often confuse what is good in absolute terms, such as smarts and conscientiousness, with what is undervalued in the market. Pursuing the absolutely good traits is futile unless one has a comparative advantage in recruiting those people, so the real opportunity lies in differential factors one can spot or recruit better.
“So not to confuse absolute importance with what's underrated. So you have to look for the differential factors that you can either spot better or recruit better.”
Talent assessment needs a durable, physics-like science that can predict how a person will respond to different environments over years from roughly an hour of interaction, and such a science could accelerate the economy and technological progress if it existed.
“I'd like to have something that's as durable, meaning it's a science that'll last at least a hundred years.”
Talented people have a greater default optimism, expecting things to be great and continue to be great given the same stimulus, and it is an open question whether this is biomechanical, such as a gut-brain or vagus nerve effect, or psychological.
“the default optimism I find talent people have is a bit greater. The belief that things will be great, things are great, things will continue to be great.”
What someone does with unobligated time is a revealed-preference test of talent, because it shows how they weigh opportunity costs among many available options, and it is more informative than a test.
“the real test is what you decide to do every given second or moment and how you weigh your opportunity costs.”
How readily a person figures out his or her role on a team is harder to measure than smarts or conscientiousness, and this subjective skill is where an organization might retain an enduring comparative advantage.
“does the question, how readily can a person figure out his or her role on a team? Much harder to measure than smarts or conscientiousness”
Mid-century American university presidents at top schools were often in their late twenties, but today a 60-year-old might be considered too young, an example of one of dozens of institutional rigidities that prevent dynamic talent allocation.
“mid-century it was not so unusual to have American university presidents at top schools in their late twenties. Now, if you're 60, you might be considered too young.”
The base rate of humans who combine strong engineering or numerical ability with strong interpersonal skill is very low, and society creates outliers in one direction far more readily than in both, which is a real constraint on great companies.
“It's very rare to find someone that has both. That's actually a real constriction.”
Many talented people are sorted into the wrong jobs because the operating system used by those who find people is outdated, and better routing of people to roles would produce more successful companies and lower startup failure rates, as an inefficiency that will only be fully understood later.
“there's a significantly outdated operating system running on the minds of different people around the world who are responsible for finding the right person for the right job.”
A hiring manager should feel better about a candidate once they understand what is wrong with that person, and should push referrers to explain how the candidate is uneven. Candidates who look perfect on paper but whose weaknesses are not understood can be costly mistakes.
“I feel much better about a hire when I know what's wrong with the person.”
Successful tech founders skew young largely because those who become founders are often outsiders, such as misfits fired from large companies or who did not fit into college, and the internet provides them an employment and company-building route, so youthfulness is a correlate rather than the innate cause.
“Now, part of it is that those people don't actually fit into any other part of society. So it's not something innate with the youngness.”
Outsiders take more chances because they are not surrounded by peers who expect a specific path; Cowen's most successful career risks stemmed from being at George Mason rather than Harvard, where status games and high-impact publishing expectations can be a dead end for real influence.
“If you're an outsider, you take more chances. You're not surrounded by peers who want you to do a specific thing.”
To learn how a candidate works in small teams, ask about a domain they know well, such as a sports team's failure or dynamics in a film, because candidates cannot prepare for such questions, and their answers reveal their understanding of hierarchies, resentments, and psychological mechanisms.
“ask them about a sports team. If they know basketball, the Los Angeles Lakers, why didn't they do well this year? Now you're not testing them about the knowledge of that.”
The signal from a reference is a function of how close the referee is to the candidate, so closer referees give more honest reads, and asking referees to stack-rank a candidate on a one-to-ten scale and then probe why that number was chosen yields more signal.
“The signal that you're going to get from a reference is strictly a function of just how close you are to that person.”
References are weak evidence for very young candidates and for creativity, because it is unclear whether the people who know a teenager or a future-famous figure like Mark Zuckerberg or Peter Thiel at 18 had the judgment to recognize their talent, so interviewers must use their own judgment.
“why should you trust their judgment over yours? Say it's someone who's a teenager. Who is it you think they know that has better judgment than you?”
For leadership roles, a person should never feel that they have achieved enough; a restlessness that always seeks the next thing is desirable, and a person who can be made happy by a set of achievements is a negative signal.
“For leadership roles, it should be there's never enough.”
After building his own psychometric test and then reassessing, Gross greatly reduced his belief in psychometric tools, keeping them only for team taxonomy and language, and increased his weight on prior body of work instead.
“I've greatly reduced my priors about all of those things with the exceptions that I mentioned earlier, so as a way to provide taxonomy and language to a team”
Talent searching is more like an art than a formula, since fixed rules fail, and the most successful talent scouts such as Peter Thiel operate with intuitive, almost moral judgment rather than formulaic methods, which is notable given his humanities background.
“more and more I think of the quest for talent as a kind of art.”
Elon Musk's ability to attract executive talent is a reflexive loop: intellectual gravitas and market success draw talent, which creates a star-power moat together with regulatory benefits, so the shortage may not be at the executive bench.
“a lot of what's going on for Elon today is he can attract talent no-one else can attract. That's part of the star power moat that he has”
Chasing publications in high-impact journals is often a dead end for real influence, because talented people who play the prestige game can win it without improving the world, while outsiders from other places more often do totally different things.
“So if you're at Harvard, you want to publish a great number of papers and high impact journals... for a lot of people, it's a dead end, in terms of actual influence or improving the world.”
Certain very high-level jobs follow a multiplicative model in which the person must be in the top tier on seven or eight distinct qualities, such as energy, work ethic, and IQ, and if any single one falls away, the person cannot succeed at that level, as with Bill Gates.
“some jobs that fit what I would call the multiplicative model. Take someone like a Bill Gates, someone like that has to be in the very top tier of at least seven or eight different qualities.”
Durability, the likelihood a person will persist, is the hardest trait to assess in a brief interaction, harder than energy, smarts, or conscientiousness, because older candidates have track records while younger candidates lack one.
“I find that much harder than assessing energy or smarts or even conscientiousness. Durability. They'll stick with it. They'll persist. Very hard for me to get a handle on that.”
Organizations should be structured so that highly specialized, unevenly skilled people, such as a von Neumann-type who only does math, can focus on one thing and be shielded from HR and general organizational demands, especially since such people often stall in large companies and thrive in small teams.
“The best way I think to go about that is to have an organization where you can structure it such that the von Neumann is only doing math and it is not doing any HR.”
The key objective in a reference call is to make the referrer feel safe enough to criticize, by asking them to rank the candidate against others they know, which forces a downgrade in praise that must be taken seriously.
“Making the referrer feel safe to criticize, which is very hard to do. I wouldn't say you need to trick them, but you have to go overboard and ask something like, Would you work with this person?”
Talented people tend toward type-two happiness, meaning difficult experiences that are rewarding in hindsight but unpleasant in the moment, like a hard hike or climbing Everest, rather than type-one momentary pleasure.
“My sense is certainly successful and I think talented people are more of the type two happiness.”
Cowen looks for a trifecta of positive agency, the ability to experience and enjoy flow, and a lack of contentment, and argues that each person has physical requirements, such as squirming in a particular way, that must be satisfied for productive work.
“I look for a positive sense of agency, ability to experience and enjoy flow. And the person not being that content, that trifecta together, I think is very powerful.”
Talent evaluators should look for people who value the activity itself, such as someone who says they love to drum, rather than people who state an end goal like wanting to be the best drummer of all time, because intrinsic love of the means is more predictive of eventual mastery.
“I look for people who really value the means. So if someone says to you, I want to be the best drummer of all time. Well, that's definitely better than no ambition at all, but I actually want to hear someone saying, I love to drum.”
When assessing candidates, a recruiter should first ask who is the most energetic person to talk to, because raw physical stamina and energy give a person more attempts at manipulating their environment, whereas credential signals like school or GitHub profile are less informative.
“I just ask some of your top three people just by energy, who is the most energetic person to talk to? And I think if we take our previous definition of talent and reapply it here, someone who just has energy to try many different ways to positively manipulate the universe, to do their bidding is just going to have more shots on goal.”
Curiosity can be tested by how many different areas a person can speak about intelligently, and each candidate should be allowed to pick one area they know well, with the interviewer testing how well they think about what they know best rather than their knowledge base.
“You want to let them pick the area. You have to know something about it, but you're not testing their knowledge base. You're testing how well they can think about what they know best.”
The strongest candidates for an opportunity are not avid fans who have consumed all of one's work, but those who learned about the person through a friend and have a working mental model, since fandom signals idle curiosity and the best signal comes through the soft network.
“When I interview candidates for emergent ventures and they say, Oh, I'm your biggest fan. I've read all your this, heard all your podcasts, they are not usually the best candidates.”
Cowen judged chess players by how they physically pick up and move pieces, reading energy, attentiveness, and focus from that physical behavior, on the view that most nonphysical domains have physical parallels that reveal these traits.
“the thing I would look at is how does the person pick up the chess pieces? Do they seem energetic in their moves? Do they seem focused? Do they seem like they give a damn?”
In the world of ideas, people aged 14 to 19 are underrated because errors can simply be ignored at low cost and young people are fresh, willing to take chances, and full of energy, so those who want to shape ideas should engage that age bracket rather than established authors aged 47.
“something like 14 to 19 year olds remain underrated. And in particular, I'm thinking about the realm I work in, which is the world of ideas, where if you get something wrong, it can just be ignored.”
Universities react slowly to emergencies, as seen in their failure to quickly open a few slots for talented Ukrainian students, and schools take many years to move even 10% of their education online, which makes them poor at attracting talent that does not want daily classes.
“Schools are not set up to adapt quickly to make, say, 10% of education at a school online can take schools many years to do.”
When assessing a candidate's downtime interests, specific answers naming particular titles, such as The Matrix or Pulp Fiction, are a better signal than generalizations like action movies, because the ability to articulate specific proper nouns reveals genuine engagement.
“the first layer that's quite important is looking for specific answers instead of just generalizations.”
Experienced people provide pattern recognition and infrastructure taste, so young founders often succeed in spite of their age, making mistakes such as poor infrastructure decisions, and the best teams combine a young founder with one or two executives who have seen similar situations before.
“you tend to succeed, in spite of the age sometimes, not because of. Because you make terrible infrastructure decisions, you have no infrastructure taste.”
The relationship between age and talent depends on the sector: chess players can decline by 35 or 40, whereas novelists, philosophers, and people who hire talent tend to improve with age until some extreme point, and firms in highly regulated sectors like life insurance may want steady, low-energy people.
“if it's chess players, you can be washed up at 35 or 40, but if it's novelists or philosophers or in fact, people hiring talent, you tend to get better at that with age up until some extreme point.”
High-stakes professions such as neurology or anesthesiology should not be staffed by young people with little experience, since the cost of error is too large for enthusiasm to substitute for experience.
“You would certainly not want your neurologist or anesthesiologist to tell you, 'Hey, I just got into Y Combinator and I read a few good Wikipedia articles and I'm ready to go.'”
Mood and drive may depend on physical conditions such as sleep, illness, or gut biology, so being unhappy by default could be partly a bodily state rather than a psychological one.
“to what extent is that something physical, some type of vagus nerve, biome thing?”
Public health institutions like the CDC and NIH failed at every level during the pandemic because they run on automatic pilot and only show failures when a discrete emergency occurs.
“Some of our public health institutions failed pretty badly. The CDC, the NIH, they failed at every level, they're on a automatic pilot.”
Talent is the ability to positively manipulate whatever current environment you are in to achieve your desired outcome, achieved through intellect, charisma, or physical strength.
“I think it is the ability ultimately to positively manipulate whatever current environment you're in to achieve your desired outcome.”
Sorting people into roles should not be adversarial, because people perform best when they enjoy what they do, so a talent science can align the interests of individuals and firms.
“No one wants to be sorted or selected for a job that they do not enjoy. And people seem to be at their best when they enjoy what they do.”
A reference call is an opportunity to source new candidates by asking the referee who else should be contacted, and to recruit the referenced person directly if they seem strong.
“it's always a great opportunity to use that as a way to source candidates, too. People don't take advantage of this, but let's imagine you're looking to hire someone”
The frontier question in evaluating NBA talent is which elite athletes will be knuckleheads and which will be team players, a question that requires judgment from tape and observation that metrics alone cannot resolve.
“In the NBA, the frontier question is which of these tremendous athletes is actually going to be a knucklehead and which will be a team player?”
Reports from women suggest that Zoom and remote work make conversations more egalitarian, since they feel cut off less, though Cowen admits he is not sure enough to form his own judgment.
“I hear from a large number that it's a plus. I'm not sure I know enough to have my own judgment but I take those reports seriously, from women.”
Competency tests such as the CFA exam can serve as self-serve filters that signal a candidate's ability to work hard and is smart enough, and should be expanded into more scalable designations, though the fix must address many layers at once.
“I'm all for competency testing, but there's so many layers at which the problem has to be fixed.”
The kindest thing Cowen has experienced was his mother's unconditional support, including driving him to chess tournaments at age ten when he could not get around by himself, which was formative for his development.
“she drove me around to chess tournaments when I was 10 years old and couldn't get around by myself. That was formative for my later development”
Elon Musk, who has described himself as having Asperger's, is among the most successful entrepreneurs of his generation, which shows that unusual neurology can coexist with extraordinary engineering command and public achievement.
“Elon Musk gets on Saturday Night Live and says he's Asperger's, which is essentially the same as autism. He's the most successful, greatest entrepreneur of our generation.”
Focus on process and quest rather than happiness, since happiness may follow; people who think a lot about unhappiness often wallow and become complainers, and a good thing can be too complacent a goal.
“don't focus on the happiness, focus on the process, focus on the thing, focus on your quest. Maybe you'll end up happy.”
Among personality frameworks, the Big Five Aspect Scale is the least bad, and although the Big 16 may have higher explanatory power in statistical terms, five concepts are usable for teams while sixteen are not.
“you can't use 16 concepts with your team. So big five it is, with a ton of qualifications.”
Excellent performers in a field often appear to not try that hard, like Usain Bolt smiling through a nine-second sprint or Kipchoge breaking the marathon record while making it look easy, because people who truly excel are simply good at it.
“people that truly excel at something, they're not trying that hard, they just are good at it.”
Productive people often hold a private, secret interest in self-optimization such as quantified-self practices while publicly appearing to do little, and avoiding performative routine talk is a healthy balance.
“most productive people have very deep, very private and secret interest in the field”
Admirers who treat a founder as a hero often reveal their own unmet needs, and a platform like YC learns to be wary of them, since the best people use the founder as a stepping stone to their own goals.
“you realize the fact that you are filling some type of heroic protagonist role model in their life speaks to something that they're missing more than something they have.”
Everyone should exercise daily, both for investment reasons and because understanding one's own physicality and how it interacts with mental output is a route to greater productivity, though the right physical routine is individual.
“I think everyone should exercise every day, if only for investment reasons, but at a deeper level, to understand your own physicality and how it interacts with your mental output is a great way to become more productive.”
Talent is a creative energy that can alter outcomes and make projects better in ways that matter.
“We treat talent as having a creative energy that can alter outcomes and make projects better in ways that matter.”
Only institutions like Harvard have a real comparative advantage at recruiting the smartest people, so most other organizations should not assume they can hire the smartest workers just by pursuing smarts.
“So maybe Harvard has that comparative advantage, but most of the rest of us do not.”
Elite performers like Elon Musk, Bill Gates, and John Carmack consume Diet Coke, which signals a form of raw physical vitality and resilience that is more useful to observe in candidates than credentials, even though the drink's popularity makes it a weak metric by itself.
“I think it's interesting and abstract that you have Elon Musk, Bill Gates, Florence Summers, John Carmack, all these different titans of industry across very different spectrums, who all have at some point in their life, and some of them, even to this day, seem to be guzzling Diet Coke.”
The kind of energy that drives talent is a fixation on reaching the next destination, which can be modeled as the combination of extreme habit formation, an impulse within oneself, and a caffeine-like stimulus, a combination Diet Coke fits better than almost any substance.
“So at some mix of extreme habit formation, a kind of impulse within yourself, plus caffeine, is how to model that.”
Cowen's life philosophy of always starting a new project, captured in the baseball saying 'let's play two,' keeps his energy high, which is why he is always writing another book or starting another podcast.
“It's part of my life philosophy, let's play two.”
A hire who was perfect on paper and in interviews, with multiple advanced degrees from prestigious schools, turned out to do no work and have no ambition, and Cowen could not ex ante answer what was wrong with him.
“Once hired someone who on paper was perfect and the interview was perfect. He showed up, really just did no work, period”
Cowen describes Daniel Gross as extremely intelligent from birth and able to see problems differently, and says his physical achievements, including running marathons in Antarctica, show that he converts ideas into tangible ambition.
“So I think the physical is a big part of Daniel's excellence. He's run marathons in Antarctica.”