Agustin Lebron - Trading, Crypto, and Adverse Selection

The conversation opens with a teaser montage of themes that recur later: a long view of life, software development as sociology, and whether finance's share of GDP is too high a price for liquidity and price discovery. Dwarkesh then introduces Agustin Lebron, author of The Laws of Trading, as one of the books that Tyler Cowen calls "quake books" because they shift the models a reader holds of the world.

Dwarkesh's first question frames the whole book. Peter Thiel has said the Straussian reading of Zero to One is that one should not start a startup; Dwarkesh suggests the parallel reading of Lebron's book is that one should not trade. The reasoning is that most people have no edge because they are no better than a marginal trader, and anyone who believes they do has probably not counted risks and other costs. Lebron says this "pretty much" hits the nail on the head. As people think harder about trading, they see how difficult it is to do well, and for anyone smart and hardworking enough to make a living in financial markets, there is usually an easier way to earn money and have a satisfying life. (The Thiel and Straussian names are as transcribed; the audio rendering is garbled.)

Lebron describes a path that connects three fields. He trained as an engineer and spent about six years as a chip designer while playing a lot of online poker, which he says was then profitable and arguably legal. When engineering grew boring, he asked what lay halfway between engineering and poker and settled on quantitative trading. In January 2008 he told his boss he was leaving for finance, and the boss asked whether it was a good time. A few months later he was hired at Jane Street, where, in his joking phrase, he watched the implosion of western civilization from a trading desk.

After several years he left Jane Street and started a consulting company that helped tech companies with growth, management and hiring. In the last few months he has started a new company in crypto. He gives only the outline: a crypto protocol with cryptographic guarantees against things traders dislike about crypto trading.

Dwarkesh asks whether the winner's curse implies that the average worker is overpaid, if hiring is treated as bidding among employers and the true value is the average bid rather than the highest. Lebron agrees that from the employer's side adverse selection is everywhere, and he gives two separate mechanisms. First, people who are very good are likely known to be good by their current employers, who have strong incentives to keep them, so applicants on the market are at the margin weaker. Second, the employee has the final say in whether a hire happens. Strong candidates receive many offers and choose among them, while weaker candidates have few, so employers are selected against systematically.

Lebron separates this from the claim that employees are systematically overpaid. Companies should have a decent idea of the marginal value of an extra employee. People also give things up for the security of working at a company, which might offset the adverse selection in pay. He says it is not clear which way the net effect falls.

Dwarkesh relays a point from a blog post by Byrne Hobart (transcribed as "Bern Hobart"), made almost in passing in a footnote: adverse selection should be stronger in industries like finance, where money is the main motivation. A person who wants to work at SpaceX has a story about why they chose you, but many finance firms bid for the same talented people, so a candidate who ends up with you invites suspicion. Lebron agrees there is something to this. From his own hiring experience, one of the biggest near red flags is a candidate who says they have wanted to be a trader their whole life. They cannot really know what trading is, because it is not tangible the way being a doctor is. What they are actually saying is that they have wanted to make a lot of money their whole life. In some jobs that is a good motivation, but he does not want it to be the sole one at the outset.

Dwarkesh points out that the book's chapter on motivation seemed to imply that wanting money is the motivation to seek, because people motivated by emotion will lose to people motivated by money. Lebron says the desire to win the game, with money as the scoreboard, makes sense for a trader. But a pure wish to make the most money possible correlates with less attractive traits. Much of the job needs inherent curiosity about random things, and a person who asks only where they can make the most money today is not necessarily optimal over the long run. The motivation needs balancing with enjoying the game for its own sake and as exploration. He admits this may be slightly inconsistent with what he wrote in the book, but argues that at the margin people need to hear this other side more.

Asked how to tell whether someone enjoys the game for its own sake, and about the idea that a firm like Jane Street holds retail trading experience against a candidate, Lebron offers a heuristic he has stated before. He would love to talk to the person who is the third best in the world at some obscure chess variant. That person likely has the willingness to grind at something and get very good at it, not for a pot of gold at the end of the rainbow but because they find the effort inherently enjoyable. Beyond the mathematical and risk-taking parts, which he treats as separate, he also looks for a strong desire to be in a competitive environment and to enjoy it, which can take many forms.

the third best player in the world at some weird obscure chess variant