Love for the Game (Hiring Signal)

The signal

"Did something at a moment in their life where they had absolutely no other reason besides love for the game."¹ That is the factor Pedro Franceschi says Brex weights most when hiring, in a world where AI already does much of the initial candidate evaluation. The examples he offers are people who coded in high school ("people that are coding in high school don't have any reason to code, you can just tell based on the kinds of things they've built and the relationship they have with the craft") and a career-changer who did not have to change, "someone that was in finance, but ended up going super deep in AI and Claude Code and built this massive thing to automate their jobs. They didn't have to do that, but they did it anyway." Franceschi cites Paul Graham's bus-ticket theory of genius as his frame, the idea of uninteresting obsessions compounding into an interesting place, and describes his own childhood building as the purest kind of creative energy precisely because there was no other reason for it.

Why he frames it as an AI-era doctrine

Franceschi's sharper argument is about what models can and cannot judge. He observes that because of post-training, recruiting models carry a relatively uniform view of a good candidate and generalize toward the well-rounded profile, adding that "it's very hard to make a model say no."¹ Against that, he argues that "what is a good candidate for us" is a choice rather than an inference, since, in his words, "focus and constraints are what give meaning to your choices": deciding where a hire must spike and what ethos the company screens for is an inherently human decision, and love-for-the-game is Brex's particular choice while a different company might legitimately prefer the PhD heading to a research lab. His claim is that when models guarantee a technical floor, technical excellence stops discriminating between candidates and the ceiling is set by ethos, so humans should spend their judgment on what remains after the models have done their part.

The interviewer sharpens the point with a tail-risk observation: a lack-of-weaknesses model would reject a young Franceschi or George Hotz for looking too unusual, structurally filtering out the spiky outliers who go on to build companies.

Connections

The pattern is the AI-era companion to hire for spikes, to which Franceschi adds the mechanism for why spikes must remain a human call, namely that the model cannot say no. It screens for the same intrinsic-care trait as hire for give-a-damn, tested through what a person did without being asked. Franceschi also links it to a calling orientation via a line he attributes to Jensen Huang, that it is easier to learn to fall in love with what you do than to do something you already love, suggesting mastery can generate the love rather than only the reverse.

Set against the other two hard-to-fake signals in the archive, what this one isolates is the absence of a reason. The file on depth as hiring signal tests how far a person went and whether they can turn it over, a question this filter never asks, since the high-school coder's project may be shallow and still count. The file on chip on the shoulder selects for the very thing Franceschi's filter screens out, a person answering something external, which is why the two can identify the same candidate for opposite reasons and disagree completely about what the company is buying.