Depth as Hiring Signal

The search pattern

Bank of America CEO Brian Moynihan, asked whether he has a standard interview question, describes a search pattern rather than a single question: "I'm always looking for something the person's gotten into a depth that they can talk about, that I can see the depth of their thinking, because it may not be that topic they're going to do here. But if they've had the courage to go deep into something and learn about it and turn it over and think of ways it applies, that shows you that they're going to keep digging and digging and digging, and have an eye to detail and information."¹

Three things in that answer are load-bearing. The topic is explicitly irrelevant: "it may not be that topic they're going to do here." He is not testing domain fit; he is testing whether the behavior of going deep is present in a person's history at all, on the assumption that it transfers. The second requirement is turning it over and thinking of ways it applies: depth alone is not the signal, since a memorized specialty is not it either. The signal is depth plus the reflex to generalize from it, which is what separates someone who learned a subject from someone who learned how to learn a subject. And the word he chooses, courage, is unusual for what looks like a curiosity trait: going deep means spending time confused and visibly not producing, and passing up the shallower coverage that reads better socially in the moment.

The signal is hard to fake for a specific reason: shallow knowledge collapses under two or three follow-up questions in a way that is immediately audible, and no amount of interview preparation manufactures a history of having actually gone deep into something. Moynihan names the same quality as his own criterion for staying in his job: "if you don't try to be curious and drive it, if you're blocking, that's when you got to go," and, on leadership generally, "if you aren't curious about how things work, that's the key."¹

Musk's version: the four ways you screwed it up

Elon Musk's interview filter is a proximity test rather than a knowledge test: "If you can't tell me the four ways you fucked something up, then you weren't the one doing the real work." It works for the same underlying reason: a failure inventory is very hard to fabricate, because only the person who actually did the work knows where it broke. Musk pairs it with detailed technical questions about a specific hard problem the candidate has solved, though the questions are portable in a way the interviewer's own judgment is not, since scrutinizing the answers requires being nearly as close to the work as Musk is himself.²

Two related practices round out his approach: he biases hard toward young, unproven engineers given a striking amount of accountability, on the theory that no bad habits beats good habits and that a fast iteration rate reveals who is right quickly; and he has been known to compress hiring to the point of interviewing a machinist, negotiating salary in the room, and signing a filled-in offer on the spot, at 6pm on a Saturday.²

Why it matters

It is the natural companion to dropping credential filters: once a degree is not the gate, something else has to carry the signal, and demonstrated depth is credential-independent by construction, since it can come from a hobby, a trade, a game, a side project, or a job. It is also cheap and portable in a way structured technical assessment is not, since it requires no test design and can run inside an ordinary conversation, which matters for an executive who interviews rarely or is hiring outside personal expertise.

Depth belongs to a small family of hiring signals that work by being expensive to fake, and what distinguishes it from the others is that it reads a behavior and stays silent about the motive behind it. It registers that someone dug and generalized, without asking what made them dig, which is exactly the question the other two answer. The file on love for the game looks for work done at a moment when nothing external rewarded it, evidence of the motive depth cannot see, and the file on chip on the shoulder reads the opposite motive, a grievance supplying the reason to push. A candidate can present real depth acquired under pure obligation and pass this test while failing both of theirs, which is what makes the three complements rather than substitutes.

The signal has real limits. It rewards people who had the slack to go deep in the first place, since depth in a subject usually requires unstructured time, so someone who worked through school may present less of it while carrying more of the underlying trait, which makes this a less egalitarian signal than it first appears. The interviewer's own depth also bounds the test, since a good second and third follow-up question requires knowing enough about the topic to ask one; a narrow interviewer can mistake fluency for depth. And the practice is self-reported and low volume by Moynihan's own admission, since an executive in his position does not interview many people, with no claim offered anywhere that depth-screened hires actually outperform, only that the signal is what he personally looks for.¹