Last week, entrepreneur Brett Adcock pulled back the curtain on the AI talent war, and the numbers are absurd. He watched Meta snatch up a senior AI infrastructure engineer with a $36 million RSU offer spread over four years. Not a typo. Thirty-six million dollars, guaranteed cash, all to outbid Adcock's own competitive offer.
This isn't just about senior hires. Adcock notes that junior engineers are also commanding multi-million-dollar annual packages. Mark Zuckerberg, he says, is "smart" for buying his way into the AI race with these tactics. But Adcock sees a deeper problem. He estimates only 20 to 30 people in California truly know how to build good AI models. "Most of my time is trying to find those folks," Adcock says. “I found that even in the Bay Area where it's probably like the richest AI and engineering like folks in the world, 90% of everybody out here is not good at their jobs.”
His company, Figure, has been running “10 case studies a week for six months” and still hasn't hired anyone. This scarcity means you can't just throw money at the problem if you want the best. Adcock prefers to find "obsessed" individuals deeply committed to the mission, not "mercenaries" driven purely by money. He won't play Meta's game. To identify these rare, truly skilled engineers, Adcock uses a specific method.
The Brett Adcock's Technical Assessment for Hiring Top AI Talent
Principle 1: Identify "Done the Work" vs. "Watched the Work": I need to know like if a if you did the work or if you like watch somebody do the work.
Characteristic 1: Deep Detail Recall (Like a Scar): If you've done the work, it's like it's like it's like a scar you carry with you. It's like it's like dug into you. Like you know all the details. You can talk about it freely. You don't need to think.
Characteristic 2: Ability to Reverse Engineer: You'll understand how to like reverse engineer everything you've done and discuss it.
Characteristic 3: Depth of Knowledge (Avoid "Blow Up" at First Layer): The folks that haven't done it can't do that. They just like they can't even go like they get one layer and they just like instantly blow up. They can't talk about it. They don't know why.
When This Works (and When It Doesn't)
This method is effective for identifying genuinely experienced and capable individuals in highly technical fields like AI and engineering, distinguishing them from those who may present well but lack deep practical knowledge. It focuses on probing for granular understanding and real-world problem-solving experience. Adcock's approach works when the talent pool is small and expertise is critical, but it requires interviewers who themselves possess deep technical knowledge to properly evaluate candidates. It might break down for more generalist roles or in fields where innovation sometimes comes from combining disparate ideas rather than just perfecting a specific craft.
What to Do With This
Next time you interview for a highly specialized role—especially in AI or deep tech—pull out a candidate's resume and pick one of their biggest past projects. Ask them to describe a specific challenge they faced, then use Adcock's framework. Demand the "scar" details: "What was the exact bug? What line of code? How many hours did you spend debugging?" Then push them to reverse engineer it: "If you had to rebuild this system today, what's the absolute first decision you'd make, and why?" If they can't peel back at least three layers of detail without getting vague, they probably just "watched the work."