Every team I talk to asks the same thing about candidates: did they use AI?

By 2026 the answer is always yes. AI wrote part of the resume, drafted the take-home, and sat open in a tab during the interview. Trying to detect it is a losing game, and catching it tells you nothing about whether the person can build.

Better question: who was in charge?

Two engineers, same model. One types a vague prompt, pastes the first answer, ships it. The other reads what comes back, spots the confident hallucination, throws out the eighty percent that's plausible nonsense, keeps the twenty that's right, and can say why. One directs an intern; the other is the intern.

You can't see that gap on a resume, a take-home, or an AI-detector. And it's most of the job now. Engineering used to mean typing the right characters from memory. Now it means knowing what to ask, what to trust, and what to throw away at speed.

That's what Crackd measures. Candidates can use AI on anything. We log every move and grade the judgment around it: how they verify, how they sort good output from bad, when they hold back, what each correct answer cost them.

The teams hiring well right now are picking the engineers who direct AI best. AI is their intern. Make sure it's not the other way around.