CriticCode Blog
Writing on interviewing
in the age of AI.
Ideas, opinions, and lessons on how engineering interviews should work when every candidate is already using AI. Published irregularly, written slowly, no SEO listicles.
-
· 4 min read
Live coding rewards composure, not engineering
Live-coding interviews test whether a candidate can stay calm under surveillance while regurgitating a pattern they happened to remember. That's not a signal about engineering. Here's what we think is.
interviewing hiring engineering-management -
· 4 min read
How to ask a question a candidate can't rehearse
Every question in your loop sits on a spectrum between 'the candidate has answered this thirty times' and 'the candidate has never thought about this before'. The second end is where the signal lives.
interviewing hiring engineering-management -
· 5 min read
What a good challenge statement looks like
The challenge statement is the hardest part of CriticCode to get right. Here's the shape of one that works, with worked examples for three common roles.
hiring interviewing challenge-design -
· 4 min read
How to read a candidate's AI transcript
The final written answer is the tidied version of the thinking. The AI chat transcript is where the actual thinking shows up. Here's how to scan one in under ten minutes.
ai interviewing hiring -
· 4 min read
Take-home interviews that respect everyone's time
If your take-home is longer than two hours of real work, you're filtering for candidates with free weekends, not candidates with judgement. Here's what a good one looks like.
take-home hiring interviewing -
· 4 min read
The paste event isn't cheating
Every time a candidate pastes a chunk of text into a take-home, some hiring manager reads it as a confession. It isn't. Here's how to read a paste event like a grown-up.
hiring ai interviewing -
· 4 min read
Candidates already use AI. Most interviews haven't caught up.
The 'no AI' clause on a take-home started as a quality filter. In 2026, it mostly ends up measuring whether the candidate is willing to pretend. Here's what we think it's worth measuring instead.
hiring ai interviewing