I had done a few remote coding interviews in recent months where I suspect the candidate was cheating using AI. It's a bizarre experience: each individual answer is produced confidently and quickly, makes sense in isolation, occasionally is even optimal, but the different answers don't connect into a coherent whole. Contrived example: the candidate confidently states that one should use algorithm X to solve a particular type of problem because of such and such reason - and then five minutes later when it comes time to write some code, they rapidly type in, with no erasing or backtracking, a solution which uses algorithm Y, and seemingly no awareness that they switched from X to Y...