AI has transformed tech interviews into a game of cat and mouse.
-
AI has transformed tech interviews into a game of cat and mouse.
-
AI has transformed tech interviews into a game of cat and mouse.
@carnage4life I would like to know why, if this thing is coming for all our jobs, candidates aren’t allowed to use it for applications.
Is it bad or something?
-
@carnage4life I would like to know why, if this thing is coming for all our jobs, candidates aren’t allowed to use it for applications.
Is it bad or something?
@palendae that's a good point.
I recall once people were forbidden from looking up documentation/googling during interviews, and then it became routinely allowed because "you'll do that on your job too anyway".
I think interviews today still try to distinguish between "intrinsic skills" and "ai user skills" because we feel the former is better somehow, but this will likely change. -
@palendae that's a good point.
I recall once people were forbidden from looking up documentation/googling during interviews, and then it became routinely allowed because "you'll do that on your job too anyway".
I think interviews today still try to distinguish between "intrinsic skills" and "ai user skills" because we feel the former is better somehow, but this will likely change.@palendae mind you, I just had to deal with some ai-generated code that was crap, so I do think one needs to prove they can code to be hired as a dev, I'm just unsure what this will look like.
-
@palendae mind you, I just had to deal with some ai-generated code that was crap, so I do think one needs to prove they can code to be hired as a dev, I'm just unsure what this will look like.
@riffraff @palendae
Just spitballing but what if the interviewer (or someone else qualified) evaluates the code the applicant delivered?If it's good, the worker should keep using whatever tools they can.
If it's not, it doesn't matter if the applicant wrote the code themselves or used an AI to do it (or copied it from SO).
-
@riffraff @palendae
Just spitballing but what if the interviewer (or someone else qualified) evaluates the code the applicant delivered?If it's good, the worker should keep using whatever tools they can.
If it's not, it doesn't matter if the applicant wrote the code themselves or used an AI to do it (or copied it from SO).
@dbrand666 An issue with this approach is that live interviews have limited time so you generally have toy problems which are easy for LLMs, but in real work you may not (sometimes).
I had a trainee who did a task using LLMs. I told them it was ok, if they understood how the code worked. They couldn't, as it was buggy.
Take home tasks on real projects aim to solve this, but they select people without "a life" (a family, a job) which can't spare days to finish something.
I honestly don't know.