@jon many orgs have senior+ people managing a team of juniors and this is all they do anyway.

Except AI makes the feedback loop nearly instant, so I think this would be very attractive to certain types of roles.

I think those type of roles are a certain sort of smell to begin with though

@jon cool, so more and more code being generated by AI, with humans reviewing only, resulting in diminishing knowledge of the code base so more issues slip through. New models are trained on the AI generated code and learn the bugs which they then contribute more of in future code and the humans know less and don’t pick them up. Then the cycle continues, until it’s all just shitty code that nobody can fix.

Thanks Google!

Sincerely

Nobody

@jon And the kicker? The more you review, the less you'll get to do in the future. If you're lucky, you'll get a bonus in your last paycheck once the company realizes that the LLM's error rate is within an acceptable, profitable range and you're no longer needed.

@jon The redeeming factor of code review is that at least you get to share knowledge and experience with other humans.