My company is all-in on agentic coding. We've been building a new product with it for the last 4 months. They talk about "AI governance" but I think nobody knows what that means. They have the AI writing the code and doing code reviews and everything really.
replyWell, one engineer decided to take a look at the code and found one file was 20,000 lines long. It was an awful mess with too much going on. So much for "AI governance". I knew this was going to happen but these guys don't want to hear anything about being cautious with AI. The solution is always more AI.
I am looking for a new job.
We're going to the same direction in my company, AI produces and AI reviews it which I feel is kind of like shaking your own hand.
replyFor now there's still human reviews but soon it will be all automated. I estimate that the AI review (latest Opus) is around 15% to 20% accurate
I'm still holding down keeping a human in the mix where I'm at now but it's a tough battle. Other problem with AI review is it will try to find issues that are not issues and you can get in this loop of an agent fixing something, triggering another review, fixing those findings and basically going in a circle. Has happened multiple times now which ends up with a massively overbuilt task that would have taken a human a few minutes to just do to begin with.
replyYes indeed, we have the same problem. You can run these agents 10 times and 10 times they will find different issues. It makes me trust the AI reviews even less, is the new feedback accurate if the previous one didn't catch it?
reply