GitHub just released their new AI coding assistant for public testing. This bot automatically creates pull requests in major repositories, including Microsoft’s main .NET codebase.
I’ve been watching some of these automated contributions and honestly, the quality is pretty concerning. The AI seems to miss important context and creates changes that don’t make much sense. You can see several examples in the recent PR history where human developers have to spend time reviewing and explaining why the suggested changes won’t work.
What really bothers me is thinking about the developers who now have to review all these automated submissions. They probably didn’t choose this workflow but management decided to implement it anyway. It must be frustrating to deal with low-quality contributions that waste review time.
Has anyone else noticed this trend with AI-generated code contributions? I’m starting to wonder if we’re moving in the right direction with all this automation in software development.
I’ve dealt with this exact thing at work. Management rolls out AI tools without asking how they’ll actually affect us developers. The AI isn’t the real problem - it’s how they implement it. These systems spam PRs without any filtering or understanding of context, so you can’t find the stuff that actually matters. What bugs me most? Management thinks all this automated activity means we’re being productive. Meanwhile, we’re spending more time fixing the AI’s mess than we’d spend just writing the code ourselves. These tools should make our jobs easier, not pile on more work. The tech could be good, but it needs way better guardrails and has to actually work with how we already do things. Right now it’s like being forced to beta test broken software in production.
this is classic corporate ai hype without any real thought behind it. sure, the metrics probably look amazing - “look at all these prs we’re generating!” but nobody’s tracking the developer hours wasted reviewing garbage. we dealt with the same thing when automated testing tools just created noise. the tech might improve eventually, but right now it’s just creating more work for everyone.