I've been noticing that many AI tools end up being just shelfware, and I'm curious if anyone else has observed this trend. For those of you responsible for making AI tools work together in your workflows, how do you handle interoperability and automation between them? What strategies or solutions have you found helpful, or what has not worked for you in your pipelines?
1 Answer
Honestly, I think many AI tools don’t really address the core problems; they often create more issues instead. In my field, we didn't have challenges that needed AI to tackle. Plus, since many of these tools operate as black boxes, it’s no wonder they just gather dust on a shelf.
Right! In DevOps, for instance, automation is crucial, but AI tends to complicate this. It’s all about having static and repeatable processes, which AI just isn’t built for. Tools like Copilot are cool for code assistance, but when it comes to actual automation, it clashes with the DevOps mindset.