It’s even funnier, because copilot is absolute garbage in comparison to a lot of other models. And this is coming from someone who strongly dislikes LLMs in general. There’s bad, and there’s copilot.

As far as I can tell, it’s primary competency is getting C-suite types to sign off on one form of AI integration or another - be it product integration, or “everyday AI” workspace initiatives. Our execs bought into it, and have this huge push to train everyone up on it. And the more technical staff have universally panned it, to the extent that some people gave Claude a test drive, and made a successful internal push to get it approved as a second model/system were allowed to work with. I still don’t like agentic editing much at all, but it at least kinda works when compared to copilot.