I use a meeting-minder that is LLM-backed. We use it specifically to avoid pro/con bedlam in meetings with opinionated and excellent engineers and devs who think they have the right answer, but we need to settle and support one.
The LLM-led pro/con discussion has been extremely helpful to curtail individuals in meetings from taking over a discussion when pros and cons come up, which has helped reign in some valued and valuable but ... sometimes unhelpful teammates.
I treat it like a new intern who just started that day.
When I see a bug, I tell it and it often fixes it.
So I don't know if I would be super impressed with the generated code being fixed; I'd be curious why someone would expect any code to be perfect and not improvable.
I use a meeting-minder that is LLM-backed. We use it specifically to avoid pro/con bedlam in meetings with opinionated and excellent engineers and devs who think they have the right answer, but we need to settle and support one.
The LLM-led pro/con discussion has been extremely helpful to curtail individuals in meetings from taking over a discussion when pros and cons come up, which has helped reign in some valued and valuable but ... sometimes unhelpful teammates.
I settle debates by finding all the bugs in the generated code.
I treat it like a new intern who just started that day.
When I see a bug, I tell it and it often fixes it.
So I don't know if I would be super impressed with the generated code being fixed; I'd be curious why someone would expect any code to be perfect and not improvable.