Still trying AI.
Last week: Oh. Claude Code seems really clever. It’s guided me to best practice for a thing I don’t know much about.
This week: Me consulting the docs and finding out this is not best practice at all, and Claude apologising and tearing its own recommendations to shreds.
It’s still just an over-confident bullshit machine isn’t it?
This happens every time.
You folks having success… are you lying? Or are you not validating the results? I don’t get it.