@f4grx @freecad This is disappointing. I was just learning #freecad for 3D print models coming from #onshape.

LLMs are pushed heavily at my work. They seem productive but all it does is push the bottleneck to reviewers, who now have the burden of making sense of slop. You're in a situation where you're tempted and pushed to ship faster: read less, skim more, and just approve. The LLMs are great at adding lines of code, but every line is liability, they miss obvious bigger picture abstractions.

Things _will_ slip by, you're facing cognitive bias no one can completely overcome. Please reconsider.