The interesting part is that the Rust compiler ends up acting as the reviewer the author can't be, since the borrow checker rejects a lot of what a model gets wrong. The risk is code that compiles while the logic is still off and nobody on the team reads Rust well enough to notice.

Show Post

I think it's kind of telling that everyone keeps thinking the problem with their harness is the language, so ofc they want to rewrite it in Rust, a language most of them do not know with any real proficiency.

Show Post

sharing an internal take externally. we love takes.

Show Post

hardest part i think is trying to slow roll in new ways of working, and i think just jumping in fast would render better results

Show Post

lol i made an edit to it after posting here that said “i want to note that i don’t mean all of literally and that we should be on the furthest end of the curve here, but rather this is a purposefully asinine take so we can find a place to meet in the middle that works”

Show Post