Damn those are a lot of reasons. The accountability one is particularly messed up to me....If you cant defend your own work and blame it on Claude, The question arises why are you even being hired for
"Helpful to the writer, useless to the reader" is such a good way to put it. The cost just moved downstream. I've started skipping the doc entirely and just reading the code, feels faster at this point
Agreed, the trust thing is what gets me too. when a review or an issue might be ai generated noise you end up re-reviewing everything by hand, so it adds work instead of saving it. The volume of meaningless issues is lowkey insane...
Nice work. One thing I didn't see covered: what happens when a #[hsrs::function] panics? Unwinding across the FFI boundary into the Haskell RTS is UB, so I assume there's a catch_unwind somewhere converting it into a Haskell exception, or do panics just abort the process?
well i wouldn't call him that but he has played his cards right everytime. He's more of a business focused guy. Btw, i had the legit same thoughts 2 days ago but yeah he's just a great leader.
yeah i feel it too..it's thinking has been way too shallow recently..i was on xhigh doing deep reasoning & it couldn't remember the context from previous text. Didn't push back at me at all, kept changing it's opinions to whatever i said. Also the usage dies so fast.