pull down to refresh

15 sats \ 2 replies \ @optimism 8 Jun \ on: DeepSWE: Measuring frontier coding agents on original, long-horizon tasks AI

Has anyone gotten a 70% one-shot success rate in practice? If yes, in what language? For what task?

write
compose
84 sats \ 1 reply \ @m0wer OP 8 Jun
reply
37 sats \ 0 replies \ @optimism 8 Jun

Okay. I guess it's just my "skill problem" and these guys are my betters.

I'll put on a collar and hand them the leash.

reply