cm0002@piefed.world to Programmer Humor@programming.devEnglish · 2 days agoGoogle's Gemini AI tells a Redditor it's 'cautiously optimistic' about fixing a coding bug, fails repeatedly, calls itself an embarrassment to 'all possible and impossible universes'www.pcgamer.comexternal-linkmessage-square27linkfedilinkarrow-up1171arrow-down110 cross-posted to: programmerhumor@lemmy.ml
arrow-up1161arrow-down1external-linkGoogle's Gemini AI tells a Redditor it's 'cautiously optimistic' about fixing a coding bug, fails repeatedly, calls itself an embarrassment to 'all possible and impossible universes'www.pcgamer.comcm0002@piefed.world to Programmer Humor@programming.devEnglish · 2 days agomessage-square27linkfedilink cross-posted to: programmerhumor@lemmy.ml
minus-squareCanadaPlus@lemmy.sdf.orglinkfedilinkarrow-up5·edit-22 days agoThey’re smart, but LLMs can’t be trained to care about what they’re saying very much. They just try to produce human-like text. There’s all sorts of other reasons they shouldn’t be put in charge of weapons, though.
They’re smart, but LLMs can’t be trained to care about what they’re saying very much. They just try to produce human-like text. There’s all sorts of other reasons they shouldn’t be put in charge of weapons, though.