cm0002@piefed.world to Programmer Humor@programming.devEnglish · 2 months 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-square32linkfedilinkarrow-up1199arrow-down112
arrow-up1187arrow-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 months agomessage-square32linkfedilink
minus-squareCanadaPlus@lemmy.sdf.orglinkfedilinkarrow-up5·edit-22 months 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.