return2ozma@lemmy.world to Technology@lemmy.worldEnglish · 3 days ago‘The People Building AI Earnestly Believe That It Could Kill Us All’: Anthropic Researcher Quits Dramaticallygizmodo.comexternal-linkmessage-square190linkfedilinkarrow-up1638arrow-down123
arrow-up1615arrow-down1external-link‘The People Building AI Earnestly Believe That It Could Kill Us All’: Anthropic Researcher Quits Dramaticallygizmodo.comreturn2ozma@lemmy.world to Technology@lemmy.worldEnglish · 3 days agomessage-square190linkfedilink
minus-squareabc@suppo.filinkfedilinkEnglisharrow-up1·2 days agoWhat do you seriously think the result of that would be? I know it’s a hard ask for a lemmy.ml account to be serious, but are you capable of that?
minus-squarebalsoft@lemmy.mllinkfedilinkEnglisharrow-up2·2 days agoIf you just pipe an LLM straight into a feedback loop with a console, it is going to destroy things rather soon, depending on the context window and other factors. All the agentic fluff is basically the workaround for LLMs going off the rails.
What do you seriously think the result of that would be? I know it’s a hard ask for a lemmy.ml account to be serious, but are you capable of that?
If you just pipe an LLM straight into a feedback loop with a console, it is going to destroy things rather soon, depending on the context window and other factors. All the agentic fluff is basically the workaround for LLMs going off the rails.