▲ 1026 ▼ how things become science (thelemmy.club) submitted 5 months ago by not_IO@lemmy.blahaj.zone to c/science_memes@mander.xyz 121 comments fedilink hide all child comments https://www.nature.com/articles/d41586-026-01100-y https://bsky.brid.gy/r/https://bsky.app/profile/did:plc:s6yp6jam5og3tftozaw7pjth/post/3mj34sn6kyk25
[–] MalReynolds@slrpnk.net 1 point 5 months ago (1 child) Valid use case, still a significant problem. permalink fedilink source parent hideshow 1 child comment replies: [–] GaMEChld@lemmy.world 2 points 5 months ago But not really a NEW problem. We knew LLM's are trained on aggregate human data. We know aggregate human data is fundamentally flawed, inconsistent, unreliable, etc. Like was there a point at which people just decided, nah AI is just plain accurate? Or is that just what morons always thought despite the permanent warnings plastered everywhere saying THIS AI CAN MAKE MISTAKES, CHECK EVERYTHING! permalink fedilink source parent
[–] GaMEChld@lemmy.world 2 points 5 months ago But not really a NEW problem. We knew LLM's are trained on aggregate human data. We know aggregate human data is fundamentally flawed, inconsistent, unreliable, etc. Like was there a point at which people just decided, nah AI is just plain accurate? Or is that just what morons always thought despite the permanent warnings plastered everywhere saying THIS AI CAN MAKE MISTAKES, CHECK EVERYTHING! permalink fedilink source parent
