Interesting #AI hypothesis:- It is known, interpolation is easier than extrapolation.- In realms learned from humans, #LLM outputs can be considered interpolation, thus reasonable + helpful.- Yet outside these realms (new questions, research) LLM output is extrapolation, might make no sense.RE: https://bsky.app/profile/did:plc:s27fo7dzltw3acjxxld3rjmb/post/3mr5eja6xjs23
Related
📰 Game Pass adds Halo: Campaign Evolved, Beast of Reincarnation, and more in late July to early AugustMicrosoft has anno...
📰 Game Pass adds Halo: Campaign Evolved, Beast of Reincarnation, and more in late July to early AugustMicrosoft has announced new titles coming to Game Pass in late July to early A...
Zamiast obiecywanej wolności, narzędzia AI wprowadzają nas w pętlę niekończących się poprawek. Mechanizm „automatu do gi...
Zamiast obiecywanej wolności, narzędzia AI wprowadzają nas w pętlę niekończących się poprawek. Mechanizm „automatu do gier” skutecznie rozbija skupienie, zamieniając pracę w serię ...
These camera-free smart glasses want to give you a secret advantage in meetingsThe new Halliday G2 brings live translati...
These camera-free smart glasses want to give you a secret advantage in meetingsThe new Halliday G2 brings live translation, instant research, automated summaries, ad more to your e...