@iWatch_AAPL@aliceisplaying@DhruvJain08 In case you're not joking, they do train on it. Perversely their TOS allows training on any conversation when it's been flagged for safety.
@can1357 "Frontier"Harness minus the frontier models
I'd rather a harness that makes frontier model use more efficient than one that handholds open models into success
@doodlestein@austeane@betterclever Aren't they increasing it? It used to be implicitly 20x plans only gave 1.7x more weekly quota compared to 5x and this was never stated. Now they're explicitly stating it gives 2x more.
many people are asking - now that opus 5.5 is so good, when would we ever use fable
it’s very important to understand that bigger models come with more “wisdom” that can’t be replaced by a smaller model trained to behave well
@nomadgeist@LByron00@angeldevenganza@maverick_zip Mens evaluation of other mens looks is weird. This guy is obviously not a 3, his eye area is at the least decent, can't tell from this angle, and that's 90% of looks. It's like people calling pete davidson ugly bc of his ears.
@angeldevenganza@nomadgeist Banning or unplugging AI won't happen. It outcompetes previous paradigms the same way capitalism, industry, agriculture, humanity did. There's no way to avoid losing our leverage to it, and with it our lives.