Matt Kane @mk.gg · 08/08/2026Did I say they were? They're both transformer models though and fundamentally are autocomplete, except this one works by predicting the next atmospheric state rather than the next token. 5830
fry69 @fry69.dev · 08/08/2026Just want to throw in that "autocomplete" is a powerful concept in itself, given enough (all?) prior knowledge. At least Laplace's demon haunted quite a few scientists for a long time.en.wikipedia.orgLaplace's demon - Wikipedia 1301
Tom Hicks @hicksyfern.bsky.social · 08/08/2026Any complaint of “it’s just predicting the next token in a stream of text” then it begs the question “given what?” Like, “given that it has a ton of knowledge and the next token prediction is just the way it emits that knowledge” then great! 2200