Many users reacted skeptically to speculation that Anthropic used looped transformers in Sonnet 3.5, warning of psyops and criticizing the lack of open sharing.
Based on 7 visible X reactions from 28 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@scaling01 Btw, Sonnet 3.5 was the first usable model for coding. While OpenAI released o1-preview, it was slow and extremely limited. Maybe we'll never see such progress again, baseline is too high and suturated.
@kalomaze GraphWalks truthers in shambles
beware the psyop
"Anthropic discovered a magic thinking feature that they dialed up for Sonnet 3.5" ahhhh, good times :)
"mythos is a looped transformer" is the new "o1 uses test time MCTS"
@kalomaze 1 loop is still a loop. We want to believe.
Many users reacted skeptically to speculation that Anthropic used looped transformers in Sonnet 3.5, warning of psyops and criticizing the lack of open sharing.
Based on 7 visible X reactions from 28 accounts; directional sample.
Ask a question below.
Published answers will appear here.