@algekalipso My understanding is they were locally more efficient than RNNs and introduced the scaling thesis. If they didn't exist we would just be on the same trajectory but slower once scaling is…

Posters: jd_pressman. jd_pressman

3 likes recorded by the source corpus.

Topics: AI & LLMs

View Post: https://x.com/jd_pressman/status/1810234087724212360