Schmidhuber strikes again: Transformers are just 1991 fast weight programmers

zainhas · x · 2026-11-03

Jürgen Schmidhuber replied on X that the new work is 'nothing new'—it's exactly what the neural fast weight programmers did in the 1990s, including the 1991 unnormalised linear Transformer, linking to an overview of the original papers. Observers quip about getting 'Schmidhuber'd,' the researcher's signature priority claims.

Related event: Schmidhuber Reiterates 1991 Fast Weight Programmers Were the First Transformers(5 posts)→

Original post →

More from Fun

Fun channel →