If leading labs were doing research then we would have already transitioned to post transformers.
The biggest thing that has happened so far in my opinion so far is o1.
8:27 PM · Jun 24, 2026 · 41.7K Views
Former Google DeepMind distinguished engineer Rohan Anil argues AI labs neglect fundamental research, pointing to their continued reliance on transformers
Anil called OpenAI's o1 the most significant AI development.
25566128844.3K
Sentiment
Sentiment building, check back later.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS4.7KBOOKMARKS4LIKES62REPLIES4
rohan anil@_arohan_
Ok I kept bar higher, there has been continued improvements on sparsity, optimizers, and training scaling.
rohan anil@_arohan_
If leading labs were doing research then we would have already transitioned to post transformers.
The biggest thing that has happened so far in my opinion so far is o1.
1dViews 4.7KLikes 62Bookmarks 4
Ahmad Beirami@abeirami
@_arohan_ And inference
rohan anil@_arohan_
Ok I kept bar higher, there has been continued improvements on sparsity, optimizers, and training scaling.
1dViews 298Likes 2Bookmarks 0