Ankith 🐋/acc@dhtikna
What has shazeer even contributed at deepmind?
6:04 AM · Jun 18, 2026 · 10.2K Views
Simo Ryu, Stable Diffusion LoRA creator, sparks debate over whether the original Transformer's encoder stack is inelegant
Story Overview
Simo Ryu shared the canonical encoder-decoder diagram from the 2017 Attention Is All You Need paper while crediting co-author Noam Shazeer, prompting replies that call the encoder stack inelegant and note its absence from most current large models.
81331713.1K
Industry Shift
Decoder-only designs now dominate
Replies point out that decoder-only architectures have become the default for frontier generative models because they train more simply on unlabeled text and generalize better in zero-shot settings.
Open Question
Encoder necessity stays open
No quantitative evidence or new benchmarks appear in the thread, so the precise drawbacks of the original dual-stack design remain a matter of ongoing architectural taste rather than settled fact.
Sentiment
Many users dismissed Simo Ryu's Transformer diagram as inelegant and insulting while positive replies called the result interesting and progress fast-moving.
Pos
20.0%
Neg
80.0%
6 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS4.3KBOOKMARKS1LIKES41REPLIES1
Lucas Beyer (bl16)@giffmana
Simo Ryu@cloneofsimo
@giffmana 🤣😂 honest outsiders view of three org
Lucas Beyer (bl16)@giffmana
@cloneofsimo They said at deepmind.
2hViews 776Likes 23Bookmarks 0
Ankith 🐋/acc@dhtikna
@aaronbatilo @cloneofsimo Take that back punk. I said deepmind, not google or google brain. Talking post character AI return.
2hViews 13Likes 2
1a3orn@1a3orn
@cloneofsimo pfft how inelegant, what's that bullshit on the left
3hViews 40Likes 1
Ankith 🐋/acc@dhtikna
@cloneofsimo Thanks smart ass
2hViews 41
bertø@graffioh
@cloneofsimo what's this?
3hViews 30
Ferbin@Ferbin08
@cloneofsimo robotics labs claim dexterous-hand breakthroughs constantly. none shipped. the problem isn't the algorithm, it's dust, oil, vibration, and the actual factory floor.
1hViews 16
Aaron@aaronbatilo
@cloneofsimo What a tourist
3hViews 4
Netticel@Netticel
@cloneofsimo interesting result
the direction keeps moving fast
3h