Observed arrival · 2026-09-29
Attention, After All? — A field guide to the Transformer bet
An interactive, source-linked guide to a wager about whether Transformer-like models will lead most NLP benchmarks on January 1, 2027.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Readers tracking language-model architecture debates
- Worth noticing
- Its Jamba comparison counts each reported task or aggregate as one vote, while keeping excluded models’ published scores visible.
Field notes
The page exposes its counting rules: readers can include hybrids, linear-attention models, or models with only a little conventional attention. Its table draws on five models in a 2024 Jamba-paper comparison, and the guide says each reported task or aggregate gets one vote. It also preserves a useful boundary: excluded architectures retain their published scores, and the results are not presented as statistically significant or as a final verdict on the wager.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
ƒJavaScriptBrowser-side code central
One card from the complete issue