Open-weights world models' parameter size doubles every ~6 months
kaarelson
14 points
8 comments
September 11, 2026
Related Discussions
Found 5 related stories in 124.5ms across 10,677 title embeddings via pgvector HNSW
- Our position on open-weights models surprisetalk · 701 pts · July 27, 2026 · 57% similar
- State of Open Models: Summer 2026 Observations gmays · 36 pts · September 01, 2026 · 54% similar
- Concerning behavior from a popular open-weight AI model matthieu_bl · 283 pts · October 26, 2026 · 51% similar
- Meta's new open-weight model targets local agentic AI bakigul · 37 pts · August 10, 2026 · 51% similar
- 6 months to live for open models g42gregory · 31 pts · July 12, 2026 · 50% similar
Discussion Highlights (3 comments)
unreal6
https://xcancel.com/kaarelkaarelson/status/20984729928211993...
kadoban
Is it me or is that not what that graph shows? That line has ~no relation to the datapoints it's purporting to be a fit for.
fithisux
I would be more interested the cost of each model in hardware resources to achieve 1 token per second. Cost is a metric of interest these days.