OpenAI RSI blog data: 124x more tokens yield maybe 10% faster AI capability gains
sebkrier · x · 2026-09-13
Ramez strings together numbers from OpenAI's recent RSI blog post to illustrate diminishing returns: a 124x increase in token output led to only a 7x increase in lines of code, which led to a 1.6x increase in experiments run. Based on power-law extrapolation, he estimates this translates to perhaps just a 10% increase in the AI capability improvement rate.
The thread argues that massive compute and output gains shrink dramatically as they pass through each layer of the R&D pipeline. While the final figure is the author's own extrapolation, the underlying data comes from OpenAI's official post, giving the scaling-diminishing-returns debate a concrete quantitative anchor.
More from AGI Musings
- Game theory of AI slowdown pledges: nobody slows down internally, everyone loses — haider1 · 2026-09-13
- The next SaaS shift: buying software that lets your agent work autonomously — eptwts · 2026-09-13
- AI solved a Millennium Prize Problem while he wrote about it: "We will never have normal times again" — voooooogel · 2026-09-13
- UK pilots three-week AI bootcamps in Preston to turn Neet youths into apprentices — nordicinst · 2026-09-13
- Video argues AI can't take everyone's jobs: excess demand means no mass layoffs — EGarrett28 · 2026-09-13
- Altman, Musk, and Hassabis back Amodei's call for independent AI oversight; OpenAI IPO delayed to 2027 — The Decoder · 2026-09-13