# Nathan Lambert says benchmarks are a bottleneck for intelligence explosion

> Original title: imo a lack of good enough benchmarks is a bottleneck. At least for the type of RSI describe as an intelligence explosion. You can use exi...

- Source: Nathan Lambert
- Published: 2026-11-05T19:29:24.000Z
- Bad Influence: https://badinfluence.news/items/riwqr836y418fas8z16if4c31
- Original: https://x.com/natolambert/status/2118425465988014534

## Summary

Nathan Lambert says a lack of good enough benchmarks is a bottleneck for the kind of recursive self-improvement described as an intelligence explosion. He adds that existing benchmarks can still be used to make models much cheaper.

## Post

imo a lack of good enough benchmarks is a bottleneck. At least for the type of RSI describe as an intelligence explosion. You can use existing benchmarks to make models way cheaper, though.

![](https://badinfluence.news/api/img-proxy?u=https%3A%2F%2Fpbs.twimg.com%2Fmedia%2FHWYolLOXMAALqQe.jpg&mode=thumb&exp=1794096000&sig=203c8d2e47de656a)

## Quoted post @METR_Evals

> Since August, a team within METR has worked with internal Anthropic information to assess whether AI itself is accelerating AI development.
> 
> We are now publishing a report summarizing the results of this risk assessment.

https://x.com/METR_Evals/status/2118412183860351084