If you’re in a world where they can operate effectively over three months, but the model release cycle is every two months, then you don’t have a way to evaluate the models at the full length of their capabilities before the next model rele
Noam Brown, OpenAI · made Sep 17, 2026 · target Dec 2026 · Open
Unverified Only a secondary source was found; the original words were not read on a page.
If you’re in a world where they can operate effectively over three months, but the model release cycle is every two months, then you don’t have a way to evaluate the models at the full length of their capabilities before the next model release cycle.
The prediction
- Who
- Noam Brown, Researcher, OpenAI
- Made
- Sep 17, 2026 11 days ago
- Target
- Dec 17, 2026 in 2 months (stated as "over three months")
- Status
- Openin 3 months
- Source
- Dwarkesh Podcast: Agent swarms, alignment, and recursive self-improvement Primary: their own words · original recording
- Topics
- Agents
More from Noam Brown
| Prediction | ||||
|---|---|---|---|---|
| Sep 17, 202611 days ago | By the end of 2030 each leading lab will have compute to run hundreds of millions of human-level AIs.Open · Unverified · target Dec 2030made 11 days ago | Dec 2030in 4 yearsby the end of 2030 | Compute | Open |
| Sep 17, 202611 days ago | So I was like, “I don’t think we’re going to get it in 2026, probably not in 2027, maybe in 2028.Open · Unverified · target Dec 2026made 11 days ago | Dec 2026in 3 monthsin 2026 | Science | Open |