Once AIs roughly match top human experts in AI R&D, could kick off a feedback loop where AIs doing AI research produce smarter AIs, yielding four or five years of AI progress in a single year.
Ryan argues AI R&D is unusually tractable for AI because it's verifiable and hill-climbable, so once AIs match top human experts it could trigger a feedback loop giving ~4-5 years of AI progress in one year. ✦ AI generated
Ryan Greenblatt · Dwarkesh Podcast · 2026-08-11 · original ↗
plays this moment only · 0:00 — 16:52
“Whether or not this turns out to be the case is probably the most important question in the world right now. And historically, I've been quite skeptical that this kind of thing happens, but, you seem to think that it might be plausible, and so I wanted to hear the case for it.”
I think once you have AIs which are roughly matching the top human experts in AI R&D, that could kick off a feedback loop where the AIs are doing AI research. That produces smarter AIs. That feeds back in. That feedback loop could be strong enough that you end up with a lot of progress in a short period of time. Maybe my median expectation is something like four or five years of AI progress in a single year. This requires really overcoming a huge amount of diminishing returns in research and basically doing the equivalent of the progress we would have gotten after a really large compute scale-out. So this is a pretty impressive, big thing.
verbatim transcript · starts at 0:00
0:00– Is AI R&D verifiable enough to unlock recursive self-improvement?
16:52– Is AI progress bottlenecked by human expert data?
0:00– Is AI R&D verifiable enough to unlock recursive self-improvement?
16:52– Is AI progress bottlenecked by human expert data?
- ·Once matching top human AI researchers, AIs could research and improve themselves.
- ·Smarter AIs feed back into research, potentially accelerating progress dramatically.
- ·AI R&D is unusually tractable because it is verifiable and hill-climbable.
- ·Ryan’s median expectation: four or five years of progress in one year.
- ·Achieving this requires overcoming substantial diminishing returns in research.
- ·It would resemble progress from a very large compute scale-out.