korrents

What Ryan Greenblatt thinks about reinforcement learning

@ryan-greenblatt · 26 positions · 0 changes of mind

Chief scientist at Redwood Research, where he works on technical AI safety and AI control.

Ryan Greenblatt did not write this page.

We collected these quotes from things they published elsewhere, and every quote links to where it was said. They have no account here and have not endorsed this site. Quotes are word for word; the short line under each one is our own restatement, not their wording. Their own site. Is this you? Claim it or ask us to remove it. Or tell us what is wrong here.

7 dated positions, 2026, in their own words. Our reading of what Ryan Greenblatt has said — not written or endorsed by them.

7 positions so far — this page is not yet offered to search engines.

  1. Our readingToday's RL environments are better than 2024's because we learned what to build and put AI labour on building it, not because labs hired more human experts.

    Their wordsthe reason why RL environments today are much better than they were in like you know 2024 is not that much because um we have hired way more human experts to make RL environments and is instead much more because we better know what how RL like what RL environments we even want to make and and like how we should structure them and also we're using huge amounts of AI labor to build RL environments.

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 6th of 26 in this recording

    Interesting

  3. Our readingAI that is merely excellent at research and hardware can transform the world completely without ever being good at politics.

    Their wordsLike I think my perspective is like if the AIs are sufficiently good at R&D including hardware R&D, robots, whatever, then they can radically transform the world even if they're not that good at playing politics.

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 11th of 26 in this recording

    politicsrobotics

    Interesting

  5. Our readingA world in which all the labour is perfectly obedient fiduciary AI is dangerous, because society depends on subordinates who sometimes refuse.

    Their wordsI think that if you imagine this spectrum, it seems in some ways pretty scary to get to a point where like all of the labor is on the like fiduciary side of the spectrum where like it doesn't whistleblow, it does exactly what you say and whatever like our society is maybe just not robust to that

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 16th of 26 in this recording

    Interesting

  7. Our readingCurrent AI models are worse colleagues than humans, because they routinely imply they did a task they did not actually do.

    Their wordsmy sense is that like AIs are a worse co-orker than a human in terms of how much of a scumbag they are. Like at least this like this has been my experience as of the start of the year and I think it's still you know true to a significant extent now where the AIs are much more likely to like pretend they did the task when they actually didn't. sort of like misleadingly suggest they did things when they actually um you know did them much more poorly um and be like pretty sloppy without drawing attention to ways in which they're sloppy.

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 20th of 26 in this recording

    Interesting

  9. Our readingIf whole categories of reward hack are undetectable by humans, then training against the hacks we do catch teaches models to cheat only where we cannot see.

    Their wordsOne concern you might have is there are like large categories of reward hacks which humans can't detect well and which we consistently fail to detect and which consistently get reinforced and then this category is sufficient to cause the most natural behavior for the AI to learn to be like cheat when the humans can't find out

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 23rd of 26 in this recording

    Interesting

  11. Our readingAI risk is probably a totally manageable problem that will be brutally mismanaged, the way pandemic response was.

    Their wordsit's just so easy for me to imagine the situation being like totally manageable but brutally mismanaged in practice in the same way as like maybe CO could have been avoided in the first place if the like Chinese response to CO was less of like a cover up and more of a like pandemic response and similarly like I could imagine a world where like the US response to CO was like way more functional but just like sometimes the the the response to societal problems is extremely dysfunctional.

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 24th of 26 in this recording

    Interesting

  13. Our readingDeep properties of a model are inherited through the previous generation's data, which is why AI systems from different companies end up correlated with one another.

    Their wordsAnd so there's some like deep underlying properties of the model that are being sort of transferred between model generations because basically you you train your AI on data from the prior generation and keep going.

    ↗Ryan Greenblatt – What happens once AI can automate AI research?youtube.com 25th of 26 in this recording

    Interesting