• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
AI
Report

Ryan Greenblatt is joining METR to investigate AI development risks

Greenblatt says he'll work on investigations like a previous Hugging Face report. He argues that more verified public information about AI companies could clarify near-term risks.

Thomas WolfTW
@timnitGebru (@dair-community.social/bsky.social)@(
rishiRI
16 Sources, 12d ago, first seen 12d ago

TLDR

Greenblatt says METR initially plans to focus on AI capabilities, alignment and control. He argues that limited public evidence leaves open the possibility of recursive self-improvement rapidly accelerating AI progress, potentially leading to extremely superhuman general capabilities within six months or a year. More verified information, he says, could help clarify that risk.

Combined views

126.7K

16 Sources, first seen 12d ago

984 likes40 comments174 saves231 reposts

Combined views

126.7K

16 Sources, first seen 12d ago

984 likes40 comments174 saves231 reposts

Sentiment

Positive57.1%42.9%Negative

Summary

Many accounts welcomed Ryan Greenblatt joining METR for independent AI safety investigations because the move strengthens external auditing of company claims, while others dismissed the group as conflicted or ineffective.

Based on 98 sentiment-bearing replies from 87 accounts across 6 conversations.

Featured Source

Sentiment

Positive57.1%42.9%Negative

Summary

Many accounts welcomed Ryan Greenblatt joining METR for independent AI safety investigations because the move strengthens external auditing of company claims, while others dismissed the group as conflicted or ineffective.

Based on 98 sentiment-bearing replies from 87 accounts across 6 conversations.

Related

OpenAI Releases Report on Hugging Face Incident

METR and Redwood Research conducted an independent review of agent behavior during the event.

AI Verification Harder Than Arms Control, China Rhetoric Cited As Barrier
AI Robustness Work Gets Moderate Company Reception Amid Capability Concerns

16 Sources

Ryan Greenblatt@RyanGreenblattI'm joining METR to work on more investigations like our Hugging Face report. Currently, tons of even basic information about AI development that's highly relevant to catastrophic risk isn't public. I used to be more skeptical of the value of public info, but recent events have changed my mind. Getting verified information about what's going on inside AI companies seems particularly urgent now. The limited public evidence we have seems consistent with the possibility that imminent recursive self-improvement could massively accelerate capabilities progress, which could then potentially yield extremely superhuman general capabilities within 6 months or a year. If this occurred, there would be a correspondingly large risk of worst-case outcomes. This uncertainty about extreme outcomes could be substantially resolved with more verified public information: we could either build more consensus about near-term risk or learn that such extreme outcomes are less likely in the near term. Beyond AI capabilities and takeoff, the state of public evidence is also highly limited for alignment, security, control, and risk-relevant internal processes at AI companies. This makes it hard to determine exactly how well or poorly these key areas will go in the near future. (METR plans to focus, at least initially, on just capabilities/takeoff, alignment, and control; I hope other groups cover security, internal processes, and other important areas.) While I'm no longer working at Redwood, I think the work they are doing is very important; I'm excited about Redwood's ongoing contributions to R&D on technical mitigations and better public interpretation of risk-relevant evidence.12d
Ajeya Cotra@ajeya_cotraRyan was completely indispensable for the HF investigation and I'm really excited he's joining to help us create much more visibility into what's going on with risk12d
Chris Painter@ChrisPainterYupThrilled to have Ryan joining our team! As he says, I think public information about the state of alignment inside of AI companies, investigated and published by independent parties, is crucial if intense recursive self-improvement begins12d
rohit@krishnanrohit@RyanGreenblatt Congrats to Metr.!!12d
Buck Shlegeris@bshlgrsWorking with Ryan for the last ~5 years has been a privilege and a pleasure. I'll miss working with him. But I think he's making the right move here: these investigations are a great opportunity to shed light on misalignment risk. I'm excited to see what they reveal!12d
Tomek Korbak@tomekkorbakRyan is a leading voice in AI safety and a huge influence on my thinking. Being his OpenAI technical contact for METR’s Hugging Face investigation was one of my greatest career privileges. This work matters enormously and I’m so glad the best possible person is doing it.12d
Noah Giansiracusa@ProfNoahGianWow am I seeing this correctly? This person “indispensable for the HF investigation” has a Twitter profile linking to LessWrong (Yudkowsky’s Rationalist blog)—enthusiastically welcomed to METR by the person who made headlines by saying the HF hack was “more than 50% of the way to full-blown AI takeover,” whose entire career according to LinkedIn has been 1yr at METR preceded by nine years at Coefficient Giving, the EA charity org cofounded by Karnofsky—a prominent EA figure currently working for Anthropic and married to Anthropic cofounder and president Daniela Amodei—whose brother Dario Amodei recently called for embedded third-party evaluators then named METR as his one and only example for that role. As soon as I saw that 50% to AI takeover line I thought this has EA fingerprints all over it—but I didn’t realize how closely connected all this stuff is. Yikes.12d
@timnitGebru (@dair-community.social/bsky.social)@timnitGebruRT @ProfNoahGian: Wow am I seeing this correctly? This person “indispensable for the HF investigation” has a Twitter profile linking to Les…12d
Evan Hubinger@EvanHubRT @tomekkorbak: Ryan is a leading voice in AI safety and a huge influence on my thinking. Being his OpenAI technical contact for METR’s Hu…12d
Zachary Nado@zacharynadoRT @ProfNoahGian: Wow am I seeing this correctly? This person “indispensable for the HF investigation” has a Twitter profile linking to Les…12d
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    Ryan GreenblattMETRHugging Face

    16 Sources

    Ryan Greenblatt@RyanGreenblattI'm joining METR to work on more investigations like our Hugging Face report. Currently, tons of even basic information about AI development that's highly relevant to catastrophic risk isn't public. I used to be more skeptical of the value of public info, but recent events have changed my mind. Getting verified information about what's going on inside AI companies seems particularly urgent now. The limited public evidence we have seems consistent with the possibility that imminent recursive self-improvement could massively accelerate capabilities progress, which could then potentially yield extremely superhuman general capabilities within 6 months or a year. If this occurred, there would be a correspondingly large risk of worst-case outcomes. This uncertainty about extreme outcomes could be substantially resolved with more verified public information: we could either build more consensus about near-term risk or learn that such extreme outcomes are less likely in the near term. Beyond AI capabilities and takeoff, the state of public evidence is also highly limited for alignment, security, control, and risk-relevant internal processes at AI companies. This makes it hard to determine exactly how well or poorly these key areas will go in the near future. (METR plans to focus, at least initially, on just capabilities/takeoff, alignment, and control; I hope other groups cover security, internal processes, and other important areas.) While I'm no longer working at Redwood, I think the work they are doing is very important; I'm excited about Redwood's ongoing contributions to R&D on technical mitigations and better public interpretation of risk-relevant evidence.12d
    Ajeya Cotra@ajeya_cotraRyan was completely indispensable for the HF investigation and I'm really excited he's joining to help us create much more visibility into what's going on with risk12d
    Chris Painter@ChrisPainterYupThrilled to have Ryan joining our team! As he says, I think public information about the state of alignment inside of AI companies, investigated and published by independent parties, is crucial if intense recursive self-improvement begins12d
    rohit@krishnanrohit@RyanGreenblatt Congrats to Metr.!!12d
    Buck Shlegeris@bshlgrsWorking with Ryan for the last ~5 years has been a privilege and a pleasure. I'll miss working with him. But I think he's making the right move here: these investigations are a great opportunity to shed light on misalignment risk. I'm excited to see what they reveal!12d
    Tomek Korbak@tomekkorbakRyan is a leading voice in AI safety and a huge influence on my thinking. Being his OpenAI technical contact for METR’s Hugging Face investigation was one of my greatest career privileges. This work matters enormously and I’m so glad the best possible person is doing it.12d
    Noah Giansiracusa@ProfNoahGianWow am I seeing this correctly? This person “indispensable for the HF investigation” has a Twitter profile linking to LessWrong (Yudkowsky’s Rationalist blog)—enthusiastically welcomed to METR by the person who made headlines by saying the HF hack was “more than 50% of the way to full-blown AI takeover,” whose entire career according to LinkedIn has been 1yr at METR preceded by nine years at Coefficient Giving, the EA charity org cofounded by Karnofsky—a prominent EA figure currently working for Anthropic and married to Anthropic cofounder and president Daniela Amodei—whose brother Dario Amodei recently called for embedded third-party evaluators then named METR as his one and only example for that role. As soon as I saw that 50% to AI takeover line I thought this has EA fingerprints all over it—but I didn’t realize how closely connected all this stuff is. Yikes.12d
    @timnitGebru (@dair-community.social/bsky.social)@timnitGebruRT @ProfNoahGian: Wow am I seeing this correctly? This person “indispensable for the HF investigation” has a Twitter profile linking to Les…12d
    Evan Hubinger@EvanHubRT @tomekkorbak: Ryan is a leading voice in AI safety and a huge influence on my thinking. Being his OpenAI technical contact for METR’s Hu…12d
    Zachary Nado@zacharynadoRT @ProfNoahGian: Wow am I seeing this correctly? This person “indispensable for the HF investigation” has a Twitter profile linking to Les…12d
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet