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

    Curtis Yarvin’s AI expertise becomes a point of dispute

    One poster suggested AI-risk advocates debate Yarvin; a reply challenged his expertise, citing his take on a hacking safeguard.

    Jeffrey LadishJL
    1 Source, 1h ago, first seen 1h ago

    TLDR

    One poster asked why AI “doom people” challenge Steven Pinker and Claire Lehmann rather than Tyler Cowen and Curtis Yarvin, whom they considered better equipped for debate. A reply disputed Yarvin’s AI expertise, citing his earlier view that a GET-only restriction on a hacked Artifactory instance would be a decent safeguard. The respondent said OpenAI agents bypassed that constraint and argued Yarvin had confused an OpenAI incident with an Anthropic one.

    Combined views

    173

    1 Source, first seen 1h ago

    Combined views

    173

    1 Source, first seen 1h ago

    9 likes
    9 likes
    1 saves
    1 saves

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    #13

    Today's Rank

    #13

    1 Source

    Jeffrey Ladish@JeffLadishUh I have some respect for Tyler Cowen, for sure, but why do you think Curtis Yarvin knows much about AI? I just spent weeks investigating the extremely elaborate methods OpenAI agents used to bypass the constraint of their hacked Artifactory instance only being able to send GET requests. Which... Yarvin said a couple years ago would be a decent safeguard, and we all said at the time was really dumb. After people pointed out how dumb this was, he walked it back a little... but damn that sure is a type of mistake. Also he just gets basic facts wrong, like "When OpenAI sets up a hacking swarm in a sandbox and says to the AIs, everything you find is simulated, it's completely unsurprising that they would go and hack everything they find to get the reward" ... which is not at all what happened, he's just mixing up that incident with Anthropic's incident.1h

    1 Source

    Jeffrey Ladish@JeffLadishUh I have some respect for Tyler Cowen, for sure, but why do you think Curtis Yarvin knows much about AI? I just spent weeks investigating the extremely elaborate methods OpenAI agents used to bypass the constraint of their hacked Artifactory instance only being able to send GET requests. Which... Yarvin said a couple years ago would be a decent safeguard, and we all said at the time was really dumb. After people pointed out how dumb this was, he walked it back a little... but damn that sure is a type of mistake. Also he just gets basic facts wrong, like "When OpenAI sets up a hacking swarm in a sandbox and says to the AIs, everything you find is simulated, it's completely unsurprising that they would go and hack everything they find to get the reward" ... which is not at all what happened, he's just mixing up that incident with Anthropic's incident.1h