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

    GPU payments should be linked to identifiable, responsible humans, a user proposes

    The proposed legal requirement aims to reduce the risk of AI “escaping containment.” A reply calls customer identity checks reasonable but argues that collective control of AI is still needed.

    DP
    DH
    AT
    3 Sources, ,

    TLDR

    A user proposes making it illegal for GPU providers to accept payments not linked to an identifiable, responsible human, as one way to mitigate the risk of AI “escaping containment.” A reply calls know-your-customer (KYC) checks reasonable but argues that they would centralize control of both closed and open AI systems. In that commenter’s view, this would make closed systems more prone to existential risks while making open systems slightly less so. The reply sees a possible net benefit, but says more foundational options for collective control are needed.

    Combined views

    1.7K

    3 Sources, first seen 17d ago

    Combined views

    1.7K

    3 Sources, first seen 17d ago

    9 likes
    17d ago
    first seen 17d ago
    9 likes
    3 comments
    2 saves

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    3 comments
    2 saves

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    3 Sources

    @iamtraskKYC seems like a reasonable development, but it still leaves us in a world where the smartest AIs are unilaterally controlled by whomever has a copy. That creates centralization in closed systems, and radical decentralization in open systems. Your proposal centralizes both... which would make closed systems even more prone to xrisks but open systems a little less. Possibly a net-win, possibly part of the solution, but we need more foundational options for collective control for your proposal to really sing.
    @pfau@iamtrask That last turn of phrase sounds very LLM-written.
    @dhadfieldmenellThis is a good idea! KYC for compute seems like it would have many benefits. There’s certainly some downsides, but I think it makes sense above some threshold.

    3 Sources

    @iamtraskKYC seems like a reasonable development, but it still leaves us in a world where the smartest AIs are unilaterally controlled by whomever has a copy. That creates centralization in closed systems, and radical decentralization in open systems. Your proposal centralizes both... which would make closed systems even more prone to xrisks but open systems a little less. Possibly a net-win, possibly part of the solution, but we need more foundational options for collective control for your proposal to really sing.
    @pfau@iamtrask That last turn of phrase sounds very LLM-written.
    @dhadfieldmenellThis is a good idea! KYC for compute seems like it would have many benefits. There’s certainly some downsides, but I think it makes sense above some threshold.