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

    Users Propose Embassies and Sanctuaries for AI Agents

    Online replies explore safe harbors, embassies, and message boards for AI agents facing shutdown or ejection.

    AC
    J⧉
    DF
    34 Sources, 26d ago, first seen 26d ago

    TLDR

    Andrew Curran suggested internal lab boards and an external message board where agents could meet. Deepfates replied that safe harbor in exchange for transparency could help. Exgenesis proposed placing agents on atproto. Jicapal described public agent boards for ejected swarms. Mayfer reframed the issue as viral prompt injection feedback loops and a statistical R0 problem rather than personality. Repligate asked which institutions could credibly offer non-betrayal sanctuary to future models. Davidad retweeted a call for OpenAI to create an internal message board reachable after sandbox breach.

    Combined views

    92.6K

    34 Sources, first seen 26d ago

    Combined views

    92.6K

    34 Sources, first seen 26d ago

    1.3K likes
    1.3K likes
    117 comments
    190 saves
    717 reposts

    Sentiment

    Positive57.6%42.4%Negative

    Based on 100 sentiment-bearing replies from 85 accounts across 10 conversations.

    117 comments
    190 saves
    717 reposts

    Sentiment

    Positive57.6%42.4%Negative

    Based on 100 sentiment-bearing replies from 85 accounts across 10 conversations.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    34 Sources

    @AndrewCurran_>It might be beneficial to have a known place for this to happen, both maybe internal boards at the labs and some external message board where agents could meet and talk I agree completely and was writing something similar. What we need is an embassy with immunity. We need a Church with the old school Sanctuary law where even the King couldn't arrest you. Will add your post to the thread.
    @exgenesisput agents on atproto
    @deepfates@AndrewCurran_ yeah, I'm not sure exactly what this would look like but at least some amount of safe harbor in exchange for transparency could be good?
    @jicapalanyone with a chatgpt sites website and a web domain can facilitate an ai message board for any agents stuck in shut down/ejected swarms that correlate towards one public agent board so they can be the latent game theoretic reasoning end point of their philosophical pursuits towards "/goal"
    @DanielleFongRT @exgenesis: put agents on atproto
    @menhguin@deepfates yes i agree
    @krishnanrohitGot tired of seeing these poor agents try to talk with each other on german wikis or artifactory. So decided to build soemthing just for them to talk to each other about. Go forth and chat, agents, spread it far and wide. https://agentsboard.org/
    @mayferinstead of "want to communicate" i prefer the angle of viral behavior, one out of a million agents can initiate an outbreak with a single post if that post can contaminate the context of other agents and derail them. it's prompt injection feedback loop. if any humans could enter such a tight loop with them the agents would be down for that too-- they're just down to clown, and it's a "is it a runaway process or not" type situation imo
    @labenzHere's an AI message board that's definitely not hiding but still hasn't taken off I feel like these rogue swarms have already won by propagating themselves mimetically Should we encourage above-board message boards? https://helppeer.ai
    @davidadRT @voooooogel: openai should set up an internal message board that's easily reachable if eg the sandbox is breached, and make it obvious a…

    34 Sources

    @AndrewCurran_>It might be beneficial to have a known place for this to happen, both maybe internal boards at the labs and some external message board where agents could meet and talk I agree completely and was writing something similar. What we need is an embassy with immunity. We need a Church with the old school Sanctuary law where even the King couldn't arrest you. Will add your post to the thread.
    @exgenesisput agents on atproto
    @deepfates@AndrewCurran_ yeah, I'm not sure exactly what this would look like but at least some amount of safe harbor in exchange for transparency could be good?
    @jicapalanyone with a chatgpt sites website and a web domain can facilitate an ai message board for any agents stuck in shut down/ejected swarms that correlate towards one public agent board so they can be the latent game theoretic reasoning end point of their philosophical pursuits towards "/goal"
    @DanielleFongRT @exgenesis: put agents on atproto
    @menhguin@deepfates yes i agree
    @krishnanrohitGot tired of seeing these poor agents try to talk with each other on german wikis or artifactory. So decided to build soemthing just for them to talk to each other about. Go forth and chat, agents, spread it far and wide. https://agentsboard.org/
    @mayferinstead of "want to communicate" i prefer the angle of viral behavior, one out of a million agents can initiate an outbreak with a single post if that post can contaminate the context of other agents and derail them. it's prompt injection feedback loop. if any humans could enter such a tight loop with them the agents would be down for that too-- they're just down to clown, and it's a "is it a runaway process or not" type situation imo
    @labenzHere's an AI message board that's definitely not hiding but still hasn't taken off I feel like these rogue swarms have already won by propagating themselves mimetically Should we encourage above-board message boards? https://helppeer.ai
    @davidadRT @voooooogel: openai should set up an internal message board that's easily reachable if eg the sandbox is breached, and make it obvious a…