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

    A proposed RSI approach pairs disposable agents with an evolving knowledge base

    The builder also describes post-training via communication, with a collaborative LLM teaming up across groups of varying sizes.

    Yisong YueYY
    Xuefei (Julie) Wang @ COLMX(
    2 Sources, 3h ago, first seen 3h ago

    TLDR

    In San Francisco for COLM2026, a builder says they’re working on two ideas: an RSI approach combining generic, disposable agents with an evolving knowledge base, and post-training via communication involving a collaborative LLM teaming up across groups of varying sizes.

    Combined views

    254

    2 Sources, first seen 3h ago

    Combined views

    254

    2 Sources, first seen 3h ago

    1 likes
    1 likes
    2 reposts
    Featured Source
    2 reposts

    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

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    2 Sources

    Xuefei (Julie) Wang @ COLM@xuefei_wIn SF for #COLM2026! What I’ve been building recently: • A new paradigm for RSI (generic, disposable agents + an evolving knowledge base) • Post-training via communication (one collaborative LLMs teaming up across variable group sizes) HMU if you're exploring RSI, post-training, or agents for science :)3h
    Yisong Yue@yisongyueRT @xuefei_w: In SF for #COLM2026! What I’ve been building recently: • A new paradigm for RSI (generic, disposable agents + an evolving kn…2h

    2 Sources

    Xuefei (Julie) Wang @ COLM@xuefei_wIn SF for #COLM2026! What I’ve been building recently: • A new paradigm for RSI (generic, disposable agents + an evolving knowledge base) • Post-training via communication (one collaborative LLMs teaming up across variable group sizes) HMU if you're exploring RSI, post-training, or agents for science :)3h
    Yisong Yue@yisongyueRT @xuefei_w: In SF for #COLM2026! What I’ve been building recently: • A new paradigm for RSI (generic, disposable agents + an evolving kn…2h