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

    Deepak Pathak Calls for In-Context Robot Learning

    Skild AI CEO contrasts current robotics training with ChatGPT success model.

    DP
    CP
    KD
    6 Sources, 27d ago, first seen 27d ago

    TLDR

    Deepak Pathak, identified as co-founder and CEO of Skild AI along with a faculty position at Carnegie Mellon, posted a comparison between language model development and robotics. He noted that ChatGPT succeeded without post-training on every individual user interaction. Pathak said robotics currently requires such per-user training, which he believes prevents similar growth. Instead, he stated robots must learn in context. The post recommends an article by Chris J. Paxton that covers this emerging approach and how S1 fits into it. Attached media in the post displays three screens.

    Combined views

    21.9K

    6 Sources, first seen 27d ago

    Combined views

    21.9K

    6 Sources, first seen 27d ago

    193 likes
    193 likes
    7 comments
    80 saves
    42 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    7 comments
    80 saves
    42 reposts

    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

    6 Sources

    @deepakpathakYou don't have to post-train ChatGPT on every user. If you did, it would never have taken off. Yet this is exactly how robotics works today For robots to take off, they need to learn in-context. Great article from @Chris_J_Paxton on this new paradigm and S1's place in it:
    @ariel__shaulovGreat blog by @chris_j_paxton. The interesting shift is from programming robots to teaching them. In-context learning could turn demonstrations into the interface for robotics, making new skills something you show, not something you retrain for.
    @chris_j_paxtonRT @ariel__shaulov: Great blog by @chris_j_paxton. The interesting shift is from programming robots to teaching them. In-context learning c…
    @CSProfKGDRT @chris_j_paxton: In-Context Learning Results Hint at a “GPT Moment” for Robotics For general-purpose robots to be useful and economical…

    6 Sources

    @deepakpathakYou don't have to post-train ChatGPT on every user. If you did, it would never have taken off. Yet this is exactly how robotics works today For robots to take off, they need to learn in-context. Great article from @Chris_J_Paxton on this new paradigm and S1's place in it:
    @ariel__shaulovGreat blog by @chris_j_paxton. The interesting shift is from programming robots to teaching them. In-context learning could turn demonstrations into the interface for robotics, making new skills something you show, not something you retrain for.
    @chris_j_paxtonRT @ariel__shaulov: Great blog by @chris_j_paxton. The interesting shift is from programming robots to teaching them. In-context learning c…
    @CSProfKGDRT @chris_j_paxton: In-Context Learning Results Hint at a “GPT Moment” for Robotics For general-purpose robots to be useful and economical…