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

    Commentator Calls OpenAI Depth Choice a War Crime

    Pseudonymous LisanBench creator reacts to OpenAI depth increase and reasoning release.

    AC
    ZN
    LA
    5 Sources, 29d ago, first seen 29d ago

    TLDR

    @scaling01 posted that OpenAI's choice of roughly 240 layers of effective depth was the single worst thing the company ever did. The user compared the decision to committing a war crime and opening Pandora's box. The same posts recalled the earlier strawberry reasoning announcement that prompted broad copying across the field. @scaling01 added that the war crime framing is an exaggeration yet argued some safety approaches could merit bans once the field matures.

    Combined views

    92.8K

    5 Sources, first seen 29d ago

    Combined views

    92.8K

    5 Sources, first seen 29d ago

    723 likes
    723 likes
    67 comments
    234 saves
    30 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    67 comments
    234 saves
    30 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

    5 Sources

    @scaling01just ~240 layers of effective depth what could possibly go wrong? this is genuinely the single worst thing that OpenAI ever did. it's like they committed a war crime. they opened Pandora's box. remember what happened when they announced strawberry and a new reasoning approach? everyone, their mum and entire family tree will now search for similar techniques and we will end up with > 500 layers of effective depth within the next 2 years in open weight models. good luck everyone.
    @andrew_n_carr@pfau @scaling01 I think the worst thing was the name "text davinci 003"...
    @zacharynadoRT @pfau: @scaling01 OK I really don't think tweaking your architecture is morally tantamount to a war crime? Like maybe you need to touch…

    5 Sources

    @scaling01just ~240 layers of effective depth what could possibly go wrong? this is genuinely the single worst thing that OpenAI ever did. it's like they committed a war crime. they opened Pandora's box. remember what happened when they announced strawberry and a new reasoning approach? everyone, their mum and entire family tree will now search for similar techniques and we will end up with > 500 layers of effective depth within the next 2 years in open weight models. good luck everyone.
    @andrew_n_carr@pfau @scaling01 I think the worst thing was the name "text davinci 003"...
    @zacharynadoRT @pfau: @scaling01 OK I really don't think tweaking your architecture is morally tantamount to a war crime? Like maybe you need to touch…