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

    Susan Zhang Questions OpenAI Nonprofit Era Replication

    AI researcher questions why labs fail to match OpenAI nonprofit phase output.

    SZ
    JF
    AG
    16 Sources, 63d ago, first seen 63d ago

    TLDR

    Susan Zhang posted that no group has matched OpenAI's early productivity achieved during its nonprofit years. The thread drew replies from academics and founders who pointed to differences in research freedom, team composition, infrastructure focus, and tolerance for exploratory work. Some accepted the implied challenge while others recalled specific early demos and the long timeline before major releases. The discussion centers on whether current constraints and incentives prevent similar results rather than on any single new claim or outcome.

    Combined views

    237.5K

    16 Sources, first seen 63d ago

    Combined views

    237.5K

    16 Sources, first seen 63d ago

    1.4K likes
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    1.4K likes
    78 comments
    199 saves
    33 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    78 comments
    199 saves
    33 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    16 Sources

    @suchenzangit's kind of crazy how no one has come close to producing what openai was able to produce in its nonprofit era with ~$100mil in funding not even ex-openai "veterans" skill issue? or gluttony issue? or not-delulu-enough issue?
    @unixpickleDamn what'd I do
    @DorialexanderTime to be a bit delusional: I could do a lot.
    @pfauAI during OpenAI's nonprofit era:
    @EMostaque@suchenzang There was a bit at stability ai where we got the formula right (before I got it very wrong, soz) Lots of freedom to do delulu things and folk from all sorts of background worked well Adding the business side & big tech folk worked… very badly :(
    @j_foerstChallenge accepted 🫡 🦋
    @jsuarez@j_foerst Limited compute goes much further with fast infra and training!
    @andrewgwilsCouple points here. (1) ChatGPT didn’t come out of nowhere. It was 7 years after founding before that was released. (2) During that long period, exploratory, non-incremental, and open research was supported in a way that it typically isn’t anymore. Now there’s a herd mentality.
    @avt_imRemember the OpenAI Rubik’s Cube demo?
    @txhfView post on X

    16 Sources

    @suchenzangit's kind of crazy how no one has come close to producing what openai was able to produce in its nonprofit era with ~$100mil in funding not even ex-openai "veterans" skill issue? or gluttony issue? or not-delulu-enough issue?
    @unixpickleDamn what'd I do
    @DorialexanderTime to be a bit delusional: I could do a lot.
    @pfauAI during OpenAI's nonprofit era:
    @EMostaque@suchenzang There was a bit at stability ai where we got the formula right (before I got it very wrong, soz) Lots of freedom to do delulu things and folk from all sorts of background worked well Adding the business side & big tech folk worked… very badly :(
    @j_foerstChallenge accepted 🫡 🦋
    @jsuarez@j_foerst Limited compute goes much further with fast infra and training!
    @andrewgwilsCouple points here. (1) ChatGPT didn’t come out of nowhere. It was 7 years after founding before that was released. (2) During that long period, exploratory, non-incremental, and open research was supported in a way that it typically isn’t anymore. Now there’s a herd mentality.
    @avt_imRemember the OpenAI Rubik’s Cube demo?
    @txhfView post on X