Vigilia.
Gli avvertimenti
Rischio di estinzioneIndustriaFounder, Situational Awareness LP4 voci a registro

Seguire

Leopold Aschenbrenner

Ex ricercatore di superallineamento di OpenAI e autore di «Situational Awareness», oggi guida una società di investimento focalizzata sull’AGI e avverte che il controllo di una superintelligenza attesa entro fine decennio è un problema irrisolto e che una corsa tra Stati Uniti e Cina potrebbe finire in catastrofe.

Ruoli

  1. feb 2022nov 2022Researcher, FTX Future FundResigned with the rest of the team when FTX filed for bankruptcy.Fonte
  2. 2023apr 2024Researcher, Superalignment team, OpenAIFired in April 2024 over an alleged information leak, a characterisation he disputes.Fonte
  3. 2024oggiFounder, Situational Awareness LPAGI-focused investment firm backed by Patrick and John Collison, Nat Friedman and Daniel Gross.Fonte

A registro

  1. 4 giu 2024 · Podcast · Dwarkesh Podcast

    This whole scheme is terrible. Being in neck and neck race at this point is terrible.

    Founder, Situational Awareness LP all’epoca

    On the danger of a close superintelligence race between the US and China.

    Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history

  2. giu 2024 · Articolo scientifico · situational-awareness.ai

    If we're lucky, we'll be in an all-out race with the CCP; if we're unlucky, an all-out war.

    Founder, Situational Awareness LP all’epoca

    Situational Awareness: The Decade Ahead

  3. giu 2024 · Articolo scientifico · situational-awareness.ai

    Reliably controlling AI systems much smarter than we are is an unsolved technical problem… failure could easily be catastrophic.

    Founder, Situational Awareness LP all’epoca

    The 165-page essay series published two months after he left OpenAI.

    Situational Awareness: The Decade Ahead

  4. 29 mar 2023 · Post · For Our Posterity

    Scalable alignment—aligning superhuman AGI systems—is a real, unsolved problem. Current alignment techniques rely on human supervision, but as models become superhuman, humans won't be able to reliably supervise them.

    Researcher, Superalignment team, OpenAI all’epoca

    Written while he was on OpenAI's superalignment team.

    Nobody's on the ball on AGI alignment