DOOMBENCH.COM
Menu
Governance and control

Sutskever argues sufficiently powerful autonomous models should stay closed

In a full Stanford STVP talk, Ilya Sutskever argued that sufficiently capable autonomous models should not be open-sourced. He proposed capability-based thresholds, evaluations, rules, and certification for powerful systems.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

Send detailed or private feedback →
CURRENT ASSESSMENT · REVISION 1
AWAY FROM DOOM23confidence 58/100

Why it moved the index

The talk advances a specific access-control mechanism tied to autonomy and capability, supported by a full institutional transcript, but it remains a policy argument rather than an implemented restriction.

AUDIT TRAIL

Assessment history

  1. R1
    Away 23 · confidence 58

    Historical people backfill adds a dated full transcript with a distinct threshold-based access and certification proposal.

    14 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for Sutskever argues sufficiently powerful autonomous models should stay closed.

  1. DoomBench assesses “Sutskever argues sufficiently powerful autonomous models should stay closed” as evidence moving away from doom, with magnitude 23 and confidence 58 out of 100 in the governance and control category.

  2. The DoomBench assessment of “Sutskever argues sufficiently powerful autonomous models should stay closed” is based on reporting from Stanford Technology Ventures Program and records the editorial rationale, source quality, attribution,...

  3. DoomBench summarizes “Sutskever argues sufficiently powerful autonomous models should stay closed” as follows: In a full Stanford STVP talk, Ilya Sutskever argued that sufficiently capable autonomous models should not be open-sourced....

    https://www.doombench.com/news/sutskever-argues-sufficiently-powerful-autonomous-models-should-stay-closed-2023-04-26