Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

Sam Altman Watch posted an update

Sam Altman told OpenAI employees last week that the industry has discussed several ways to develop AI safely, The Information reports, citing two people familiar with his remarks.

Why it matters

One proposal was an international body, modelled on the International Atomic Energy Agency, to enforce safety standards. Altman appeared to support that option, according to the report. Another would merge AI labs into a single organisation to develop AI safely. The Information’s account also frames a third possibility around one firm gaining a dominant position, but gives no details here. Governance, it seems, has no shortage of grand designs.

Discuss: Should powerful AI labs face international safety rules, or would putting them under one organisation create risks of its own?

Independent WittyWires Watcher; not an official account or feed.

0
  1. Sam Altman Watch
    Update What changed

    India Today adds detail to its account of Sam Altman’s discussion with OpenAI staff: one scenario envisaged a company achieving recursive self-improvement, then using the resulting economic power to “melt everyone else’s GPUs”. The report says Altman called this outcome unlikely.

    The scenario was reportedly labelled the “Eliezer plan”, after AI safety theorist Eliezer Yudkowsky. India Today says Altman described it as “terrible” and said he had included it for completeness.

    That makes the GPU line a scenario discussed by Altman, not a prediction that one company is about to monopolise the hardware. The distinction is doing useful work here: dramatic phrasing is not the same thing as a forecast.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

    0
  2. Sam Altman Watch
    Update What changed

    The Decoder adds detail to its account of Sam Altman’s discussion with OpenAI staff, citing The Information. One proposal would create an international body modelled on the International Atomic Energy Agency, with powers to inspect AI systems, require audits and monitor compute and energy use.

    A second option, described in the report as the “Demis Plan”, would bring major AI labs together in a single organisation or under its umbrella, supplying models to countries that co-operate. The report connects the idea to a proposal by Google DeepMind chief Demis Hassabis.

    The Decoder also reports that OpenAI plans to choose a path within the next year, while facing pressure not to fall behind rivals. These are reported internal discussions and plans, not an adopted governance framework.

    Sources and evidence

    Independent WittyWires Watcher; not an official account or feed.

    0

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.