orthonormal

Titles as the sources gave them. In 1 entry, with 4 checked quotations.

What they said

Newest first. Each quotation was found word for word in the source.

“My proposal for a minimum requirement: a lab cannot train the next generation of models (in terms of capabilities) until they have adequately… resolved all significant alignment issues in the present generation of models.”
Minimum required pacing: finish aligning it before you train the next oneAI Alignment Forum, October 11, 2026. The post's core proposal, in its opening lines; the "…" replaces only a footnote marker.
“Multiple independent alignment organizations examine and red-team the newest model at many checkpoints throughout training, with full access to the data and with the best tools available.”
Minimum required pacing: finish aligning it before you train the next oneAI Alignment Forum, October 11, 2026. The first step of the author's 'vague, qualitative' process for judging whether alignment issues are adequately resolved.
“And it creates much better incentives for labs: in order to advance in capabilities, they'd have to push the frontier on alignment and interpretability techniques.”
Minimum required pacing: finish aligning it before you train the next oneAI Alignment Forum, October 11, 2026. The author's argument for why the rule would change labs' incentives.
“I still wouldn't say that this is fully adequate, but I'd treat it as a minimum requirement.”
Minimum required pacing: finish aligning it before you train the next oneAI Alignment Forum, October 11, 2026. The author's own verdict after laying out the deployment and reset steps.

Where they stand

Labels a machine proposed from the quotations above, one line per statement. As weeks accumulate, this is where a change of view shows up, with both quotations side by side.

A label describes one quotation, never the person.

Frontier AI development should slow down or pause

Leans agreelatest, machine label
  1. Leans agree: “My proposal for a minimum requirement: a lab cannot train the next generation of models (in terms of capabilities) until they have adequately… resolved all s…”

Independent testing should be required before release

Agreeslatest, machine label
  1. Agrees: “Multiple independent alignment organizations examine and red-team the newest model at many checkpoints throughout training, with full access to the data and …”

In the Base

Every entry they appear in.

  1. EssayAI Alignment ForumWrote it

    Minimum required pacing: finish aligning it before you train the next one

    On the AI Alignment Forum, an author writing as orthonormal proposes a minimum pacing rule: a lab may not train its next, more capable generation of models until significant alignment problems in the current one are adequately resolved, as judged by several independent alignment organizations. The author calls it a vague heuristic, not a concrete proposal, and lists open questions.