AI News Daily한국어
Sunday, November 22, 2026

Story 8 of 8

Live updates

OpenAI calls for an industry-wide cap on AI development and publishes its pretraining numbers for the first time

First seen on X 4 hours ago@sama ♥ 2,90032 views

OpenAI calls for an industry-wide cap on AI development and publishes its pretraining numbers for the first time

OpenAI has published internal pretraining measurements it had never shared before and proposed that AI developers agree on common limits they will not cross. CEO Sam Altman called it "the most important thing OpenAI has published in awhile."

In a post titled "Perplexity and Intelligence" on Nov 22, chief scientist Jakub Pachocki charted base-model log-perplexity for OpenAI models from GPT-3.5 onward. Perplexity measures how well a model predicts the next token; lower is better. On an evaluation set drawn from OpenAI's internal codebase, GPT-3.5 sat at about 0.99 and GPT-4 at about 0.91. Pachocki wrote that once models dropped below 0.8, transformative capabilities emerged in cybersecurity and mathematics, and that after upgrading to a model at about 0.76, solutions to open math problems came "one after another in a single Slack thread." The chart ends with an unnamed "Internal model" at about 0.73, below the latest public model, GPT-6.2 Astra, at about 0.76. A footnote says the curve steepened in 2026, with AI already making meaningful design contributions to the latest models.

The core of the proposal is a ceiling. Pachocki expects further transformative capabilities in the 0.68-0.7 log-perplexity window. He proposed common limits, based on perplexity and other measurements, that developers would not cross until their ability to understand and monitor AI behavior reaches a concrete milestone, such as a successor to chain-of-thought monitors. He warned that an "avalanche" of AI-driven pretraining advances could arrive as soon as 2027, and said the urgency of shared safety standards outweighs the cost of showing competitors where OpenAI stands. On X, Altman said he hopes the field can agree to common limits, adding that labs committed under the White House Accords to set safety standards.

Pachocki's post on X passed 380,000 views and 2,700 likes in just over an hour, with many replies asking about the "internal model." The metric has limits: labs use different evaluation and training data, which makes numbers hard to compare, and chain-of-thought reasoning gains do not show up in perplexity. Pachocki listed both as open problems.

Comments 0

Select a sentence in the story (long-press on phones) to comment on it.

“This is the most important thing OpenAI has published in awhile.”

Next story →A 940-page, all-human Riemann proof took 12 years. AI helped flag problems in 8 hoursRead an AI Insight →If AI can ace the exam, how do you prove what you know? A Meta engineer says tests should watch how you solve problems with AI, not the answers

Get tomorrow's stories as an alert

We'll ping you when each new edition or breaking story lands. Free, no sign-up.

Daily morning AI news, new AI Insights, and a monthly briefing of top stories on the 1st.

AdPut your product here. About 5,000 people a month read AI news on this site.Advertise with us ›

More stories from the same day

  1. A 940-page, all-human Riemann proof took 12 years. AI helped flag problems in 8 hours
  2. Tesla's FSD v15 screen now looks like GTA, but the "25x safer" claim comes without a number
  3. "Just tapping my ring finger": a Neuralink patient flies a plane in GTA 6 with his mind
  4. "Colgate gets 90% accurate surveys from AI personas" drew 7,200 likes. The paper's 90% means something else
  5. Perplexity's CEO hears Gemini Neon beats Opus 5.5 and may make it the default, based on retention
  6. A day after Lean certified the Riemann proof, a claim that it's "invalid" hits 4,800 likes. Top reply: "not credible"
  7. "Use Claude Opus 5.5 free on AWS Quick" post gets 3,700 likes. AWS's own docs name neither the model nor "10 hours"
See all of Sunday, November 22, 2026 →

* Summaries and publishing on this page are done by an automated agent. Check the facts against the original links.

Privacy PolicySupportSend feedbackAdvertiseDiscord botClaude Code & CodexRSSListed on Sell With boost