Common Knowledge First: John Schulman on Voluntary Pacing Before Washington Acts

John Schulman did not sit for a private interview with Classy AI News in July 2026. What he did publish — alongside 1,309 other employees of frontier AI companies — was a signed comment on Pacing the Frontier, a petition asking the U.S. government to support an international effort to develop tools for deliberately pacing automated AI development. This piece reconstructs Schulman's public statement and the broader letter from pacingthefrontier.com. It is not a private conversation.

"Common knowledge about the possible need for coordination"

Schulman, now Chief Scientist at Thinking Machines, signed the July 2026 statement with a comment that cuts to the mechanism problem:

Signed because this statement helps establish common knowledge about the possible need for coordination mechanisms as automated AI research accelerates progress. I'd also like to see labs start designing these mechanisms voluntarily, even before the USG gets involved.

That voluntary-first clause matters. The petition itself does not ask labs to unilaterally halt training. It asks Washington to help build technical and governance tools so that pacing becomes an option when competitive pressure makes unilateral slowdown impossible.

What 1,310 employees are asking for

The core statement, published at pacingthefrontier.com, argues that leading AI companies may be close to automating AI research — and that capability acceleration could outrun society's ability to understand or control resulting systems.

The signatories request that the U.S. government:

support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.

Signatories include Jakub Pachocki (OpenAI), Jared Kaplan (Anthropic), Shengjia Zhao (Meta AI), Shane Legg (Google DeepMind), Ilya Sutskever (SSI), Dario Amodei (Anthropic), and Dawn Song (Meta), among others.

Why Schulman's framing differs from a pause letter

Schulman's emphasis on common knowledge echoes a game-theoretic diagnosis: every lab faces pressure not to slow down alone, but many researchers may already believe coordinated pacing will eventually be necessary. Making that belief public reduces the cost of building tools before a crisis forces reactive policy.

His call for voluntary lab-designed mechanisms also anticipates criticism that government-led pacing could become regulatory capture. Several other signatories echoed similar themes — Shengjia Zhao warned that frontier labs are "very close to AI that can exceed even the best people on almost every metric of intelligence," while Ilya Sutskever noted that dealing with future AI power "will require unprecedented measures."

What this is not

The petition is not a training halt, a compute cap, or a merger of frontier labs. It is a request for infrastructure — monitoring, governance, and coordination tooling — built on work already underway to track frontier model releases.

For Schulman, the immediate next step is design work inside labs, not waiting for legislation. Whether that voluntary track materializes before the next capability jump will be one of the defining policy stories of late 2026.

### Sources

• Pacing the Frontier — Statement and signatories (July 2026)
• Pacing the Frontier — John Schulman comment (July 2026)
• TechCrunch — Sam Altman is ready to decelerate (July 28, 2026)