AI

Tom Milton on how we could verify a US–China deal on AI

Read full article at source ↗

Based on the episode description, not a full transcript. Source evidence retrieved 23/10/2026, 16:03:24 UTC.

Briefings

Tom Milton argues that verifying US-China AI cooperation requires shutting down data centers to prove compliance, as no alternative verification tools exist. He cites Amodo Design's prototype tools and notes that only ~50 engineers worldwide work on compute verification, compared to billions spent on Cold War nuclear detection. The episode was recorded September 29, 2026, and covers topics from evasion tactics to hardware engineering challenges.

Key points

• Tom Milton proposes shutting down data centers as the primary method to verify AI compliance during a potential US-China deal, arguing this would prevent cheating while preserving most AI benefits.
• Amodo Design's October 2026 Frontier Safety Roadmap update reports a prototype tool that runs on production inference to check if outputs match specific model weights.

Context

Tom Milton, CEO of Amodo Design, discusses how to verify AI slowdowns in a US-China deal. The evidence includes his argument that shutting down data centers is the most likely verification method and a reference to Anthropic's October 2026 Frontier Safety Roadmap update.

Limitations

Evidence is limited to a single podcast episode from September 2026. No additional background on the US-China AI deal, specific dates of the deal, or long-term verification tool development timelines is supplied.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 23/10/2026, 16:04:32 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeigkxlahoiapyqvtdnhugazj3vq2owkmd4q4doh36szwr7yawfnpw4