The Compute Verification Post

Notes from our nights and weekends

Summary: We start with a primer on compute verification: understanding what’s going on inside a datacenter and how we can prove it to someone else. This involves proving certain properties about compute, e.g. where it is, what it’s doing, and how much was used, under particular trust assumptions. We then argue: Inference-only verification is the highest-value near-term property and the highest-priority target for large-scale demo buildout, International verification will never be trustless, only low-trust, so we should focus on stacking decorrelated defense-in-depth measures, Verification…

saved by

• Will Anderson

related reading

• The Compute Verification Postfirstscattering.com
• Escape Velocitywriting.antonleicht.me
• Verification Planai-2040.com
• [2606.05433] Zero knowledge verification for frontier AI training is possiblearxiv.org
• A shallow dive into formal verificationvitalik.eth.limo
• How to pace the US frontierblog.aifutures.org
• We Should Think More Critically About Compute Verificationsudarsh.com
• Convictionconviction.com
• Oversight Assistants: Turning Compute into Understandingbounded-regret.ghost.io
• Verifying International Agreements on AI: Six Layers of Verification for Rules on Large-Scale AI Development and Deploymentarxiv.org
• [2004.07213] Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claimsarxiv.org
• When AI Writes the World's Software, Who Verifies It? — Leonardo de Mouraleodemoura.github.io