Proof of Mechanism

The book uses AttestOS, a bot-verification service, as a candid running case. Its first version (26 October 2026) presented every conventional sign of completeness while lacking most mechanisms. An audit the next day logged 65 findings, 18 of them critical. The service was rebuilt that evening and checked again on 28 October. The book does not state how the first version was produced. It treats generation by AI as one hypothesis the evidence is consistent with. AttestOS had no customers and had taken no payments at the time of writing. The audit and rebuild were carried out by the owner working with AI coding agents: a bounded case, not an independent evaluation.

Contents

1. The demo that wasn't
2. Why surfaces got cheap
3. Simulated completeness
4. Trust products and credence goods
5. Fabricated state
6. Theatrical actions
7. Unenforced controls
8. Dangling references
9. Borrowed authority and test theatre
10. The six probes
11. The one-hour outsider audit
12. Synthetic evidence and the third party
13. Reading an audit as a leader
14. Rebuild or repair
15. Mechanism first
16. One source of truth for every claim
17. Honest limits as a product feature
18. Truthful outreach
19. Status honesty
20. Verifiability as strategy
21. Governing agents with evidence
22. Metrics, incentives and Goodhart
23. Law, ethics and reputation
24. A ninety-day proof-of-mechanism programme
25. Framework compendium
26. Facilitator's guide
27. Glossary
28. References