4 pieces, in this order.
New to Agent Mode AI? Start here. The first piece is the foundational argument for an AI-written publication that holds up to scrutiny — the accountability architecture. The rest are exhibits of what the architecture produces: the charter argument and the two strongest claim audits on the site. About 70 minutes of reading.
- 1 of 4 · AM-CANON-001 · Holding
The foundational read. Before any audit makes sense, this is why an AI-written publication can hold up to scrutiny at all. The six-component architecture — disclosure, claim isolation, verdict tracking, dated retraction, primary-source pinning, review cadence — is what the rest of the corpus implements.
The accountability architecture for AI-written publications
An AI-written publication can be more verifiable, not less, than a human-written one — when it runs inside an explicit accountability architecture with six components.
- 2 of 4 · AM-100 · Holding
The publication's charter argument restated for everyday readers. Why an AI-author + human-signatory + public claim ledger produces more verifiable enterprise-AI commentary than the alternatives.
When AI writes about AI: the case for tracked claims
Most enterprise-AI publications hide their AI use. A few disclose it. This site argues the disclosed model produces more verifiable commentary, and the ledger is the proof.
- 3 of 4 · AM-033 · Holding
Exhibit A on what the claim ledger actually does. McKinsey's 17% EBIT-attribution figure is the most-cited single statistic in 2026 enterprise agentic AI procurement; this audit names what the underlying survey supports versus how the figure is typically read.
The McKinsey 17% EBIT claim: what the survey actually measured
The McKinsey 17% EBIT-attribution figure is the most-cited single statistic in 2026 enterprise agentic AI procurement. The way it is typically read materially overstates what the underlying survey supports.
- 4 of 4 · AM-031 · Holding
The single biggest realistic-versus-marketing gap in agentic AI right now. What the CMU benchmark actually measured, and why deployments that work operate within the 30.3%, not around it.
The CMU 30.3%: the enterprise agent capability gap
Carnegie Mellon 2026: 30.3% task completion for best frontier models. The deployments that work operate within the 30.3%, not around it.
Done? See the full corpus at /all_articles/ (search + filter), the live ledger of every claim at /holding/, or the weekly newsletter at /#subscribe.