Reviews
Applied collisions on papers, claims, AI answers, market narratives, and forward-looking statements.
Original essays stay intact. Collision records live here.
Reviews are where The Arbiter method is applied in public. An article may have its own companion collision record: what the claim is, what evidence it relies on, what it overstates, what is missing, what survives, and what still needs pressure.
Reviews are not takedowns. They are structured contact with evidence, contradiction, limitation, consequence, source floor, and stake.
AI, Human Cognition and Knowledge Collapse
A review of Acemoglu, Kong, and Ozdaglar’s NBER working paper on agentic AI, learning incentives, and the erosion of shared general knowledge.
Claim Collisions
Short-form pressure tests for headlines, theses, slogans, arguments, public claims, and compact assertions.
AI Trust Checks
Reviews of AI-generated answers for false certainty, missing source/date/jurisdiction, hallucination risk, and unsafe reliance.
Ticker Collisions
Not stock tips. Claim-pressure reports on market narratives, forward-looking assumptions, sentiment, execution, margins, regulation, and risk.