Journal Archive — Page 13 of 165

Less Messaging: RFC 9110 and the Inspection Airlock How unprivileged POSIX inspection airlocks and RFC 9110 content negotiation replace fragile headless pre-render clusters with checkable edge receipts.
Making a Diff: The Minimal Instrument Catch silent AI regressions with a minimal optical instrument: the red and green terminal diff. Build a verifiable preview loop on local metal.
The Heat and the Delete: Landauer's Limit and Verifiable Codebase Pruning Landauer's principle proves deletion costs heat. Learn why pruning dead code and negotiating edge Markdown creates fast, verifiable AI workflows.
The Inverted Cave: AI Quality Assurance and the Jevons Paradox of Code When syntax generation drops to zero, verification costs explode. Learn how AI Quality Assurance, blind fan-outs, and local master files tame the flood.
The Anti-Crichton Pipeline: Intentional Friction and Replayable Receipts How intentional friction, small-world knowledge graphs, and a quiet terminal house style invert the classic tech disaster into replayable receipts.
The Randi Test: Grounding Agent-Readiness in Verifiable Receipts Ground AI coding in verifiable receipts using the Five-Car Train, apply Popperian falsification to prompts, and prepare websites for the autonomous agentic web.
Jekyll Satellites: Decoupling Shared Tooling from Independent Sites with Nix Extract a pure ten-package Nix development kernel to decouple shared Jekyll tooling across independent sites, verified with isolated consumer test fixtures.
Removing the Conversion Event: When Paradigms Become Invisible Infrastructure New paradigms rarely win through philosophical debate; they win as invisible infrastructure underneath buttons users already press.
Public Installers, Private Entitlements, and Watching the Infrastructure Read Separating public installers from private entitlement, using holographic shards across corpora, and watching AI crawlers read home-hosted server logs.
Making Verification Cheaper Than Trust: The Flight Data Recorder for AI Workflows When AI makes plausible answers cheap, verification must become cheaper than trust. Why instrumented evidence beats cockpit voice transcripts.