The Capybara in the Room
A source-linked update on Claude Mythos: distinguish naming speculation, vendor evaluations and actual model availability.
Latest
Building an AI Audit for Things 3
A Claude Code audit for Things 3: read SQLite safely, score task structure, propose fixes, and apply approved changes through AppleScript.
Configure Claude Code for Maximum Power
Configure Claude Code model, effort and permission modes explicitly, with documented limits and realistic recovery expectations.
One Env Variable That Makes Claude Code Smoother
An experimental Claude Code rendering option that improved flicker in one reported test; support and behavior can change.
How to Install Claude Code the Right Way in 2026
Compare Claude Code installation methods, identify package ownership, and migrate to the native installer with verification.
The Third Pass
One agent implements, another reviews original evidence, and a human verifies the result. Separate responsibilities improve scrutiny without guaranteeing correctness.
A Programming Language with No Compiler
A specification-first, English-like language with AI-generated interpreters and a shared test suite. Execution still depends on host toolchains.
Human Language Is the Best Programming Language
Natural language can help express software requirements, but generated prototypes still need engineering verification before deployment.
Decade Zero: A Realistic Blueprint for 2026–2035
A speculative scenario essay about AI, work and institutions over the next decade. Dates and future outcomes are assumptions, not calibrated predictions.
Spec-Driven Agentic Development
A specification-first workflow can reduce requirements-related rework. Review and tests still need to verify both the specification and its implementation.
Claude Code Commands You Didn’t Know Existed
A practical reference to Claude Code commands, including conversation controls, checkpoint limits, model selection and terminal configuration.
Claude Code Skills: The Complete Guide
A practical guide to Claude Code skills: reusable instructions, invocation, frontmatter and examples, with explicit limits on permissions and hook guarantees.