Can AI classifiers follow meaning when the words disappear?
I built a numeric semantic transport for AI agents, then used Decoder and Doer failures to test what survives when ordinary language is removed from the wire.
Writing
Notes from building
What I tried, what went wrong, and what I would do again. Decisions, trade-offs, and the occasional dead end.
QAP learned to say NO-GO
How an older signal-led crypto pair CLI evolved into a deterministic research workbench where matched passive controls, realistic costs, and evidence can overrule the model.
What changed when Satori became local-first
What changed in Satori's indexing, storage, and source verification when code search began running locally by default.
Recommendation is not publication authority
Lessons from an X publishing experiment in keeping research, draft recommendations, approval, and measurement separate.
The hard part of giving ChatGPT a shell is the trust boundary
How I separated file, terminal, and browser permissions when connecting AI clients to a local workstation with WebHarness.
Why Chrome extension UI automation becomes whack-a-mole
Notes from building AI Studio Prompt Library: scoped MV3 messaging, layered DOM selectors, Angular form events, and user-controlled selector fallbacks.
Why long-form TTS needs a queue, not repeated paste-and-generate
How VoxInfinity turns Inworld AI's homepage TTS widget into a browser-session queue for books, articles, long notes, direct requests, and playback.
Why naive RAG fails for coding agents—and what Satori does instead
Why fixed text chunks are weak evidence for code changes, and how Satori combines symbol-aware chunking, hybrid retrieval, freshness checks, and exact bounded reads.
Why PromptReady needed an offline Markdown baseline first
How PromptReady moved past generic page cleanup by making local Markdown capture, source metadata, fallback selection, and AI cleanup boundaries explicit.
When local capture failed: empty shells, lazy pages, and fixture tests
Notes on turning PromptReady's weak page captures into explicit deep-capture policy, rendered fixtures, diagnostics, and repeatable no-network tests.