01
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.
Writing
Notes archive
Practical engineering notes, project retrospectives, and explanations of implementation decisions and failures.
02
What changed when Satori became local-first
Moving Satori from a connected-vector-store story to a local-first code map forced clearer ownership between source evidence, model-specific vectors, publications, runtime compatibility, and agent-facing tools.
03
Recommendation is not publication authority
How a local X publishing tool grew into a network-growth system with relationship intelligence, writing strategy, measured learning, explicit main-feed approval, and tightly bounded reply automation.
04
The hard part of giving ChatGPT a shell is the trust boundary
What changed when I stopped treating an MCP-to-WSL bridge as a connectivity problem and separated files, Bash, code intelligence, terminal lifetime, browser control, and constrained-client access into explicit authority boundaries.
05
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.
06
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.
07
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.
08
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.
09
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.
10
Why AI cleanup should not be the extraction engine
How PromptReady changed AI cleanup from raw HTML rewriting into a quality-gated pass over a local Markdown baseline.
/ Topics