Transparency
How AI is used here.
AI is mostly a topic on this site, not a runtime dependency. Where it does run, this page documents
what it does, what data it touches, and where the computation happens.
1. The AI Maturity Diagnostic
Where it runs: entirely in your browser.
What it does: scores your answers against a deterministic rubric I authored across six dimensions of enterprise AI maturity, then surfaces three priority recommendations from a fixed matrix.
What it does not do: there is no language model, no API call, no telemetry. Your answers never leave your device.
Why this design: a diagnostic should be reproducible and inspectable. A deterministic rubric meets that bar; a black-box LLM does not.
2. The Insights search and topic filter
Where it runs: entirely in your browser.
What it does: client-side keyword filter over the published insights metadata.
A semantic-search upgrade using a small embedding model (all-MiniLM-L6-v2) computed at build time is on the roadmap. When it ships, embeddings ship as a static JSON file; nothing is generated at request time.
3. The market ticker
Where the data comes from: a scheduled GitHub Action queries a public, free Yahoo Finance endpoint and writes the results to a static JSON file in the repo.
What runs in your browser: a small fetch + render. Polling is gated by visibility (no requests when the ticker is off-screen).
No personal data is collected by this feature.
4. Things this site deliberately does not do
- No conversational chatbot impersonating me. I prefer real conversations and email.
- No third-party analytics that track individuals. If analytics is enabled, it is the privacy-respecting Plausible script (no cookies, no cross-site tracking).
- No model fine-tuning on visitor input. There is no visitor input to fine-tune on.