Read every resource on the site, then download the original file when you want to use or reproduce it.
All resources are included in this website source. No account, purchase, external viewer, or new browser tab is required.
Reusable worksheets for deployment choices, data handling, model comparison, and a seven-day local AI setup.
Compare local, cloud, and hybrid options against task, data, hardware, cost, failure, and approval constraints.
Use it to: choose where an AI task should run before implementation.
Included files: ai-deployment-decision.csv.
Record data categories, processing locations, retention, access, and required handling controls.
Use it to: map what data enters an AI workflow and how it is handled.
Included files: data-inventory.csv.
Compare model version, settings, correctness, latency, memory use, and observed failure behavior.
Use it to: record repeatable evidence before selecting an AI model.
Included files: model-evaluation.csv.
A structured daily record for hardware inventory, installation, testing, privacy review, and next decisions.
Use it to: keep a seven-day local AI setup reviewable and reproducible.
Included files: seven-day-setup-journal.md.
Small deterministic examples with source, test cases, expected outcomes, and explicit limitations.
Allows one validated read operation while denying deletion, malformed identifiers, replayed requests, and instruction-like note content.
Use it to: study a small allowlist and input-validation boundary for agent tools.
Included files: README.md, permission-gate.mjs, test.mjs.
Validates required structure, exact source quotations, supported values, and a human-review flag across five synthetic cases.
Use it to: study deterministic validation around an AI-assisted summary step.
Included files: README.md, validate.mjs, cases.json, schema.json, source.txt.
Bounded exercises and observations with environments, commands, results, and limitations preserved.
A before-and-after clamp utility change that rejects reversed bounds, with a patch and six passing assertions.
Use it to: inspect a bounded code change, its acceptance test, and retained evidence.
Included files: README.md, before.mjs, after.mjs, change.patch, test.mjs, test-results.txt.
A dated Ollama test on one existing Ubuntu laptop, including hardware, settings, observed rates, output limits, and non-tested areas.
Use it to: review an example of a bounded hardware test with explicit limitations.
Included files: README.md.
