The actual working material.

Seven sections, no padding.

• 01 The Prompt Stack — full methodology. The four stages worked through with one real example end to end. The version I actually run, not the simplified one in posts.
• 02 Ten prompt templates. Copy-paste blocks for: position review, downside hunt, sector scan, earnings prep, RNS triage, valuation sanity check, competitor map, thesis stress test, exit checklist, post-mortem.
• 03 Decision checklists. Pre-trade, post-trade, monthly portfolio review, and the one that gets used most: "should I act on this AI output, or is it just well-worded noise?"
• 04 The honest limits page. What AI is bad at for investing — numbers, recency, long-filing context, anything time-sensitive. The page that stops you from making the obvious mistakes.
• 05 Tool-switching guide. One-page decision tree for when to use Claude vs ChatGPT vs Perplexity. The honest comparison, not the marketing one.
• 06 Glossary. Plain-English definitions of the terms used across the site — RNS, AIM, adversarial prompting, the Prompt Stack vocabulary.
• 07 Further reading. The handful of sources I actually trust — books, papers, and other one-person sites that aren't full of noise.

Investors who already know AI is half-useful.

If you've used Claude or ChatGPT on a real investment question and come away thinking that was kind of useful but I don't trust it — this is for you. The Lab Book is the result of figuring out which half is which.

It is not a beginner's guide to investing or to AI. It assumes you've been doing both for a while and want to use them together more carefully than the average tutorial suggests.

No tips. No tickers.

The Lab Book contains zero stock recommendations and zero positions. It's about the method, not the trades. If you want a list of stocks to buy, this is the wrong document and the wrong site. If you want a sharper way of thinking about whatever you're already doing — keep reading.