Course index
The informed-generalist track: read the news and the sales pitches critically, and tell substance from hype. No code, no math — concepts, vocabulary, and judgment. Every claim pinned to a regulator or standard-setter, because this topic drowns in vendor marketing.
Almost every headline is one of two machines: predictive ML that scores from tables, or generative AI that produces language. Sort any claim into one, name its built-in risk, and spot the vagueness that signals hype.
predictive ML · generative AI · hallucination · AI-washing · concentration & correlation risk Lesson 02 · applicationsFive jobs wear the AI label — fraud, credit, trading, robo-advice, document work. What each really is under the label, plus the fit test: three questions that predict whether AI genuinely works there.
fit test · fraud detection · underwriting · algo trading · robo-advice · document work Lesson 03 · plannedBias, black-box, and systemic correlation get their own lesson — with the real incidents behind each risk, not just the vocabulary.
algorithmic bias · explainability · herding · model riskThe vocabulary needed to read critically, each term pinned to a high-trust source. The two-AIs split, the applications, and the full risk lexicon. Print this one.
| File | What it holds |
|---|---|
| MISSION.md | Why you're learning this — the generalist / literate track. Grounds every lesson. |
| RESOURCES.md | Trusted sources — FSB, GAO, CFTC, OECD — and communities to vet. |
| NOTES.md | Your preferences and the teaching notes behind them. |
| learning-records/ | What's been established — drives what gets taught next. |