Get Andrew Ng's Prompting Course Cheat Sheet
Andrew Ng's 3-hour AI Prompting Course on DeepLearning.AI — compressed into a single PDF you can pin to your second monitor. The 9 rules from the course, top to bottom: empathy-for-the-AI framing (treat it like a smart fresh college grad who does not know you yet), the 3 information paths (pretrained → web search → deep research) with the bombshell that the model never actually reads the pages it cites, beating sycophancy with neutral framing and rubric-before-score (subjective vs objective grading differs by ~20 points on the same work), killing AI slop with progressive outlining and the named fingerprint patterns (em-dashes, "delve", "it's not X — it's Y", lists of three), Andrew's retirement of "think step by step" in favor of "think hard"/"ultrathink" plus the METR task-horizon curve, the brainstorm-loop move that beats more-context-up-front (feedback on options surfaces what context actually matters), the 750K-word context window and why new chat per topic still matters, the desktop-app footgun (delete files without recycle bin, no edit history), and AI seeing images coarsely. Plus two paste-ready prompt scaffolds — an anti-sycophancy rubric prompt that forces binary criteria and sum-at-the-end scoring, and a brainstorm loop prompt that asks for 4 markedly different options before any feedback step. A stat sheet covering the six numbers Andrew anchors the course on. And a direct link to the full 3-hour course so you can go watch Andrew teach it himself.
Free. No spam. Unsubscribe anytime.