Course 4 · Next Token Prediction × Knowledge × Working Memory × Steerability

Blind Spot Scan

Input: an AI response you're about to act on, plus (optionally) the prompt that produced it and two quick context flags.

Outcome: a routine audit against the five known compound-failure patterns from this course — run whether or not anything feels wrong. For anything flagged, you get a real verification prompt to run, a place to note what it actually told you, and — once you have findings — one final prompt that folds all of them into a genuine, corrected rewrite.

Useful for: the actual habit this course is teaching — most people don't catch AI mistakes because there's no felt symptom, not because they aren't paying attention. This runs the check regardless.

This is a heuristic scan for known danger shapes (specific facts stated confidently, agreement with an unverified premise, long unbroken reasoning chains, mixed sources, long conversations) — not a verdict on whether the answer is actually wrong. Every flagged item includes a real way to verify it, because judging correctness itself is not something this tool can fake.

Paste the AI's response you want to audit Paste the prompt/question that produced it (optional, improves the audit)

This is part of a long, multi-turn conversation

I supplied a document/source alongside my prompt