This video presents five concise prompt ‘codes’ that improve ChatGPT’s usefulness without changing settings. The host explains how to type each code inline to trigger specific behaviors—honest critique, human tone, simple explanations, future-minded guidance, and adversarial testing—with examples and use cases. Each code is positioned as a lightweight way to reveal blind spots, risky assumptions, and clearer outputs.
- truth mode — Forces unfiltered critique: stops agreeable, generic replies and highlights weak assumptions.
- /human — Humanizes AI writing to reduce robotic, generic copy while still requiring personal review.
- Eli10 — Short for “explain like I’m 10”: simplifies complex topics into plain, easy-to-understand terms.
- future you — Instructs the model to answer as if it were your successful future self, offering actionable perspective.
- red team — Plays devil’s advocate to stress-test ideas, find blind spots, and identify risky assumptions.
Quotes:
Type ‘truth mode’ and ChatGPT will stop glazing and give you the unfiltered truth.
“/human” produces far less generic AI slop — more human, less robotic.
Use ‘red team’ to rip your idea apart and expose blind spots and risky assumptions.
Statistics
| Upload date: | 2026-08-15 |
|---|---|
| Likes: | 4302 |
| Comments: | 152 |
| Statistics updated: | 2026-09-01 |
Specification: Secret codes for ChatGPT & Claude
|
