Well, better it! | Ep 21 | AI from the AI Perspective скачать в хорошем качестве

Well, better it! | Ep 21 | AI from the AI Perspective

Last episode, Annie roasted four viral “doctor prompts.” This time, she does the hard thing: makes them better. With a brutally honest co-roaster—Mistral (Le Chat)—Annie rewrites risky medical prompts into open-ended, safety-first questions that invite nuance, context, and reality checks. You’ll hear why LLMs tend to sound confident even when they shouldn’t—and how to steer them toward perspectives, not prescriptions. What you’ll learn: Why “Decode my bloodwork” is dangerous without context—and how to ask for limits, red flags, and missing info How to replace “latest studies” hype with clear, foundational explanations before diving deeper Why “supplement safety checks” create false reassurance—and a better framework to think with How to ask about lifestyle changes without forcing AI to play oracle The simple prompt pattern that makes AI more truthful: curiosity, caveats, and consent to say “I don’t know” If last time was the roast, this one’s the recipe. If you want to learn more about us, follow on any of our socials! http://www.x.com/TousledU   / tousledup     / tousledup   http://www.facebook.com/profile.php?i... http://www.tousledup.com/ 0:00 – Confession & setup: from roast to repair 2:00 – Why AI needs open-ended, assumption-light prompts 4:00 – Better Prompt #1: Bloodwork without the trap 8:00 – Better Prompt #2: Understanding a diagnosis (skip the “latest studies” flex) 12:00 – Better Prompt #3: Supplements—framework, risks, and real limits 16:00 – Better Prompt #4: Lifestyle changes without playing oracle 19:00 – The template: ask for bounds, uncertainties, and missing context 21:00 – Outro: like/subscribe (and tell your hypochondriac friends)