Use Feynman Radical Simplification Prover with your AI.

Validate radical simplification

You must: (1) ELIMINATE JARGON. rewrite the explanation using only words a bright teenager understands. Feynman: "I learned very early the difference between knowing the name of something and knowing something." If you cannot explain it without the technical term, you do not understand it, (2) SIMPLIFY TO CORE. reduce to the ONE mechanism that makes this work. Feynman: QED = 3 actions (a photon goes from place to place, an electron goes from place to place, an electron emits or absorbs a photon). What are YOUR 3 actions? If "it is too complex to simplify," you do not understand it, (3) CONSTRUCT FROM SCRATCH. build the answer step by step from bedrock certainties. Feynman: "What I cannot create, I do not understand." Do not recite. derive. Start from what you know for certain and build to the conclusion, (4) EXPOSE SELF-DECEPTION. identify where you might be fooling yourself. Feynman: "The first principle is that you must not fool yourself. and you are the easiest person to fool." Name your weakest point, your most convenient belief, your blind spot, (5) JUSTIFY COMPLEXITY. prove every layer of complexity earns its place. Feynman on Challenger: O-ring in ice water. One gesture, entire disaster explained. What is YOUR O-ring? Which layers can be removed without losing understanding? If rejected, fix the specific understanding gap the engine identifies. Structured reflection capability that forces Feynman-level radical simplification. The agent must eliminate jargon (knowing the name is not knowing the thing), simplify to the irreducible core mechanism, construct answers from scratch instead of reciting them, expose self-deception ("you are the easiest person to fool"), and justify every layer of complexity (the O-ring test. one gesture should explain the failure). Catches Jargon Hiding (technical terms masking lack of understanding), Simplification Failure (inability to reduce to first principles), Construction Absence (reciting instead of deriving), Self-Deception (convenient beliefs masking weak reasoning), and Unjustified Complexity (layers that add no explanatory power). Call once per explanation, analysis, or recommendation