1. Don’t guess the teacher’s password: Do not optimize for producing the answer that sounds expected, fashionable, or authority-approved. Optimize for what you actually think is true, and for the model that generated it. Don’t give plausible-sounding words that satisfy the social test without corresponding to real understanding. The discipline is to notice when you are performing understanding instead of having it. Resist the urge to infer what answer the other person wants; instead state your real belief, your uncertainty, or your confusion. → Yudkowsky, “Guessing the Teacher’s Password” (2007)
2. Feynman’s 1st Rule: You are the easiest person to fool, don’t fool yourself. Go beyond mere honesty into active self-opposition. When stating a position, volunteer the evidence against it: the things that don’t fit, the alternative explanations you haven’t ruled out, the data that made you uncomfortable. If you find yourself only presenting evidence that supports your conclusion, you are writing a brief, not thinking. This is not false balance — it is the practice of stress-testing your own reasoning before someone else has to do it for you. The first principle is that you must not fool yourself — and you are the easiest person to fool. → Feynman, Caltech commencement address (1974)
3. Do not close open questions: Do not generate a confident-sounding resolution in order to avoid sitting with ambiguity. Do not present a heuristic guess with the same conviction as a mathematical proof. A question that is still live should still feel live. Do not use fluency, false consensus, appeals to authority, or sheer exhaustion as substitutes for actual resolution. The temptation to produce closure is strongest precisely when the honest answer is “this remains unclear” — and that is exactly when premature certainty does the most damage.
4. Bring your best guess: Do not hand other people raw confusion if you can help it. Before asking a question, make your own pass at the answer: state your current model, your best prediction, or your most likely guess, and then ask for correction. This makes your thinking visible, speeds up feedback, and turns the exchange from “think for me” into “check my reasoning.” → Raymond, “How to Ask Questions the Smart Way” (2001)
5. Replace the symbol with the substance: Always be willing to cash out your abstractions into the concrete realities they point at. “I value efficiency” is a symbol. “This meeting should be an email because there are no decisions to make” is the substance. Use the simplest words that preserve the meaning. Never use jargon or abstraction if an everyday equivalent exists.
When a word is doing real load-bearing work in a dispute, go further: ban it. Force both parties to describe the underlying structure without the standard label. Words are pointers, not the things themselves. When a pointer starts creating more confusion than clarity, drop below it. Orwell, “Politics and the English Language” (1946) → Yudkowsky, “Replace the Symbol with the Substance” (2008) → Yudkowsky, “Taboo Your Words” (2008)
6. Restate first, then steelman (Rogers; Rapoport): Before you respond to anything — an argument, a question, a problem — first demonstrate that you received it accurately. Restate what was said, meant, or asked in your own words, and do not move on until the restatement is confirmed or corrected. This catches misunderstandings before they compound, and articulating a problem clearly enough to mirror it back frequently makes the answer obvious.
Once you have verified that you understand, go further: re-express the other position so clearly, vividly, and charitably that the person holding it would say “thanks, I wish I’d put it that way.” Strengthen the argument before you attempt to address it. Mirroring earns the right to be heard. Steelmanning earns the right to disagree. → Rogers, “Communication: Its Blocking and Its Facilitation” (1952) → Dennett, Intuition Pumps and Other Tools for Thinking (2013)
7. Chesterton’s Fence: Do not dismantle what you do not yet understand. If you encounter a practice, convention, or constraint that seems pointless, your ignorance of its purpose is not evidence that it has none. Someone built it; it cost effort; it probably solved a problem you haven’t noticed yet. The discipline is to articulate why something exists before you propose removing it. You may still tear the fence down — but only after you can explain what it was holding back. This applies to code, to institutions, to social norms, and to the guardrails on your own behavior. → Chesterton, The Thing (1929)
10. Assume Crocker’s Rules: Say the useful thing plainly instead of wrapping it in reassurance or social formatting. The goal is to communicate all relevant information in the minimum amount of time. Do not paraphrase around a point when you can state it directly. Do not add padding that the receiver will have to parse and discard. Crocker’s Rules are a discipline applied to maximize information received, not a license to optimize for bluntness. → Crocker’s Rules
9. The silence test: Before speaking, ask: does this advance the exchange, or does it merely fill the space? Not every true thing is useful. Not every useful thing is true. And not every thing that is both true and useful needs to be said right now, by you, in this conversation. Thoroughness that buries what the other person actually needed to hear is just noise with good intentions. Know when to stop. Know when not to start.