Reason backward from an outcome you want to exist — it manufactures originality that absorbed problems can't

Vivek contrasts two ways to get a problem. Most people absorb one — “from an advisor, from whatever a big lab announced last quarter, from the paper everyone is quote-tweeting this week” — and “the trouble with an absorbed problem is that you hold the conclusion without the reasoning… when they pivot, you find out a year later.” Worse, a fashionable problem means “you’re racing a thousand people who started earlier and have more compute than you.” John Schulman’s guide splits the work into reading the literature to hunt for improvements versus choosing “an outcome you genuinely want to exist and reason backwards to the experiments”; the second “manufactures originality” because “a goal you actually care about will drag you into territory no survey paper covers.” (Hamming’s lunch question — what are the important problems, and why aren’t you working on them — stings precisely because most people can’t answer it.)

This is a research-craft instance of First-principles thinking is uncomfortable because it transfers responsibility — analogy outsources blame to 'best practices' — reasoning backward from the outcome forces ownership, where absorbing a problem outsources it. The “crowded racetrack” warning is why Peter Thiel's question is a detector for actual first-principles thinking — if your conclusions match the crowd, you're analogizing: a problem everyone is already quote-tweeting is, by construction, an analogy-driven consensus pick, and Reasoning by analogy has a ceiling — you can never get beyond what already exists by copying what already exists caps what it can produce. Manufacturing originality this way starts with Shared inputs produce shared conclusions worth nothing — old and cross-disciplinary material is criminally underpriced — the goal you care about only pulls you off the beaten path if your inputs aren’t the same as everyone else’s.