# Becoming perceptive (by Henrik Karlsson) ## A concise summary Source: “Becoming perceptive” — https://www.henrikkarlsson.xyz/p/perceptive Related pieces: “Unfolding” (part 1) — https://www.henrikkarlsson.xyz/p/unfolding ## Core thesis - Perceptiveness is central to self‑actualization (Maslow): self‑actualized people live more in immediate reality and notice concrete details rather than labeling or abstracting. - Perception is a “controlled hallucination”: the brain predicts sensory input and updates those predictions only when prediction error is registered. If prediction error is small, the brain filters reality and you mostly perceive your model, not what’s actually there. - To be truly perceptive you must both build rich models and be willing/able to “see through” them — i.e., detect when your models are wrong. Links referenced: - Maslow on self‑actualization (PDF): https://www.holybooks.com/wp-content/uploads/Motivation-and-Personality-Maslow.pdf - On “seeing through” mental models: https://gwern.net/unseeing ## Mechanism (quick) - Brain predicts input → constructs a perceptual hypothesis (a controlled hallucination). - Incoming sensory data is compared to predictions; only large enough errors update the model. - Sensitivity to prediction errors varies by person and can be modulated by context, training, and state of mind. ## Practical ways to become more perceptive (actionable) - Assume you’re wrong: adopt a mental prompt like “There is something I don’t understand here—find it.” This increases attention to anomalies and odd details. - Ask “stupid” or naive questions: they surface hidden assumptions and reveal missed information. - Create tight feedback loops that confront your models with reality: - Externalize by drawing or sketching: forces you to admit what you don’t actually see. - Write to think: make interpretations explicit so they can be corrected (see Henrik’s “How to think in writing”). - Paraphrase others’ meaning aloud and invite correction. - Record predictions (about outcomes) and review them later. - Undertake real projects that produce fast, meaningful feedback (startups, experiments, competitive play). - Learn from messy, complex exposure rather than only sequenced curricula: accelerated training (military example) uses complex simulations to break early oversimplified models and improve pattern matching. References: - Writing to think: https://www.henrikkarlsson.xyz/p/writing-to-think - Accelerated expertise / non‑sequential training: https://commoncog.com/accelerated-expertise/ ## Pitfalls / obstacles - Knowledge shields: learning via neat, sequential examples can make you blind to subtle, complex realities. - Comfort bias: the brain prefers predictable inputs; avoiding surprise preserves comfort but reduces perceptiveness. - Individual differences: some people naturally register smaller prediction errors; others can cultivate sensitivity with practice. ## Practical habit checklist (useable in Obsidian) - Daily: note one assumption you’re making about a situation; list what would disconfirm it. - Weekly: write 1–3 concrete predictions about projects/relationships; review outcomes. - Regularly: practice “explain what I hear” with interlocutors and request corrections. - Periodically: take on short, real-world projects outside curricula to expose yourself to messy feedback. ## Related reading (from the author) - Unfolding (part 1): https://www.henrikkarlsson.xyz/p/unfolding - Being patient with problems: https://www.henrikkarlsson.xyz/p/being-patient-with-problems - Cultivating a state where new ideas are born: https://www.henrikkarlsson.xyz/p/good-ideas Notes - The phrase “controlled hallucination” is credited to neuroscientist Anil Seth (also used by Andy Clark).