Human Cognition Evidence layer for scientific AI Β· Updated November 18, 2026
Humans reliably hold only a small number of items in active use at once, with classic estimates near four items β strong evidence, contested exact count
People miss clearly visible events when attention is directed elsewhere, as in the invisible gorilla studies β well replicated
Fast intuitive and slow deliberate thinking describe real behavioral differences, though the strict two-system framing remains debated β useful model, mechanism debated
Systematic errors such as anchoring and confirmation bias appear across many lab and field settings, but their real-world size varies by context β well replicated, effect sizes vary
Recall rebuilds memories rather than replaying them, so suggestions can insert plausible details people confidently report β well replicated
Sleep after learning improves later recall for much of the material, with effects strongest for certain memory types β strong evidence, boundaries unclear
The adult brain continues to rewire with practice and learning, though the pace and scope of change decline with age β well replicated, limits under study
People can monitor the accuracy of their own judgments, yet confidence often tracks feelings of fluency more than actual correctness β solid findings, individual variation large
Large collaborative replications confirmed many classic cognitive effects but failed to reproduce others, often with smaller effect sizes β documented across multiple replication projects
Know what changed, what holds up, and what remains uncertain. Every Friday. No ads.