It’s Bayes All The Way Up

A lucid synthesis of the Bayesian/predictive-coding account of perception, working from Corlett-Frith-Fletcher (2009) and Lawson-Rees-Friston (2014). Perception as a 'handshake' between top-down priors and bottom-up sense-data, with neurotransmitters mapped onto Bayesian terms (AMPA/NMDA glutamate, dopamine as precision). Uses the framework to explain schizophrenia (attenuated priors / delusions of reference), the ketamine and LSD profiles, and autism as aberrant precision (narrow confidence intervals → stimming, sensory sensitivity, social difficulty). Section III is characteristically self-critical, flagging the model's unresolved tensions. Explicitly 'very speculative' and a precursor/companion to the later Surfing Uncertainty review.

Quality 73 · Strong. 73 — high Strong. A memorable, standout explanation of a genuinely hard neuroscience topic that helped popularise predictive processing in the rationalist sphere; readers point to it. Held below Excellent because it is an acknowledged speculative synthesis of two papers with Scott flagging his own confusions, and the definitive predictive-processing post is the later Surfing Uncertainty review.

Claude’s paradigm shift 59 · Moderate. 59 — Notable shift, just short of Major. The Bayesian-brain frame was fresh and non-obvious to his 2016 audience and clearly updated informed readers, but it builds on established Friston/free-energy neuroscience rather than originating the idea.

Real-world impact 3 · Moderate. 3 — moderate. Predictive processing became durable vocabulary within the rationalist/psychiatry-adjacent subculture, partly via posts like this; niche professional-sphere influence, no material-world change.