The Canal Papers

Explains two computational-psychiatry papers on 'canalization' (the new keyword for trapped priors / mental mountains). (1) The Friston/Carhart-Harris paper: model the brain as an energy landscape where steep 'canals' = over-precise priors; over-canalization may underlie the 'general factor of psychopathology' (correlated with fewer dendritic spines). (2) 'Deep CANAL' (Juliani/Safron/Kanai): maps canalization onto deep-learning concepts - overfitting/underfitting (canalization during inference) and over-plasticity/over-stability (during training) - into a 2x2 of disorders. Scott adds clear pedagogy (the inference-vs-landscape-change distinction, illustrated by his thought-loop bad-trip anecdote) and real critique (the autism-as-overfitting-vs-undercanalization contradiction; factor-analysis skepticism; the clockwork->switchboard->computer->neural-net metaphor-history reflection).

Quality 74 · Strong. Top-of-Strong / standout explanation - makes dense computational-psychiatry legible with characteristic framings (the energy-landscape pedagogy, the metaphor-history reflection, the bad-trip illustration) and genuine critical engagement. Held below Excellent because it's primarily exposition of others' papers and Scott openly admits not fully understanding the model ('hand-wavey and imprecise... I don't fully understand it').

Claude’s paradigm shift 48 · Moderate. Moderate - a clear synthesis with some original framing (the metaphor-history point), but the substantive ideas are the papers'.

Real-world impact 2 · Minor. Within-discourse computational-psychiatry/neuroscience explainer; niche professional reach, no material effect.