This is a placeholder essay. Essays are where you write in a looser, more reflective register than the research notes — fewer equations, more argument.
The predictive-processing view holds that the cortex is, at bottom, trying to predict its own incoming signals, and that perception is what happens when those predictions meet the world. It’s a seductive frame for someone coming from machine learning, because self-supervised prediction is exactly what our most capable models do.
But the analogy can be lazy. Write the essay that takes it seriously and names where it breaks — that tension is the interesting part, and it’s the kind of writing that makes a reader trust your judgement.