I think that we don't achieve stuttering remission due to the "problem" of belief updating and error minimization. So: I think that our subconscious does this completely wrong, because it's designed to minimize these errors through: (1) belief updating based on errors, and (2) action (modifying the environment to confirm predictions). In other words, as Usler calls it, (1) being reliant on outdated and imprecise predictions, and (2) precise predictions becoming too sensitive, can result in an inability to optimally update prior beliefs and inaccurate predictions. So: If we speak fluently alone, but we add a single person. Let's say that our subconscious starts fearing people's judgements, then uncertainty may occur. Meaning: a feeling that our speech-related predictions are unable to reliably minimize prediction error through perception and action. And then ultimately, it results in a defensive behavior involving the sudden stopping of speech movement to a perceived threat, and thus, we stutter. So: I think stutterers need to re-learn this, when aiming towards stuttering remission. Re-learn action-perception cycling. Because right now, I think that fluency in stutterers often occurs *by luck or coincidence*. For example in some stutterers, "reducing loud background noise in the shopping mall" decreases the precision of sensory feedback, while in others it increases it. Depending on how we interpret such stimuli associated with speech production (i.e., depending on our evaluation of the severity of the perceived conflict). *To clarify our brain's system further:* The brain functions as a hierarchical generative model consisting of prior beliefs P(x) and likelihood functions P(y|x) for the generation of updated (i.e., posterior) beliefs P(x|y) based on incoming sensory observations P(y). This hierarchy includes: Lower-level predictions: "prediction of what word you will hear next". Higher-level predictions: “prediction of self as an effective communicator,” including agentic control and action sequencing across longer timescales like sentences. Prior precision = confidence in prior beliefs. Sensory precision = confidence (or faith) in the incoming sensory data. Stuttering may emerge from 1) imprecise prior beliefs of sensory input associated with speech production; and 2) a precipitating inability to attenuate sensory precision during speech. Conclusion: So, I think we need to re-learn this action-perception cycling to succeed in stuttering remission, wouldn't you say so?