Neurotypica

The brain compresses information by grouping features into units and sequences into routines. Perceptual grouping follows bottom-up regularities---things that are similar, close together, or form closed contours get bound into units automatically. With practice, multi-step actions compress into single fluent routines. What started as deliberate steps becomes a chunk that fires as a unit.

This matters because changing what groups together---by tweaking inputs, shifting attention, or inserting edits at boundaries---causes downstream scripts to reassemble differently. The "arrive home, shoes off, pantry" sequence feels unstoppable because it's chunked. To change it, insert your edit at a natural boundary where one sub-chunk hands off to the next.

Key takeaway: chunks form through repetition across space and time; change groupings or insert edits at boundaries to reshape behaviour.

Task sets are the top-down configurations that tell the system which chunks are currently relevant. When you shift into 'making tea' mode, the brain loads a transient set of features, rules, and responses---what to attend to, what actions are on the menu. That set selects which chunks are "in play" for this episode and suppresses the ones that aren't.

The trouble is, if you don't preload the intended set at the cue, the system falls back on whatever set was active last or whatever the situation most strongly suggests. To get the right chunks to bind, you need the right set loaded before the cue arrives.

Key takeaway: preload the intended set at the cue so the right chunks bind; minimise unnecessary set switches.

Working memory is where the brain binds the few items that matter right now---the current goal, the next step, the constraint you're juggling. It's powerful but costly, so you don't want to use it to run every step of a routine. Instead, use it briefly to set up the bindings---load the right goal, the right cue, the right constraint---and then let the trained chunks carry execution.

Think of working memory as the place where you set the stage, not where the whole play runs.

Key takeaway: spend control briefly to set up the bindings; let trained chunks carry execution.

The brain's control systems are layered: higher levels hold the goal and bias which chunks should link and in what order; lower levels execute the learned transitions. This hierarchy is what lets you flexibly recombine familiar sub-routines to serve new aims.

When you're trying to change behaviour, it matters which level you adjust. Sometimes the problem is the goal (high level), sometimes it's the specific link between two chunks (low level). Clarifying the goal can reshape which links assemble; inserting a new sub-chunk can redirect the sequence without changing the goal at all.

Key takeaway: adjust at the right level---clarify the goal to reshape which links assemble.

Recall depends on cues matching how a script was originally encoded. The right context makes the intended chunk pop into place automatically; the wrong context keeps the old script dominant, no matter how much you 'know' the new one.

This is why practicing a new routine in the exact situation where you'll need it is so much more effective than practicing it somewhere else. You're not just training the routine; you're training the link between the cue and the routine. Build environmental or procedural cues into the context, and the right chunk will fire when you need it.

Key takeaway: build environmental/procedural cues into the exact context where you want the new link to fire.

Chunks don't just form through individual experience; your brain also maps the social world. You learn patterns by watching what co-occurs, and communities coordinate which things co-occur. When everyone in your social world treats 'distrust institutions' as naturally paired with 'value self-sufficiency' and 'prepare for collapse,' your brain maps that cluster automatically. You didn't reason your way into it; you absorbed it through exposure to social regularities.

This is why belief systems stack in communities. Preppers share theory stacks (collapse, self-sufficiency, weapons). Woo-woo folks share different ones (alternative medicine, energy work, manifestation). Scientism adherents share yet another (trust institutions, evidence-based medicine, mainstream consensus). None of these stacks are purely logical---they're socially mapped chunks. The beliefs travel together because the community coordinates them, and your brain learns that coordination as environmental regularity.

Key takeaway: communities coordinate patterns through shared practices; individual brains map those patterns automatically; to change socially chunked beliefs, change the coordination.