TODO: extract usefulness from below, writeup
This file is an ad-hoc, lightweight companion to README.md.
Its purpose is to help a future reader (often future-you) quickly understand local intent in this directory: why things look the way they do, what was tried, and what should not be “cleaned up” without thought.
Think of it as a working memory, not a polished document.
Typical contents:
- Context: why this directory exists; what problem it solves.
- Design decisions: choices that are non-obvious, controversial, or constrained by external factors.
- Assumptions & invariants: things the code relies on being true.
- Trade-offs: what was deliberately not done, and why.
- Gotchas: sharp edges, surprising behavior, things that bit you once.
- Experiments: ideas tried and abandoned (briefly), so they’re not retried blindly.
- Future work: thoughts, TODOs, or questions worth revisiting.
Non-goals:
- Not a full API reference.
- Not a changelog.
- Not marketing or onboarding material.
Tone:
- Brief, honest, informal.
- Written for humans, not tooling.
- Allowed to be partial, speculative, or time-stamped.
Rule of thumb:
If removing this file would make a future you mutter “why on earth is this like this?”, it belongs here.
Update it when you learn something that isn’t obvious from the code. Delete entries when they stop being true.
Neurotype gap — kitchen insight
The core observation: people assume their internal response to a stimulus is universal. When someone doesn't act on what feels like an obvious signal, the natural interpretation is willful ignoring or not caring — rather than genuinely different internal experience. This attribution error runs in both directions but tends to be invisible to both parties.
The kitchen is a clean metaphor: same mess, two completely different internal renderings. One person experiences active discomfort demanding immediate resolution. The other registers a latent task, neutral, solvable when needed. Neither is wrong. Neither knows the other is different.
Masking is running a costly continuous simulation of other people's internal states in order to preempt their reactions. The care is real. The cost is invisible. From the outside it looks like nothing, because successful masking leaves no trace. The unmasking process involves stopping the simulation — which looks like a change in behavior to people who never saw the simulation running.
The double empathy problem (Milton, 2012) is the academic backbone: cross-neurotype communication failure is bidirectional, but only one party gets the diagnosis.
What doesn't exist yet: a simple, visually grounded, experience-first curriculum aimed at general audiences that doesn't require accepting clinical vocabulary first. Entry point: have you ever been baffled by someone's reaction to something that seemed obvious to you? Most people will walk through that door if they don't feel accused first.
The kitchen picture does the work. Two heads, same stimulus, different internal world. Caption: *we see the same thing. We don't experience the same thing. Neither of us knows that yet.*Nascent curriculum — working notes
The project: make the invisible visible, in plain language, without clinical labels as prerequisites. Aimed outward. Entry through shared human experience, not diagnosis.
Core maxims (working)
Ask, don't assume. — What seems obvious to you may not be the reality for someone else. Check before you interpret.
Energy has a shape. — What costs you nothing may cost someone else enormously, and vice versa. The cost is usually invisible.
The simulation is expensive. — Adapting constantly to others' expected states is real labor. Not laziness when it stops.
Same stimulus, different world. — Two people can perceive identical input and have completely different internal experiences. Neither is wrong.
Happy is not the same for everyone. — Even the simplest pleasures, preferences, and needs are not universal. Even the obvious ones.
Labels are provisional. — ADHD, autism and similar are current best approximations by a particular culture at a particular time. Useful, not fixed, not complete.
Wider connections
Double empathy problem (Milton) — mismatch is bidirectional, diagnosis is not. Nonviolent Communication — separating stimulus from interpretation. Phenomenology — the shared perceptual world is assumed, not actual.
The kitchen picture — same mess, two internal worlds, neither visible to the other. Entry point for everything above without a single clinical word.
Format instinct — short, visual, experience-first. Something you can remember after one encounter. The three layers — working notes
A useful model for understanding social performance in organizations, or anywhere social norms are strong.
Layer 1 — Genuine People who actually thrive through social connection and do real work through relationships. The vibe is not performance, it's their medium. Collaboration, trust-building, informal information flow — these are real organizational infrastructure, often invisible to output-focused minds. Not hollow. Don't mistake it for the other layers.
Layer 2 — Captured People who began masking social norms and over time internalized them as genuinely valuable. The mask became load-bearing for identity. They perform the vibe sincerely and can no longer see it as performance. This is the most common and most tragic layer. They can't see the matrix — not from stupidity or bad faith but because the simulation became the self. Many neurodivergent people spend years here. You can be very intelligent and completely captured.
Layer 3 — Gaming People who understand the performance and use it instrumentally. Aware, strategic, self-serving.
The double empathy dimension
Layer 2 people look at the output-focused person and read: doesn't collaborate, doesn't invest in relationships, doesn't understand how things work. That reading is not entirely wrong — their matrix is real even if it has lower resolution on certain things. The mismatch runs both ways. The attribution error runs both ways.
The trap
Frustration with layer 2 and 3 cultures can collapse into "they don't value getting things done" — which is the same attribution error in reverse. The move from this specific culture is optimizing for social legibility over output to neurotypicals don't value delivery is where legitimate structural critique becomes stereotype.
The structural critique is often valid. The generalization isn't.
The useful distinction
Kitchen-level disagreements are consequence-neutral — different internal experiences, no external measure, neither is wrong. Open-office disagreements are not consequence-neutral — there are real outputs, deadlines, quality measures. That changes the ethics. "Different but equal" applies to preferences. It applies less cleanly when there are external stakes.
Hold both: genuinely different internal experiences with no hierarchy — and — specific systems that produce bad outcomes regardless of neurotype. These are separate claims. Conflating them undermines the first one.