Brain dumps, voice capture, and what on-device AI changes — written by the one person who builds the app. RSS
The ADHD tax is the real money and time lost to executive function failures: late fees, expired groceries, missed renewals, rush shipping on things bought twice. Where it comes from — working memory, time blindness, avoidance — and which parts of it planning can and cannot reduce.
A new planning system works brilliantly, then dies — around two weeks in, people usually say, though nobody has measured it. The cause is not discipline: it is novelty running out and leaving three structural properties exposed. What the system charges you daily, what it does when you skip a day, and whether partial use pays anything.
Planning the night before buys a decided morning and protects you from starting the day in reaction mode; planning in the morning buys accuracy about the day you actually woke up into. Neither has a research verdict. The honest answer is a split — decide the shape at night, confirm it awake — and the deciding factor is which failure hurts you more.
Laziness is not wanting to do the thing. Executive dysfunction is wanting to, knowing how, and still not starting — a gap between intention and action that effort alone does not close. What the difference looks like from inside, and which supports actually address it.
Paper wins on friction-free starting, freedom from notifications, and memory of the act of writing. Digital wins on capture speed anywhere, reminders that find you, and search. The honest answer is that the medium matters less than three properties either medium can have — and most setups in both camps lack.
A reminder you set for yourself should be the easiest one to obey, and it is usually the first one you swipe away. The mechanics of alert fatigue: repetition teaches your brain that alerts are safe to ignore, badly timed alerts arrive when acting is impossible, and vague ones name no action. What a reminder has to be to survive.
Folders make you decide where a thing lives before you can put it down — the exact decision an overwhelmed brain defers. Tags let you capture first and label later, if at all. Why one tag per item, a cap of twelve, and colour you never pick beat a filing system.
The brain dump works; the hour of sorting afterwards is where it dies. A four-step loop — say it, tag it, arrange it, then plan a day of 3–5 — that turns a spoken mess into a finishable day with no folders, no boards, and no system to maintain.
When a to-do list makes you anxious, the instinct is to avoid it — which makes it longer, which makes it worse. What research on open loops, attention residue and the planning fallacy says about why lists turn threatening, and the fixes that address the mechanism instead of the mood.
ADHD working memory is measurably smaller than the demands a day makes of it — meta-analyses in children and adults agree. The fix is not remembering harder but cognitive offloading: moving the remembering into the world. What the offloading research shows, including its one real catch.
Knowing exactly what to do and still not starting is not laziness — initiation is its own executive function, and it fails on predictable inputs: aversiveness, distance from reward, and a first step too big to be an action. What procrastination research actually shows, and the three levers that move it.
Multiple lists are how people actually live — a project, a trip, the groceries. They are also how to-do apps die: contexts multiply, items scatter, and the day fragments. How Stowe 1.9 does separate lists with one shared day, a visible rail, and a cap of six.
Streaks work until they break, and for a lot of people — ADHD brains especially — the break is the whole memory. In 1.3 Stowe replaced its streak card with a run that counts days planned, best and held (never missed), a shelf of things Toki keeps, and days off. Here is the reasoning.
Time blindness is a weakened felt sense of time passing — "now" and "not now" with little in between. What the ADHD time-perception research shows, and the external supports that compensate: visible time, anchored plans, and days sized to reality.
Body doubling is working on your own task in the presence of another person — live or remote, silent or not. Why presence works as borrowed executive function, what the research actually shows, and how to try it without an app.
A five-step daily planning method built for brains that resist planning: empty your head, sort what came out, pick 3–5 things, phrase the first as an if-then, park the rest out of sight. Each step exists because research says it does work the intention alone will not.
Most AI planning apps send your tasks to a server to be processed; on-device AI keeps them on your phone. What each architecture means for privacy, the five questions that reveal which one an app uses, and where the honest trade-offs are.
Three to five meaningful tasks is the defensible answer — not because a magic number exists, but because planning-fallacy and open-loop research both point at capped, finishable days. Why popular systems converge on small numbers, and how to pick yours.
To-do lists fail for reasons psychologists have measured: we underestimate how long things take, intentions alone convert to action about half the time, and long lists carry a cost. The research, and what actually fixes it.
Research suggests external planning supports — written plans, if-then intentions, visible time — help with ADHD precisely because they move executive function out of the head. The studies, honestly summarized.
Task paralysis is the freeze that hits when a long list offers too many entry points and no obvious first move. What choice research actually shows — including the replication debate — and the case for one next action.
An honest evidence review of popular planning methods: if-then planning has a large research base, time management skills have trial support, body doubling is promising but early, and several famous rules are folklore.
A brain dump is emptying everything on your mind into one place, without sorting as you go. The dump is the easy half — here is the method for what comes after.
Voice to-do apps used to transcribe what you said and leave the sorting to you. On-device AI changes the job: speech becomes structured tasks without leaving the phone.
An honest comparison of three planners: Things wins on polish and one-time pricing, Structured on timeline planning, Stowe on voice-first capture with on-device AI.
Stowe transcribes your speech and structures it into a plan entirely on the iPhone, so every feature works in airplane mode. Here is the pipeline, in plain terms.
A reference table of iPhone models, their RAM, and what local AI each can run — why 8GB is the Apple Intelligence line, what changed at 12GB, and what model sizes fit in each tier.
Stowe