The AI Learning Hub Journal

The Not-List

Everything cut to make the weekend possible comes back, wearing better argumentsthe answer is a written list of what version one refuses to do — not now, with a review date, never a neverTHE NOT-LIST — EVERY ITEM EARNS ITS PLACE THE SAME WAYVERSION ONE REFUSESBY HAND INSTEADWHAT IT COSTSTHE DATEonline paymentinvoices sent by hand, exactly as today20 min a weeklook again: month 3automatic remindersa calendar note and a text sent by thumb10 min a weeklook again: month 3a loyalty schemea card and a stampmomentslook again: month 6an appthe same page, opened in the phone’s browsernone at alllook again: month 6the wish-list is not wrong — reminders would genuinely help this business. it is early: version one has one job, becoming the way bookings are takenTHE TEST THAT KEEPS A NOT-LIST HONESTno workable by-hand version? then it may notbelong on the list — it may be version one’s real jobif the workaround truly costs minutes, building itnow buys almost nothing, and pays in ledger daysby-hand is the cheapest way to learn a feature’s worthWHAT EVERY YES ACTUALLY COSTSan afternoon to build — that part really is cheap nowbut everything you ship must then be run, supportedand maintained: it can break, confuse a customer, andturn up in the late-night email, in its second yeara little of every week afterwards, indefinitely — a subscriptionthe AI makes reminders an afternoon’s work, which makes yes feel free — every yes is bought with days from the ledgerNOBODY REGRETS A SMALL VERSION ONE THAT WORKS people regret large ones that almost do — refusing things you could easily build feels wrong at first, and it is the skill
A not-list is not a confession of weakness — each item is a decision with a date, and every argument becomes a scheduling question.

A Decision with a Date

The moment the project gets serious, the full wish-list comes back. Online payment, automatic reminders, a loyalty scheme, an app — everything that was cut to make the weekend possible returns, wearing better arguments. The answer is a not-list: a written list of what version one refuses to do. Notice the word refuses. A not-list is not a confession of weakness, and it is not forever. Each item is a decision with a date: not now, reviewed once version one has run for a while. That framing matters, because it turns every future feature argument into a scheduling question instead of a fight. The wish-list is not wrong — reminders would genuinely help this business. It is early. Version one has one job: become the way the business takes bookings. Everything that does not serve that job waits, in writing, with a review date attached.

  • The wish-list returns the moment the project gets serious
  • A not-list is what version one refuses to do, in writing
  • Each item is a decision with a date, not a never
  • Feature arguments become scheduling questions instead of fights

The By-Hand Workaround

Every item on the not-list earns its place the same way: it has a by-hand workaround that costs minutes a week and zero days of building. Online payment? The groomer keeps sending invoices by hand, exactly as she does today — twenty minutes a week. Reminders? A note in the calendar and a text sent with her own thumbs — ten minutes. The loyalty scheme? A card and a stamp. This is the test that keeps a not-list honest. If an item has no workable by-hand version, it may not belong on the list at all; if the workaround genuinely costs minutes, then building it now buys almost nothing, and pays for it with days from the ledger and one more thing to maintain. By-hand is not an embarrassment or a failure of ambition. It is the cheapest possible way to run a feature while you find out whether anyone wants it.

  • Every not-item has a by-hand workaround costing minutes a week
  • Invoices by hand, a text reminder, a card with a stamp
  • No workable by-hand version? The item may not belong on the list
  • By-hand is the cheapest way to run a feature while you learn its worth

Refusing What You Could Build

Here is the strange skill this lesson is really about: refusing things you could easily build. The AI makes reminders an afternoon's work, which makes yes feel free. It is not free. Every yes is bought with days from the ledger, and the price does not stop at building. Every feature you ship must then be run, supported and maintained — it can break, confuse a customer, turn up in the late-night email, and demand attention in its second year, all of which this course will put honest numbers on before the end. A feature costs an afternoon to build and a little of every week afterwards, indefinitely. The not-list is where you decline those subscriptions. The discipline feels wrong at first, because refusing easy things always does. But nobody regrets a small version one that works. People regret large ones that almost do.

  • AI makes building cheap, which makes yes feel free — it is not
  • Every yes is bought with days from the ledger
  • Everything you ship must then be run, supported and maintained
  • Nobody regrets a small version one that works

Prefer slides, quizzes, and saved progress? Read this lesson in the library — free, no sign-up.