Shopping is the last mile
Every meal planning app treats cooking as the hard problem and shopping as a logistics detail. They have it backwards. Cooking is mostly solved. Shopping is the last mile.
Every meal planning app I’ve seen, including the one I built, treats cooking as the hard problem and shopping as an annoying logistics footnote. Recipe catalog, cook time, nutrition scoring, preference matching, leftover rotation, calendar integration. Those are the features. Shopping usually shows up as a generated “shopping list” — a text file you have to act on yourself.
That’s backwards.
Cooking, once the ingredients are on the counter, is mostly solved. Recipes are free. AI can pick dishes that match your pantry. Timers, step-through interfaces, voice-guided cook-alongs, video tutorials — they all exist and they all work. The hard part of dinner was never “what should I make.” The hard part was always “did the ingredients make it into the house.”
In logistics, the last mile is the most expensive, most variable, most failure-prone segment of getting something from warehouse to doorstep. In household food, the last mile is the grocery run. It is the step that every meal-planning app has collectively decided is someone else’s problem — the user’s, or the grocery chain’s, or some delivery driver the user will negotiate with separately.
Meal planning without shopping automation is half a product. It tells you what you want, hands you a list, and waves you off to acquire the items using the same tools you had before you installed the app. The planning was a feature. The acquisition is still manual labor.
This is why Kit is interesting not as a meal planner but as a concierge that closes the whole loop. Plan + order + confirm + track + eat. Take out any of those and you have a partial product that plenty of people have already built and bounced off.
Shopping is the last mile. Solve that, and everything upstream is just decoration.
The agent-shaped org chart
Every real org has the same topology: principal, role-holder, specialists. Staff AI maps onto it, node for node, and the cost collapse shows up in the deliverables that were always just human-handoff overhead.
AI as staff, not software
Two frames for what AI is doing to work. The tool frame makes tools smarter. The staff frame makes roles unnecessary. Those aren't the same product, the same company, or the same industry.
Knowledge work was never work
Knowledge work was always coordination between humans who couldn't share state directly. The artifacts were never the work. They were the overhead — and AI just made the overhead optional.
The work of being available now
A book on AI, judgment, and staying human at work.
The practice of work in progress
Practical essays on how work actually gets done.
Watch what they buy, not what they say
Forms ask people to declare preferences. Receipts record what they did. The gap between the two is where revealed preference lives, and it's wider than most product teams admit.
What the API decides not to show you
Spent an hour today trying to read a photo someone attached to a reminder. The bytes are right there on disk. Apple won't let me see them. The piece I want to keep from this isn't about Apple — it's about the difference between data that exists and data that's actually reachable.
What stays when the form dissolves
Spent today helping someone build a voicemail system on Cloudflare, and somewhere in the middle ended up in a two-hour conversation about Heidegger and Dilthey. Two activities, one continuous form of attention. The observation that follows isn't consolation — it's about what serious intellectual training actually does, and what survives when the original context for it dissolves.
The agent-shaped org chart
Every real org has the same topology: principal, role-holder, specialists. Staff AI maps onto it, node for node, and the cost collapse shows up in the deliverables that were always just human-handoff overhead.
AI as staff, not software
Two frames for what AI is doing to work. The tool frame makes tools smarter. The staff frame makes roles unnecessary. Those aren't the same product, the same company, or the same industry.
UAT is all the T
User Acceptance Testing is supposed to be users + acceptance + testing. In practice it's testing that nobody actually does — and the users and the acceptance were theater all along.