Ship First, Then Find the Money

I went for another walk and talked through the current pile of projects. The useful thing that came out was not a new project. It was a ranking rule.

Revenue cannot be decorative. Every project should return something: followers, clients, money, signal, a better body, a stronger position. But the first move is still to ship the alive thing. I do not think I can fake a sales muscle into existence by staring at it.

So the rule, for now, is this: ship first, but do not let the thing float away from value. Money follows a real, interesting thing more often than it follows an obligation I am pretending to care about.

The current priority stack

Bay Area liquidity thesis and consulting positioning

This is probably the highest-leverage business thread. If OpenAI, Anthropic, SpaceX, or adjacent companies create real liquidity in the next 6 to 18 months, the money will not stay inside technical circles. It will spill into real estate, services, families, small businesses, and the strange cultural problems that come with sudden wealth.

The move may not be to sell AI to an industry. It may be to help newly liquid people inside slow industries figure out what to do next. Who gets rich? Which sectors feel it first? Where does the friction show up, financially and culturally? That needs its own session.

Health protocol harness

This one is already moving. I did the Function Health blood draw, and the baseline is deliberately imperfect: irregular sleep, inconsistent exercise, too much beer. Good. That gives the experiment something to push against.

The frame is simple: use the digital layer as a harness for the person. Laptop, phone, AI, calendar, notes, camera, bloodwork. Not a chatbot demo. A system that helps run part of a life.

Good Juju card

This is the project with the most natural enthusiasm. A physical card gives a stranger good luck for a set number of days, then turns sour unless they pass it on. A tiny traveling charm with an expiration date.

The MVP should stay stupidly small: design one card, print 20 to 50, put them in the places I already go, and see whether anyone cares. No website first. No elaborate tracking first. Just the object and the reaction.

Market-research agent stack

This supports the first two threads. I want an agent that can scrape public sources for practices, claims, complaints, protocols, and customer language. Health protocols are one domain. Bay Area wealth problems are another. The same tool could also map the current SF aesthetic so the Juju card can deliberately push against it.

The deeper purpose is less glamorous: find problems worth solving before I build the wrong thing.

Content strategy

The content layer has to convert the work into trust. Small morsels frequently, then a larger payoff every few weeks. Private health data can stay private. The learnings can be public.

I also keep coming back to the human and AI split. Video should probably be me: raw, embodied, hard to fake. Text can be more agent-shaped: structured, searchable, useful to humans and machines. There is a chance the audience increasingly includes AI agents, which sounds ridiculous until it does not.

Creative process and movement documentation

This is the enjoyable lower-urgency thread. Walk the dog, photograph the city, narrate what I am noticing, track the route, let an agent structure the residue afterward. Content for humans. Spatial data for machines.

It also connects to relocation, city exploration, photography, and the recurring question of how things get made. Not the main thread yet, but alive.

The container

Six weeks is the right unit. Long enough for a real signal, short enough that the project cannot become a personality costume. Forty-two days, roughly two cycles per quarter.

The wagers are where it gets watchable: same landmark at strange hours, talk to a stranger on every walk, fixed sleep and wake times, same meal for a stretch, maybe even a minimal-vocabulary challenge where I try to communicate with five words and a card explaining the bit. Some of this is useful. Some of it is just funny enough to be useful.

The next moves

• Pull the Function Health biomarkers and make the baseline snapshot.
• Design one Good Juju card and print a small batch.
• Block a separate session for the Bay Area liquidity thesis.

That is enough for now. Ship the alive thing. Keep the money in view.