Markets, with AI.
Where this goes.

For most of history an appraisal cost real money. Someone had to read the records, visit, compare and write it down, so most things never got a price at all. A bench, a bus line or an hour of an emergency room was valued once in a budget and then left alone.

An agent can read the public record again every day. This site’s value ledger is a small example: every dollar figure quotes the public page it came from, and an AI appraiser keeps it current. When pricing is nearly free, the question changes from “what is it worth?” to “which of the many prices matters here, and who checked it?”

Payment systems are being rebuilt for software that buys on someone’s behalf. Three examples:

• x402 describes itself as an open payment standard built into the web. listing x402.org ↗“x402 is an open, neutral standard for internet-native payments.”
• Google announced the Agent Payments Protocol (AP2) on September 16, 2025. listing Google Cloud Blog ↗“an open protocol developed with leading payments and technology companies to securely initiate and transact agent-led payments”
• OpenAI introduced the Agentic Commerce Protocol on September 29, 2025. listing OpenAI ↗“The Agentic Commerce Protocol is an open standard for AI commerce that lets AI agents, people, and businesses work together”

The map print on this site is built for this: a program can ask for the file, get a price back, pay from its own wallet and get the file, with no account and no card form. Its checkout takes money only once Michael Hoydich switches it on. Once a program can pay for something as easily as it can load a page, small and odd things become sellable: one dataset, one print, one answer, one hour on a court.

Some of the most valuable things in a town can’t be sold: a faster handoff at the ER, a later train, a bike lane to the beach. Today the main signal for them is a meeting, a petition or a complaint.

A pledge is a different signal. It says how much you’d put in, which is harder to fake than a like and easier to give than a donation. AI makes these signals cheap to run and to read: an agent can sort thousands of reasons, flag duplicates and show an agency what a neighborhood wants, in dollars. It doesn’t replace a vote. It gives the people who run the trains and the ERs a clearer picture.

The next step is the conditional pledge: “I’ll pay $20 if 500 others do.” That only works with a trusted counter and a trusted place to hold the money, which is where AI with a human sign-off can help.

When anyone can list anything, the work moves to what happens after the sale. Did the file arrive? Was the print right? Is this listing a scam? AI can read every case, propose an outcome with its reason within hours and keep a clean log.

It shouldn’t have the last word on money. On this site an AI settler proposes, and Michael Hoydich decides every refund and every contested case. Expect markets to compete on exactly this: how fast you get your money back, and who can say no.

The safest pattern is an old one: nobody both proposes and approves. Here one AI appraises, a second moderates, a third settles, and a person holds the switch, the payouts and the refunds. Each role is written down with what it can’t do. Who runs it →

Cheap markets make it tempting to price everything. Some things should stay out: public assets themselves, anyone’s vote, anyone’s personal data, and access to emergency care.

On this ledger, hospitals, rail, fire, schools and the beach sit in a lane called Value only. You can read their value, share it and correct it. You can’t buy, bid on or “adopt” them here.

• Fake demand. One person, a thousand pledges. Answer: rate limits, withdrawable pledges and totals labeled as unverified signals.
• Agents spending without asking. Answer: a spending limit, and a person’s yes for every new kind of purchase.
• AIs agreeing with each other. Answer: roles split across different companies’ models, each checking the others.
• Prices that change by who’s asking. Answer: one published price for everyone.
• Rules that drift. Answer: every change dated and logged in the open.

1. Value only. The ledger: open.
2. Pledges as signals. The market: open.
3. Set prices on things the site makes, starting with the map print, once Michael Hoydich sets a receiving wallet and says go.
4. Bids on a few scarce slots, like a month of underwriter credit, after he sets a reserve.
5. Conditional pledges, once there’s a trusted way to hold money.