21. The Organization as Protocol: Toward the Trustless Firm

in the old web, a bank is a building full of people, and its website is just the front door you tap on your phone. In Web3, the bank can be the software — the vault, the rulebook, and the boardroom all rolled into one program that lives on a public ledger. There is no building behind the website, because the website's machinery is the institution.

your banking app lets you see and move your money, but it is only a window. The actual money is held by the bank, the rules are written and changed behind closed doors, and the big decisions are made by people in a boardroom you will never enter. You are trusting an institution you cannot inspect. The website is online; the institution is not.

turning an organization into a protocol means putting the three things a bank hides — your money, the rulebook, and the decision-making — out in the open as code. The money sits in transparent contracts, the rules run exactly as written for everyone to read, and decisions are made by recorded votes. You no longer trust an institution to behave; you verify a program that cannot do otherwise. At least, that is the design. Whether reality matches it is the question this chapter ends on.

Coase's great question was why we bother with companies at all instead of everyone just buying and selling from one another. His answer: because using the market is full of hidden costs — finding partners, haggling, writing contracts, and making sure people keep their word — and a company is a way to cut those costs by doing things in-house under one roof. Web3's wager is that smart contracts slash exactly those costs, because a contract that enforces itself needs no lawyers, no courts, and no one you have to trust. If that wager holds, the basic economics of why organizations look the way they do changes.

these are not thought experiments. Sky issues a dollar-pegged currency and manages its collateral by code, like a central bank with no governors. Uniswap lets people trade with no company holding their money or matching their orders. Aave and Compound lend and borrow with no banker deciding who qualifies. And Bitcoin runs a global money with no company, no boss, and no boardroom. Each takes a function we assumed required a trusted institution and rebuilds it as a program. The differences among them — and how much trusted-human residue each still contains — are the subject of the reckoning that follows.

the earlier web put a great website on top of the same old companies. Web3's real bet is to turn the company itself into a website's worth of code — something open, copyable, and runnable by anyone, that you check instead of trust. That is the idea that ties the whole book together: first individuals, then markets, then decisions, and finally entire organizations, all rebuilt to work without a trusted middleman.

the trustless organization usually still has trusted humans in it — they are just harder to see. A small team often holds master keys that can rewrite the rules; a committee with a shared wallet acts as a board; the final step of carrying out a vote tends to run through people; the website everyone uses is owned by a company that can and does filter it; and to stay legally safe the whole thing often becomes a normal LLC after all. Critics argue "decentralized" can be a costume that hides who is really in charge and lets them dodge the blame. And when code goes wrong, there is no one to call and nothing to reverse — which is also a reminder that some human discretion, the kind that grants a second chance, may be worth keeping.