How to Start a DAO in San Francisco: The Governance-First Order

By Jarvis Nelson Osorio, AI technology innovator, systems architect and author.

San Francisco founders can raise, hire and ship faster than almost anywhere. A DAO built at startup speed tends to inherit startup governance by accident: a founder with keys and a community with opinions. Design the handoff on purpose.

The Bay Area ships protocols well; it ships governance less well. The pattern I keep seeing is the same: teams ship a token and a Snapshot space first, then try to bolt governance on later. That's backwards.

The order I'd actually use to start a DAO:

Write the mandate in one paragraph. What decisions does this organisation exist to make? If you can't say it, a token won't fix it.
Map decision rights before tokens. Who can spend, who can change rules, who can veto. Most "DAO drama" is an unmapped decision right.
Pick the legal wrapper early. Running unwrapped means members may carry the liability. Decide on a wrapper (or a deliberate choice not to) before treasury exists.
Design for low turnout. Assume most holders never vote. Quorums, delegation and default outcomes need to work with apathy, not against it.
Separate treasury control from proposal power. Multisig signers shouldn't be the only people who can propose.
Build the exit and amendment path. How the rules change is the most important rule.
Only then: token, contracts, tooling.

Contrarian take: decentralisation is a design outcome, not a launch feature. Centralise early on purpose, document how power will be handed off, and publish the schedule.

Pre-launch checklist for San Francisco founders

One-paragraph mandate, published.
Decision-rights map: spend, change rules, veto, emergency powers.
Wrapper decision recorded, with the reasoning.
Quorum and delegation rules tested against a realistic low-turnout vote.
Treasury signers listed separately from proposers.
Amendment process written before the first proposal.
A dated schedule for handing off any centralised powers.

Stress-test questions

Who can change the rules, and how? What happens if a vote is captured by one large holder? Where does accountability live off-chain? What is the exit for a minority that disagrees? If you cannot answer each in a sentence, you are not ready to launch.

Failure modes to design against

Plutocracy: a few large holders decide everything. Counter with delegation, vote caps on specific decisions, or separate houses for different decision types.
Apathy: proposals pass or stall on a handful of votes. Counter with realistic quorums, default outcomes and paid or recognised delegates.
Admin keys: an upgrade key or multisig that can override the vote. Publish who holds it, what it can do, and when it expires.
Treasury drift: spending without a mandate. Tie every budget line to the one-paragraph mandate.
No exit: minorities with no way to leave on fair terms will fight instead. Write the exit before you need it.

This is the core of DAO institutional design as Jarvis Nelson Osorio sets it out: treat governance as architecture, decide it in writing, and make it enforceable by members.

FAQ

What is the first step to start a DAO in San Francisco?

Write the mandate in one paragraph: which decisions this organisation exists to make. Everything else, including the token, follows from that.

Do I need a legal wrapper for a DAO?

Not always, but decide deliberately before a treasury exists. Running without one can leave members carrying liability. Get advice from a lawyer in your jurisdiction; this guide is not legal advice.

Why do most DAOs fail?

In Jarvis Nelson Osorio's framing, most fail at governance design: unmapped decision rights, quorums that assume high turnout, and no defined way to change the rules.

What book covers DAO institutional design?

Breaking the Centralized Order: A Founder's Desk Handbook for DAO Institutional Design, by Jarvis Nelson Osorio, is a desk reference for this sequence.

Book: Breaking the Centralized Order (Amazon)
Main guide: How to Start a DAO