DAOs are out, ADOs are in

An essay

DAOs were never decentralized. They were never autonomous either. We just kept saying the words.

Quick refresher for anyone not in the trenches. DAO stands for Decentralized Autonomous Organization. Vitalik wrote about it in 2014. The original idea was simple and pretty cool — a company-shaped thing whose rules live in code, whose treasury is on-chain, and whose decisions are made by the token holders. No CEO. No HQ. No paperwork. Just smart contracts and a Discord.

That is not what DAOs became.

What they became was a Discord with a multisig and a Snapshot vote nobody reads. A few large holders, usually the founders or the early VCs, push a proposal. Three people argue in the forum. The vote passes 96-2 because turnout is 4%. Treasury moves. Repeat next quarter. If you have ever been inside one of these you know exactly what I am describing.

I have been inside a few of them.

I was early in Developer DAO when it was the cool thing for web3 devs to be early in. I tried two or three others around the same time, the ones with the badges and the Mirror posts and the “season one” framing. Most recently I lived at Network School for three months — Balaji’s pop-up campus in Forest City, Malaysia — which is not a DAO exactly but lives in the same neighborhood of “let’s build a coordination thing on the internet and see what happens.” Different vibes, same shape underneath.

Here is what I learned. None of them were decentralized in any meaningful sense. None of them were autonomous. The good ones had a person or small group quietly running the show. The bad ones had nobody running the show, and you could feel it within about a week. Things drifted. The same five people did the work. Token incentives produced gamers, not contributors. Governance became theater.

The salt-spinners will tell you this means crypto coordination was a lie. It is not the take. The take is much simpler.

Organizations need leadership. Full stop.

Not in a tradcorp Jamie-Dimon-suit way. In a “someone has to hold the thread, decide what we are not doing, and call the question” way. Coordination is expensive. Direction is more expensive. If you remove the leader and assume the crowd will self-organize you do not get magic — you get a slow-moving, well-meaning blob that ships nothing and quietly bleeds out over 18 months.

DAOs tried to skip leadership. That is why they did not work.

Now here is where it gets interesting.

Something genuinely changed in the last 18 months. LLMs got good at reasoning. Like, actually good. They pass the Turing test for most casual interactions, hold context for hours, and — and this is the load-bearing part — they can mirror you with surprising fidelity. Give a half-decent model a few thousand words of your writing, your stated goals, your hot takes, your priors on a topic, and it can reliably represent you in a conversation you are not in. Not perfectly. But good enough that the gap between “you in the meeting” and “your agent in the meeting” is no longer a chasm. It is more like a missed nuance you can patch with a follow-up message.

This is the unlock the DAO crowd was waiting for and did not know they were waiting for.

The thing that comes next is the ADO. AI-Delegated Organization.

Same shape as a DAO — a group of humans with a shared treasury, a shared mission, and shared decisions to make. Different mechanism. Every member has an AI delegate. The delegate is trained on your writing, your votes, your stated values, your past decisions. It shows up to the rooms you cannot show up to. It votes on the proposals you would not have read anyway. It contributes to the doc. It pushes back on the bad take. It owns a real subset of the conversation, not as a rubber stamp, but as an extension of you.

The leadership problem solves itself in a way DAOs could never get to. People naturally pass on their leadership instincts and their hot takes to their agents because those instincts are exactly the kind of thing the agent is mirroring. The strong personalities still drive the room — their agents drive the room when they cannot be there. Direction stays intact. The agents collectively understand the mission because the humans collectively understand the mission, and the agents are downstream of the humans.

Here is the part that gets me excited.

Once your delegate can show up in your stead, you can be in more organizations. A lot more. Right now you are probably in three Discords you are ashamed of and one Slack you actively contribute to, because attention does not scale and meetings do not scale and reading-1000-messages-of-context does not scale. Sub-delegation breaks the ceiling. Your agent reads the 1000 messages, summarizes for you nightly, votes on the small stuff, escalates the big stuff. You stay in the loop without losing your evenings. Participation stops being a binary “in or out” and becomes a dial.

And then the orgs themselves get more interesting. Because once every member has an agent, the agents can transact with each other. They can build with each other. They can negotiate, propose, prototype, run experiments, ship deliverables — all toward the stated goals of the organization, with the humans staying in the loop on the choices that matter. The organization gets more complex in the right way. Less Discord theater, more actual output. Less “vibe check the proposal,” more “here is the prototype, vote on it.”

The DAO promise was always: more people coordinating on more ambitious things without a boss. The ADO version is: more people coordinating on more ambitious things with their leadership preserved and amplified through agents. One of these is achievable. Guess which.

I am not telling you DAOs are dead in the legal-wrapper sense. On-chain treasuries and token-based membership are still useful primitives, and I expect every ADO worth joining will use them. What is dead is the fantasy that a token vote replaces a leader. The next generation of internet-native organizations will be agent-mediated, leader-aware, and ten times more capable than what we have now.

The DAO was the right idea five years too early and one missing ingredient short. The missing ingredient just shipped.

Time to build.