The following text was first posted back in July 2016. Does it hold up today?
There's been a lot of talk about the idea of a DAO, but the closest we come to a "real" DAO (a "Decentralized Autonomous Organization") is the Bitcoin network itself. But – of course, it's not entirely autonomous. It needs humans to run the nodes, the mining, the wallets (the inspectors) and the monitors, etc.
First, let's define what the dream of an actual "Decentralized Autonomous Organization" isn't.
It's not a "smart contract," not a script of any kind, and not a "Dapp" (a "Decentralized App"). An "Organization" is something different (the word implies a collection of humans involved, by definition).
In this piece, in Coindesk, the author argues that it's possible to sue a DAO. Well, then it's an "old" kind of a DAO (like an SPV, the "Special Purposes Vehicle" in use in some jurisdictions). Why is that? Well, it's because if you can find someone "responsible," someone to sue... it's not decentralized. In Bitcoin, there is nobody to sue. In BitTorrent, there is nobody to sue. If IP packets get lost on the Internet, there is nobody to sue. See the pattern? If you can sue somebody to stop/affect a significant change, it's not decentralized. And if you roll back a complete production blockchain with value in it (like Ethereum is trying to do), it's not decentralized.
Slock.it is an excellent example of a company trying to be a DAO (based on Ethereum, of course). Here is a quote from Slock.it:
"A DAO is an organization that's self-governing and that isn't influenced by outside forces: its software operates on its own, with its bylaws immutably written on the blockchain, not controlled by its creators. DAOs are formed by groups of like-minded individuals with specific projects and goals in mind . . . A DAO purely manages funds: in itself it does not have the capabilities to build a product, write code or develop hardware. It requires a Service Provider for this purpose, which it hires by signing off on a proposal."
That last part sounds very much like the Bitcoin network, where the miners are the "service providers" and the protocol is the DAO (the whole network of participants is the "organization"). Miners submit proposals (solutions to the next block), and the network/protocol (miners) "sign" the winning "proposal" and ads it to the chain. Even if individual miners can leave and come as they see fit, the protocol stays up and running all the time (as long as there are at least a few miners willing to do the work, and there has been an increasing number of miners willing to do that, for more than seven years now).
Voting not enough to achieve "autonomy"
The so-called "The DAO" is therefore stupid in its design if the goal was to create something "autonomous" as it requires a lot of engagement from token holders for anything significant to happen (they need to vote on a multitude of proposals continuously). As we've learned from BitShares, people have other things to do (like make money). People just don't enjoy voting all the time.
If we apply this thinking to the stupid idea of the Ethereum/"smart contract" confusion that is Slock.it, where the founders think we need to add a (slow) blockchain connection to locks – so we can rent stuff out (like that problem wasn't solved already), using "the blockchain," we find another typical example of founders/owners dreaming of creating an autonomous vehicle (a DAO) that'll make them rich without them needing to do any work. Sorry, guys, but the universe just doesn't work that way.
And "smart contracts" as a function are nothing new really (even if the term is relatively new) – financial institutions have been constructing them for many years (albeit not decentralized). Dapps (when truly decentralized) would be something new. But a real DAO is something *entirely* new, and the Bitcoin network itself fits the description more than anything else (even better than BitTorrent and/or TCP/IP).
Self-regulation is key
This all actually comes down to the concept of self-regulation. To achieve autonomy, you need self-regulation, right? The code needs to be designed so that interactors with the code will benefit from improving the code (devs/miners/inspectors owning bitcoins in the case of Bitcoin). That's a difficult task to achieve, and in a Turing-complete environment where it's tough to predict/foresee all types of effects (as so-called "smart" contracts can reference themselves or anything else), it seems more or less impossible (so good luck, Ethereum). The simplicity of Bitcoin is a feature, not a bug. Bitcoin is integrated, all its parts fit together seamlessly, and the only connection the protocol has to the outside world is the mining (miners buy electricity from the "real world" and need to sell their bitcoins on exchanges in the "real world"). Everything else happens within the protocol. That's a strength as you don't have to worry about all those yet-to-be-found attack vectors you'll find in the "smart contract"/Turing-complete world (just look at what happened to the "The DAO").
The self-regulation in Bitcoin is evident. Not even the halving caused any problems. The exchange rate even adjusted beforehand to compensate for the decreased number of bitcoins coming into circulation each day (1800 now, instead of 3600). If miners can't make money, they drop out, and others take over. Simple as that. If there is a proposal (like BIPS supporting Seg-Wit) that benefits everyone, you see a +95% support in block voting and software upgrading across miners/nodes running the network. Simple as that.
What if all these self-proclaimed DAO-"experts" would focus on actually solving a real-world problem (using the excellent functionalities of the Bitcoin blockchain) instead of confusing "old-world" concepts (like "organization") with made-up counterparts in this shiny new world of crypto?
After all, nothing is autonomous except nature.
Ps. The far future holds some promises on independent sentient AI:s (which would be needed for an autonomous "organization"), so maybe we'll see those DAO's around the year 2035 (if Kurzweil's theories on exponential growth and accelerating returns are correct). Dapps, on the other hand, may see the light way sooner, but let's return to that subject in another post.