Getting to know DAO 2.0
Original article on DAOhub.
The DAO contract, or “DAO 1.0”, is an innovation in governance because it is decentralized, provably democratic, immutable, resource efficient and has a relatively low barrier to entry. Putting these feats of engineering aside, it is still a completely flat shareholder democracy, something which by market standards isn’t much of an innovation performance wise.
Improvements in the performance of DAO governance are currently in development, and these are what generally fall under the title “DAO 2.0”.
As one would expect, there’s a lot of noise surrounding the topic. Many think a flat shareholder democracy is doomed for failure, some think it is the pinnacle of fair governance, some think we need liquid democracy, or delegative voting, or perhaps vote markets, quadratic voting, prediction markets, contracted hierarchy, range voting, etc. etc. For all we know, there are as many opinions as there are token holders.
The idea that has surprised me most though is that we should do nothing — or at least nothing for the time being. This is not because of a disagreement with a particular governance proposal, but because radical change is viewed by some as something we should avoid in our early stages. Many seem to hold DAO investment as a silent agreement that the DAO is a flat shareholder democracy and should stay this way; changing this too soon would be a betrayal of sorts. Furthermore, rapid change will confuse investors new to the technology and tarnish our image in the media.
While I agree that token holders shouldn’t be rash and vote for new governance mechanisms without due consideration, this shouldn’t be confused with keeping the DAO homogeneous — if in 6 months time the DAO has branched off into 10 different implementations, that would be a sign of progress, not failure.
What is DAO 2.0?
DAO 2.0 refers to the upgrading of the current DAO smart contract deployed at 0xbb9… . Because smart contracts are immutable, ‘upgrading’ means the creation of a new contract, either for the DAO 1.0 contract to interface with, or to completely move over to (all existing funds/tokens minus those that split). There are currently two pre-proposals for this: one from Backfeed and one from GroupGnosis. Many more are expected.
DAO governance also includes offchain interactions. These currently take place at DAOhub.org, which is also in the process of evolution — a “DAOhub 2.0” if you will. Features being explored include resistance to censorship, downtime and sock puppets.
These DAO 2.0 solutions can be thought of applying to two separate layers of cryptogovernance: voting and census (for a broader look at cryptogovernace, take a look at my previous post The DAO: A Venture in Cryptogovernance). Liquid democracy, delegative democracy and quadratic voting all belong to the voting layer, because, as you would expect, they directly modify the rules of voting. On the other hand, Futarchy creates a secondary layer of betting for the purpose of aggregating information, or “census”.
Getting to Know Futarchy
Disclaimer: Futarchy is rooted in some very complex maths, something I’m certainly not qualified in; refer to Vitalik’s An Introduction to Futarchy for technical details.
A Brief Introduction
Futarchy is a model of governance that puts prediction markets at the core of decision making. This is based on the underlying assumptions that:
- Democracies fail by not aggregating enough available information
- Speculative markets are the best known method of aggregating available information
When originator Robin Hanson first described Futarchy, he meant it as a means of nation-state governance, however the term has generally stuck to the use of prediction markets in any governance model (especially in the field of cryptogovernance). Most people in the space are already familiar with the design and function of prediction markets (Augur, GroupGnosis); if not, prediction markets are simply any speculative market were people can bet on the state of the future and are rewarded/punished for their accuracy, hopefully generating accurate predictions.
There are several ways in which prediction markets can be applied to cryptogovernance. Here I will cover three approaches that help explain some of the DAO 2.0 proposals coming our way.
DAO, meet Prediction Market
1. Prediction Markets for Aggregation of Information
Prediction markets can and will be used as aggregators of information, providing the DAO with a decentralized alternative to in-house analysts (e.g. due diligence teams). With the successful launch of Augur and GroupGnosis, token holders will be able to establish markets for any forecasting they require, the information serving purely as reference for decision making.
2. Prediction Markets for Delegation
Prediction markets can serve as just another governance mechanism in our repertoire: proposals can be made to defer a decision to the result of a prediction market. For instance, Slock.it could present three funding options for the Ethereum Computer in a single proposal — conservative, balanced and aggressive. When the proposal is approved by the DAO, the contract triggers the creation of three prediction markets (with an argument like ‘expected ROI within 36 months’), the outcomes of which will trigger the best funding option.
3. Prediction Markets as Intelligence Selective Agents
The issue with flat organisations is that all shareholders are punished for bad decisions, and all shareholders are rewarded for good decisions: if 5% of funds are invested on a 23% yes/22% no vote, and that investment fails, 100% of investors lose 5% of their funds. This is of course unavoidable: having ‘skin in the game’ is fundamental to intelligent operation in the first place. Used alone however, this leaves the organisation with a single, very crude selective agent for intelligence. In a market where every organisation is flat, the selection for intelligence (aka performance) rests solely between organisations, such that organisations with smart investors profit, and organisations with dumb investors suffer losses.
Traditional organisations have a tried and tested model for overcoming this: performance-based hierarchy. This adds a further selective agent between shareholders by establishing responsibility for decision making: a CEO doesn’t just perform because he has stock options, but because he won’t have a salary and bonuses if he makes bad decisions. This is of course not what we should aim to do with the DAO, as we can achieve the same effect, if not greater, without inviting centralization and communication inefficiencies onto our turf.
The way we do this is by using prediction markets as a secondary and intrinsic layer to decision making, one that rewards token holders based on the accuracy of their prediction. This can be done by automatically generating a prediction market for each and every proposal, running either in parallel to or as an argument to the vote itself.
Going back to the 5% of funds lost to the bad investment, token holders who made the right prediction now have the ability to capitalize on their sound judgement. This not only incentives token holders to put more thought into their decisions, but can also contribute to the power of modified voting mechanisms like delegative democracy: token holders that consistently make correct predictions gather reputation and thus voting power through delegation.
Here we start to see a holy grail of governance emerge, where prediction markets, reputation, delegative democracy and vote markets come together to draw the greatest wisdom from the crowd. It is even possible, using prediction markets, to circumvent the very DAO itself! Instead of advanced prediction markets being built on top of the DAO and its voting mechanisms, we can build the voting mechanisms on top of the prediction market. This is arguably an inevitable end for the technology, if what we are trying to achieve is freedom of capital and the best extraction of wisdom from the crowd.
To illustrate this, let’s consider the Ledger proposal: 6,000 Ethereum Hardware Wallets at a 25% ROI for €120,000. The market could use this investment model:
- Ledger submits the proposal directly to a sophisticated prediction market.
- The market assesses its viability.
- The market confirms its viability (and to a degree).
- The payment contract is launched, capped to the amount asked for in the proposal, and Reward Tokens are generated.
- Done!
Not only have we simultaneously streamlined and decentralized investment on the blockchain (no more single-address DAOs to point at), but we’ve circumvented the problem of individuals accepting proposals they don’t like for the sake of sticking around for proposals they do.
Issues with Futarchy
There are many criticisms of Futarchy; I’ll address the two most relevant.
It’s a Zero-Sum Game
This argument is an interesting one: there is no strict financial object produced by prediction markets (one man’s gain is another man’s loss), and therefore no rational investor would engage in them. For instance, election polls are almost always paid for; if I were to lose money upon making a wrong prediction, I probably wouldn’t participate.
This argument depends very much on how the markets are set up. If the markets are used to decide the outcome of a proposal, then DTH’s have an intrinsic incentive to participate — it’s their investment that is being decided upon. If the markets are used in conjunction with a proposal, purely as an optional opportunity for DTH’s to make gains on better decision-making, then yes, we run into a few incentive issues.
Solutions to the zero-sum game problem include allocating a small percentage of each proposal to rewarding market winners (a just sacrifice for more intelligent decision making), or by allocating reputation to winners such that they can bolster their influence (assuming some form of delegative voting is activated).
Attack Vectors
This ties nicely into our attack vectors: isn’t this starting to get too complex, disrupting the democracy of the organisation? Doesn’t this give whales new avenues to psychologically distort markets?
The answer to this is that any new governance solution will create new attack vectors. And because any upgrade to DAO governance will require votes, it is in our interest as token holders to learn the ins and outs of whatever technology is being proposed.
Conclusion
At the end of the day, it is up to individual token holders to learn how DAO 2.0 solutions work and decide whether they would like to adopt them. That said, there are two appeals I would like to make:
Firstly, don’t get stuck looking at existing governance models for solutions. Yes, VC and corporate governance is tried and tested, but they are models built in a world without turing-complete blockchains. It is easy to emulate the past, but that’s not where gains are made: gains are made by looking to the future, and thinking about how far our current technology can take us.
Secondly, although it is ill-advised to rush into new governance solutions, let’s not confuse this with the idea of keeping the DAO as one homogeneous organisation. The DAO has not ‘failed’ if it splits into different implementations, or ceases to exist in its current form as investment on the blockchain becomes further decentralized. The DAO is an experiment (a rather expensive one at that) so let’s not push aside innovation for the sake of perceived stability.
The rest of the world is watching: let’s show them how far we can take this!
