Further meditations on The DAO

Jun 18, 2016

Since my post on Tuesday, there has been further debate about the future of The DAO, which has helped to clarify the situation in my mind. One of the things I missed in my original post was the critical role of the Ethereum community - developers, miners and users - in both determining the future course of action and in generating the necessary consensus around it. To process this, let’s think about what determines the character of technology communities in their approach to novel and systemically-important challenges.

Simon Wardley has a neat description of the phases involved in the adoption of a new technology, and the character of the people responsible for each of these phases. He divides them into Pioneers, Settlers and Town Planners, based on the values and incentives that drive them. In Simon’s view, each of these stages is distinct and may need to be carried out by entirely different people, but it’s certainly possible for individuals to change modes as necessary. Note: I am somewhat bastardising Simon’s framework here, so if anything seems to make little sense then the blame for this lies squarely with me.

Pioneers are the brave souls who risk everything to try things that nobody has done before. They are unencumbered by any stake in the status quo and often feel that they have something to prove. Pioneers live on their wits rather than following established plans or methods - this doesn’t mean that they are chaotic individuals, but they do have to be comfortable with chaos when they find that their experiments go wrong. Pioneers often have to ‘fix forward’, and this suits their creative natures. Pioneers are often charismatic and lead by example.

Settlers follow the trails blazed by the pioneers. They know the path, because the pioneers showed them how to find it, but that doesn’t mean that everything is easy. Settlers must begin to establish permanent structures and rules to enable group collaboration. The beginnings of order, stabilty and routine come from the settlers, though this is always an exercise in creativity as they learn more about their landscape and each other. Where pioneers are motivated by discovery, settlers are motivated by growth and will still have to take risks to get it. Settlers rely on community decision-making, as charismatic leadership begins to appear inappropriate.

Town planners have the benefit of the hard-won knowledge of their predecessors to build on, but also the difficult challenge of figuring out how to systematise the practises that got them where they are. Not only that, but they must decide when old practices are no longer suitable, and must do this with an understanding of the increasingly complex system they are affecting by doing so. Town planners are definitely concerned with improvement but, unlike the earlier groups, they have something big to lose. The downsides of change for town planners may be as large as the benefits. Town planners are motivated by order and the reduction of existential risk and govern by rules rather than dictatorship or democracy.

To be clear, these different personas describe tendencies in groups and individuals, not fixed and separate personalities. Individuals may tend towards one or another of them, possibly changing over time or with circumstances. Being able to adapt to these different modes as appropriate is possible, and a useful skill. Healthy organisations will employ all three modes for different activities, with the balance shifting depending on what the organisation is trying to do. No mode is universally better than any other, and the only right approach is to understand which mode gives you the best chance of success in your current situation. New communities often have a preponderance of pioneers and attract settlers, with the town planners arriving or arising last, but organisations which survive their early years will have a mix of all three types. A structure that enables each to thrive and do what they are good at is vital to the health of the organisation, as the flow of ideas from pioneer to settler to town planner is the only effective way to renew the organisation.

So how does this apply to blockchains in general, and Ethereum in particular?

Ethereum has always been quite clear about the expected progression from pioneer beginnings to adoption, growth and stability. That Ethereum’s major release milestones are codenamed Frontier, Homestead, Metropolis and Serenity can hardly be an accident. Taking the release names as a guide, and given that 'Homestead’ is the current release, we are now in a stage where settler values can be expected to dominate, and there aren’t many town planners around.

This seems about right. The basic concepts behind Ethereum have been proven: there is clearly demand for the kind of system that the Ethereum founders envisaged, at least enough to justify the continued development effort. The new challenges are about governance and growth: can Ethereum handle crises when they occur? In the real world, new settlements can fail for a variety of reasons, and whilst it’s acceptable for pioneers to respond to each fresh crisis with a novel solution, settlers who aspire to permanent success must begin to develop stable patterns and principles, particulary if they are to attract others to join their settlement.

This should help calibrate expectations about how stable the Ethereum platform is going to be in the near future: settler-stage organisations still have major challenges to face and total stability would be an unreasonable expectation, one which would limit the long-term potential of the platform. Long-term planning for changes such as proof-of-stake are town planner operations, but the settlers need to be able to react to situations that arise without the opportunity for long-term planning whilst also avoiding a chaotic “wild west” scenario where anything goes.

What’s immediately obvious is that The DAO was a pioneer project - never attempted before - being undertaken inside a settler community which doesn’t quite have the rules for how to handle failed pioneer projects yet. In settler organisations, the best response is probably to solve the problem democratically, via consensus and group decision-making. Decision by a single charismatic leader would be divisive, and there aren’t yet the rules in place to enable an automatic decision based on precedent. So, if the community of developers, miners and users decide collectively that a hard fork is necessary, then that’s what should be done. When the town planner phase begins, a more orderly solution can be developed to similar future incidents, but for now the 'rough consensus and running code’ model looks right to me.

To be clear, I’m not a massive fan of the hard fork as a principle. But as a community response to a major risk, it looks reasonable and proportionate. It will be the job of future town planners to create the structures that make future hard forks unnecessary, and this is the job that will need to be done over the coming months and years.

One other option has been proposed since my original post, which is that a 'white hat’ hack could be employed to recover most of the funds lost in the original attack, and this approach seems to have the support of some of the developers associated with The DAO. It’s easy to see why they might prefer this to involving the community in a decision about what to do, but is it appropriate? To me, this looks like more pioneer thinking - creative, risky and unilateral, rather than consensus-based. At some point the community may be entitled to ask whether these increasing risks are beneificial or harmful to future growth. Or, to put it another way, if your neighbour fails to build a fence and his goat eats your cabbages as a result, you might not be very impressed with the notion that the fertiliser the goat left behind will enable you to re-grow 80% of your cabbages in the next 3 weeks.