Ha! Lots to think about. (where to start…)
DAOs are governed by a set of rules that are functionally immutable, so at least some bits of entropy are minimized. In essence, mechanism design is tokenomics crossed with game theory.
I’m essentially a creature of the Santa Fe complexity group. In full disclosure, my philosophical biases are heavily influenced by the belief that complex adaptive systems cannot be measured through mechanistic designs. This is an a priori argument for me, and makes it difficult to evaluate such systems. To be clear, it is not a criticism, but simply the issuance of a point of view, but also helps us understand where we may disagree.
(i read through your link btw - thanks!)
The core premise behind much of complexity based research is the evaluation around ontology, or causality. Where many see a mechanistic, Newtonian world that can be understood given enough study and investigation, others, like myself, see only probabilities and dispositionality. These probabilities are distinctly non-predictive in complex systems. The world-view models can be differentiated between ‘deterministic and reductionist’ and ‘non-deterministic’. I’m squarely in the latter camp. It doesn’t mean that Mechanism Design doesn’t have utility, only that i recognize limitations in how effective its modeling can be in predicting outcomes.
For instance, the mathematical notation used to measure a system’s utility and the desire that it exceeds the cost is very troublesome. (in the embedded paper attached to your article link)
- This doesn’t account for hierarchical benefits through multiple systems. A system can never be evaluated within itself, but only in the system harbouring it. A car for example, is a system. It has a goal however, which can ‘never’ be determined by reductively examining a car. Only when you zoom out to scan transportation systems, supply/demand, commuting, freedom of travel, do you being to understand what a car truly represents. Measurement of this network effect is impossible in a complex system.
- Humans make everything ‘complex’, even in a DAO with immutability built into it. This forum, governance discussion around the first proposal, is proof of that. If you widen the system to include economic modeling, threat modeling (SEC coming in and poking around), viability of the crypto-anarchist world view given current political environment, utility of DeFi, general prosperity of crypto natives and non-natives, etc… The point of this list is its length. You begin to see that variables contributing to outcomes are non-deterministic. The problem with attempting to place too much value on modeling is that you restrict the possible outcomes to the number of variables the model can account for.
This is my fundamental issue with these approaches.
- Many will claim success when they find a ‘use case’ to demonstrate that a certain type of modeling worked, but this is when i will fondly recall George Box’s quote (famed statistician)… “all models are wrong. some are simply more useful than others”. How i interpret this is to concede that modeling the universe or human behaviour is impossible. Worthwhile? Yes! But only when you recognize the limits. As a panacea it becomes a blunt weapon that dulls understanding.
I thank you for giving me the benefit of the doubt with
When i said i didn’t care about the ‘specific’ governance changes, it is only because i care far more about the ‘governance engine’ itself. I’m always zooming out to evaluate double-loop thinking, or problem dissolution - not resolution. Dissolution is when you ‘reframe’ the system to ensure the problem never returns. In other words, i am challenging the DAO foundation, and privileging that concern over the individual proposals that operate on top of that framework. So, hopefully you now see elevated concern where you may have seen a lack of it earlier.
I agree with your suggestion to marry the methods of inquiry together (mechanism design with complexity). If i knew exactly how i would be writing papers about it. This helps reveal that what we are doing is fundamentally ‘experimental’ in nature. Humans simply have never successfully implemented this type of social organization before. If we maintain that perspective on ‘experimentation’, we are likely to be surprised by ‘beneficial’ outcomes. However, if we remain mired in trying to make the existing framework function, we limit the variation in responses, and almost ensure our demise. So, a round-a-bout way to answer your question
It’s a mindset first and foremost. To accept that we cannot possibly predict outcomes through design, only ‘nudge’ towards our north star (goals). Once you accept the interminability of outcomes, then we can skillfully use ‘mechanism design’ to tinker with the gears and pulleys to catalyze change.
Finally, most people are unaware that evolution (nature’s mechanism for adaptation) works mostly through exaptation, which is impossible to model. For example, the wings of a bird are an exaptation. The reason is that feathers were initially evolved to provide warmth - not flight. The subsequent adaptation to flight would only have happened if feathers had evolved a utility for that species. Most of evolution works in this manner. An experiment is not strictly useful because it proves or disproves a hypothesis. An experiment in complexity-based science, is useful because it catalyzes non-predictive changes, many of which may result in a beneficial adaptation. Trying to ‘stabilize’ Uniswap governance is the equivalent of slowing adaption and removing any possibility of a beneficial exaptation. In my philosophical model, this is deadly.
(we should create a separate post to continue the discussion - i fear we are going well beyond the OPs intent - in Meta-Governance? really enjoy the honest dialog)
