This is an idea paper. Use github to participate.

Why DAOs are not it

01 Oct 2020

Success of an org depends on its:

  1. Decision Velocity
  2. Decision Quality

DAOs in their current form suffer from a principle-agent problem.

They work well for simple low-velocity coordination within a group of principals.

Not so much for complex high-velocity coordination with a multi-variate group of principals and agents.

There are 4 kinds of people in each team:

  1. People who don’t care for success of the business.
  2. People who care about business success when the ball is rolling.
  3. People who care about business success (even) when the ball gets dropped.
  4. People who care about business success irrespective of if the ball exists (or will exist) or not.

#1 need to be fired. #2 are short-term agents.

#3 are long-term agents and need to be empowered. #4 are principals.

Treating everyone equally, irrespective of their (type and scope of) contribution, creates system collapse. This is what is also preventing DAOs from scaling.

Thought trigger: Imagine the journey of Tesla, if it was built using a single-layer consensus driven DAO. It will start failing apart at the nth participant. Think about the likely value of n.

This essay is an attempt to explore the issues (if solved) would create the next version of DAOs.

1. Single layered decision-making

Not all decisions in an org are the same. They do not need to incur the same cost.

Visualizing an org as a decision tree:

If a decision that is solvable by a node at depth 4, should not need to travel to depth 1 where a general consensus would be invoked to solve it.

A possible solution is a multi-layer architecture. Perhaps, that would be true-er decentralization.

Interesting permutations are starting to emerge here:

However, they are not enough. The problem is much more core.

2. Consensus-only decision making

Consensus is a sub-optimal decision making process. It has its usecases, but isn't the gospel that is taken to be. Why?

Democracies lead to politics. Politics shift the winning strategy from truth seeking (and enterprise) to narrative building.

In democracies the loudest voice wins; not the truest.

While value-creation requires innovation, taking a risk and being right. All of these are individual pursuits. Groups instead gravitate towards ideology, stability and survival.

Consensus has its place, but there also is a need to delegate majority decision-making to methods more suitable to enterprise. An idea is here.

Highly recommended read:

DAOs & Token-based Governance is a fantasy | @verbine

3. Lack of jurisdictions

Governance without a power to enforce rules, is useless. In crypto systems, because identity is anonymized, the worst you can do is slash a bad actor's balance.

Read more:

Web3 governance may be overrated| @verbine

The problem is perhaps more core. The web2 methods of startup wealth creation, aren't directly applicable to web 3.

The idea of venture capital with company as its primary enterprise form uses control, ownership and intellectual property to succeed.

While web3 uses community, decentralization and network effects as its moat.

If the companies of tomorrow are going to be networks, we do not know how to build them effeciently, yet. The optimal solution would require new thinking.

I reckon the solution would be radically less violent than what is available today.

Until then, there is perhaps a stop-gap solution.