Week 9
Mechanism Design And Plural Governance
- Decentralized systems from zero, week 9
- Shape the rules so good behavior is the easy choice.
Outcomes
What You Will Be Able To Do
- Explain mechanism design.
- Describe a Sybil attack and defenses.
- Compute quadratic voting and funding.
- Compare conviction voting and futarchy.
- Apply cosmo-local ideas.
Last week
Recap Of Week 8: Where DAOs Fail
- Plutocracy: money decides.
- Voter apathy: few participate.
- Attacks: flash loans and hidden keys.
- This week offers rule changes that respond.
Core idea
Mechanism Design
- Game theory predicts behavior under given rules.
- Mechanism design chooses rules for a wanted outcome.
- Work backward from the goal.
- Nobel Prize 2007: Hurwicz, Maskin, Myerson.
Checklist
Three Design Questions
- What behavior does the rule reward?
- Who can fake that behavior cheaply?
- What does an attacker gain?
Identity
Sybil Attacks
- One attacker creates many fake accounts.
- New addresses cost almost nothing.
- Counting addresses does not count people.
- Every one person one vote system faces this.
Identity
Defenses Against Sybils
- Cost: make identities expensive.
- Identity: check for one unique human.
- Social graph: members vouch for each other.
- In person: verify at gatherings.
Voting
Quadratic Voting: The Rule
- Each voter gets a budget of credits.
- n votes on one issue cost n squared.
- Strong preferences cost more per vote.
- Voters spread credits across issues.
votes cost
1 1
2 4
3 9
10 100
Voting
Quadratic Voting: The Catch
- Splitting credits across fake accounts is cheaper.
- It needs Sybil resistance to work.
- Collusion and vote buying remain risks.
- MACI-style privacy reduces proof of how you voted.
Funding
Public Goods
- Many benefit, few pay.
- Open-source software is a typical example.
- Markets underfund them.
- Quadratic funding targets this gap.
Funding
Quadratic Funding
- Donors give small amounts to projects.
- A matching pool tops up each project.
- Match rewards the number of contributors.
- Gitcoin Grants ran many rounds.
| Project | Donors | Each | Matched |
|---|
| A | 4 | $1 | $16 |
| B | 1 | $4 | $4 |
total = (sum of sqrt(d_i)) ** 2
[100] -> 100
[1]*100 -> 10000
Voting
Conviction Voting
- Stake tokens on a proposal.
- Support grows the longer the stake stays.
- A threshold releases the funds.
- Moving stake resets that part of the conviction.
conviction(t+1) = a * conviction(t) + stake
0 < a < 1 (decay)
Markets
Futarchy
- Vote on values, bet on beliefs.
- A vote sets a measurable goal.
- Markets predict the goal under each policy.
- The better predicted policy is adopted.
Markets
Futarchy: Open Questions
- Liquidity must resist manipulation.
- The goal metric must be hard to game.
- Members must trust an abstract system.
- MetaDAO ran early experiments on Solana.
Plural
Plurality And Cosmo-Local Ideas
- Pol.is clusters opinions and shows common ground.
- vTaiwan used it for policy talks.
- Cosmo-local: light global, heavy local.
- Share designs globally, build locally.
Summary
Recap
- Design rules backward from the goal.
- Sybil resistance underpins fair voting.
- Quadratic tools reward breadth of support.
- Conviction and futarchy change timing and information.
Coming up
Next Week: Interoperability And The Open Web
- Oracles and cross-chain messaging.
- Mastodon, ActivityPub, and Nostr.
- Local-first software.
Wrap up
Questions And Exercise
- Exercise: compute quadratic costs and matching.
- Simulate a Sybil split and note the effect.
- Read the Liberal Radicalism paper's introduction.