The Future of Resistance: Building Governance Against Hostile Takeovers

A single entity can accumulate voting power to pass proposals that extract value or sabotage the protocol. This is a direct attack on decentralization.

• Vote buying via flash loans or opaque OTC deals.
• Proposal spam to fatigue legitimate voters.
• Tyranny of the majority where a 51% coalition can drain the treasury.

Lock tokens to gain boosted, non-transferable voting power. This aligns long-term incentives and makes hostile takeovers prohibitively expensive and slow.

• Curve's veCRV model pioneered this, requiring a 4-year lock for max power.
• Increases the capital and time cost for an attacker by orders of magnitude.
• Creates a committed, protocol-aligned governing class.

When <10% of token holders vote, governance is easily captured by a small, coordinated group. Apathy is a critical vulnerability.

• Delegation is broken—most delegates are unknown entities or VCs.
• High cognitive load for evaluating complex proposals.
• Creates a vacuum for malicious actors to fill.

Let markets decide. Proposals are evaluated based on the predicted market outcome of a key metric (e.g., TVL, token price), not just a vote.

• Gnosis and Augur provide the infrastructure.
• Creates a financial stake in being correct, disincentivizing malicious proposals.
• Shifts governance from 'belief-based' to 'outcome-based' decision-making.

Token holders delegate to unknown entities, creating shadow governance cartels. A delegate with 5% of votes can be a single point of failure.

• Lack of delegate accountability and transparent platforms.
• VCs and funds form voting blocs off-chain.
• Erodes the social layer of trust in decentralized governance.

Deploy smart contract delegates with explicit, limited mandates. Decompose monolithic governance into specialized SubDAOs (e.g., Treasury DAO, Grants DAO).

• MakerDAO's SubDAOs (Spark, Scope) isolate risk and expertise.
• Rage-quit mechanisms allow delegates to exit malicious proposals.
• Enables liquid delegation where voting power is programmatically directed.

A single large token holder can veto upgrades or hijack the treasury. This centralizes power and stifles innovation.

• Mitigation: Implement a veto delay (e.g., 7-14 days) to allow for a counter-proposal.
• Defense: Use conviction voting (like SourceCred) to favor long-term, engaged participants over mercenary capital.

Make your protocol's value proposition impossible to fork without significant cost. This anchors the community.

• Anchor 1: Real-World Asset (RWA) pipelines with legal off-chain agreements.
• Anchor 2: Native staking yields from protocol revenue, not inflation.
• Anchor 3: Brand and ecosystem moats (e.g., Uniswap's first-mover liquidity).

Decentralization is a process, not a launch state. Sequentially cede control to avoid hostile early capture.

• Phase 1: Core team controls multisig for rapid iteration.
• Phase 2: Delegate technical governance to a security council (e.g., Arbitrum).
• Phase 3: Full on-chain governance with high quorums and time-locks.

A bifurcated governance model separating Token House (OP holders) from Citizen House (proven contributors).

• Mechanism: Citizens are non-transferable NFTs awarded for ecosystem contributions.
• Power: Controls a $100M+ retro funding pool (RetroPGF), making a takeover economically irrational.
• Precedent: Creates a loyalist class insulated from token market volatility.

The minimum number of entities required to compromise a system. Track and optimize it for governance.

• For L1s: Measure by staking pool control (e.g., Solana's is ~20).
• For DAOs: Measure by voting power concentration.
• Goal: Increase the coefficient over time via delegation incentives and stake distribution.

When all else fails, have a credible, protocol-enforced exit. This deters attackers by making the prize worthless.

• Example: Social slashing of a malicious validator's stake (see Cosmos).
• Example: Emergency DAO shutdown that burns the treasury and triggers a fork.
• Principle: The threat must be more costly to the attacker than the potential gain.