The extractive philosophies have all been tried. They all failed structurally. What remains is the one philosophy the space has not yet built at scale.
1. The pattern after ten years
Crypto has spent more than a decade trying to design financial systems from scratch. In that time the space has tried, in order, four distinct economic philosophies. Each was promoted by serious people as the path to a fairer system. Each produced systems whose dominant strategy was extraction.
The first was pure libertarian permissionlessness. The premise was that removing intermediaries would remove rent. The intermediaries did get removed. What replaced them was Maximal Extractable Value, sandwich attacks, frontrunning bots, and the routine wealth-transfer from retail users to professional searchers operating in the microsecond gap between transaction submission and execution. Permissionless without fairness guarantees produced extraction at network speed.
The second was regulated clearinghouses, attempted by every centralized exchange and most jurisdictional crypto frameworks. The premise was that licensed gatekeeping would prevent exploitation. Gatekeeping appeared. So did jurisdictional arbitrage, custodial failures (FTX, Celsius, Voyager, others), and a system where 1.4 billion unbanked adults were excluded not because they were risky but because the cost of evaluating their risk exceeded the profit of serving them. Regulated without permissionlessness produced exclusion at institutional speed.
The third was commons-based peer production. The premise was that voluntary contribution would sustain shared infrastructure. Bitcoin and Ethereum both started here. Both ended with mining cartels, validator concentration, and core developer teams whose decisions affect billions of dollars and answer to no one. The commons did not survive contact with financial incentives. Voluntary norms get arbitraged by anyone with skin in the game.
The fourth was platform cooperativism, attempted by user-owned DAOs across DeFi. The premise was that user ownership would prevent platform extraction. Then came governance token concentration, voter apathy, multisig captures, treasury embezzlements, and bribe markets. Every major DAO has been captured at least once. Democratic governance does not survive contact with whales or paid voting infrastructure.
The space has tried four philosophies. Each is a different answer to the same question: how do you design a system where rational agents do not extract from each other? Each was correct about what it rejected. Each was wrong about what it offered as a replacement.
2. What is structurally left
If we honestly enumerate what has been tried and failed, the remaining option is also the only option that has not yet been built at scale. It is cooperative capitalism: a system where cooperation is enforced by mechanism design, not by social norms (which get arbitraged), not by legal structures (which get captured), not by governance votes (which get bribed). The cooperation lives in the math.
The phrase sounds ideological. The implementation is not. Cooperative capitalism is a mechanism-design pattern, not a political program. The pattern has one rule:
Mutualize the risk layer. Compete on the value layer. Design the boundary between them deliberately.
Most current crypto systems mutualize nothing and compete on everything, including the risk layer. That is why every market crash cascades, every oracle failure becomes catastrophic, and every governance attack succeeds. Risk that should be shared across participants is instead borne by whoever happens to be holding the bag when the music stops.
Traditional finance does the opposite: it mutualizes through gatekept insurance regimes and competes within an exclusive class of pre-approved participants. The risk is shared, but only among those allowed inside. The 1.4 billion unbanked are not protected; they are excluded.
The third path, the one that has not been built at scale, is to mutualize the risk layer permissionlessly while keeping the value layer competitive. The risk layer covers losses (insurance pools, treasury stabilization, impermanent loss protection, retroactive clawback when exploits land). The value layer covers gains (arbitrage, liquidity provision, priority bidding, plugin innovation). Anyone can participate in either. The mechanism design ensures neither layer contaminates the other.
3. Why “structurally left” matters
The framing matters because “we should be more cooperative” is a moral claim that crypto has heard for ten years and ignored. Cooperative capitalism is not a moral claim. It is a process-of-elimination claim: every other philosophy has been tried at scale and produced extraction; this one has not been tried at scale; the only honest experiment remaining is to actually try it.
The argument is closer to game theory than to ethics. In any sufficiently large system of self-interested agents, dominant strategies dominate. If the dominant strategy is extraction, extraction is what you get, regardless of how earnestly anyone wants something else. The way to get cooperation is not to ask for it. The way to get cooperation is to design a mechanism where cooperation is the dominant strategy and extraction is not.
The mechanism design has four moving parts.
First, Shapley value distribution. When the system distributes rewards (trading fees, liquidity mining, governance allocations) it uses Shapley values, the unique distribution rule from cooperative game theory satisfying efficiency, symmetry, the null-player property, and additivity. Shapley is sixty-plus years old. It has been hiding in plain sight as the answer to “what is the fair share of a coalition’s surplus for each participant” for that entire time. The null-player axiom alone closes the sybil-extraction loop: a contributor with marginal contribution zero earns zero, by axiom. No social norm needed. No detection mechanism. The math is structural.
Second, mutualized risk pools. Insurance is permissionless. Anyone holding a covered asset is automatically insured. Underwriters earn premium for taking the risk; they bear losses if the trigger fires. The pool is the insurer. The community insures itself. There is no third-party adjuster with an incentive to deny claims, because there is no third party. Payout conditions are pre-committed in the smart contract before the event, not adjudicated case-by-case after.
Third, retroactive security. Even with the best pre-emptive defenses (audits, formal verification, access controls), patient attackers with operational security eventually find a path through. Pure pre-emptive security has a structural ceiling. Retroactive security accepts that ceiling and makes the successful attack unprofitable in retrospect. When a bad actor is flagged, taint propagates through the transaction graph; rational counterparties refuse to interact with tainted addresses because their own attribution gets contaminated downstream. The defector does not get punished, they get geometrically uncoupled from the honest counterparties’ value flow.
Get Blockchain Philosophy’s stories in your inbox
Join Medium for free to get updates from this writer.
Subscribe
Remember me for faster sign in
Fourth, layer separation enforcement. The competitive layer (arbitrage, liquidity provision, priority bidding, plugin innovation) generates revenue that funds the cooperative layer. The cooperative layer (Shapley distribution, insurance pools, treasury reserves, loyalty programs, retroactive clawback) provides stability that makes the competitive layer viable for participants who are not professional extractors. Each layer feeds the other. Neither contaminates the other.
4. What the mechanism design produces
Once all four moving parts are in place, the system has a property that classical economics calls incentive compatibility: cooperation is the individually-rational strategy, not because of altruism, but because the math makes cooperation more profitable than defection.
The whale who arrives for one block to farm rewards earns a high direct-contribution score and near-zero scores in the other three dimensions (time-in-pool, scarce-side, stability-during-volatility). The smaller liquidity provider who shows up for six months, supplies the scarce side of a lopsided market, and stays during a crash earns lower direct scores but higher composite Shapley values. The math rewards the loyal LP more than the mercenary whale, automatically, without any committee or voting process.
The MEV searcher who pioneered sandwich attacks finds that uniform clearing prices in batched commit-reveal auctions provide nothing to sandwich. The searcher’s profit from sandwiching converges on zero by construction. The same dollar value that used to flow to MEV searchers now flows to priority bidders, who pay it into a transparent treasury that funds backstop liquidity and impermanent loss protection. MEV is not eliminated. It is redirected.
The DAO whale who used to control governance proposals finds that math-enforced invariants (Shapley uniqueness, supply conservation, value-at-risk bounds) cannot be overridden by a 51% vote. Governance retains discretion over parameters within math-bounded ranges. It does not retain discretion over the fairness axioms themselves. Capture becomes a smaller problem because there is less to capture.
The system that emerges is not utopian. It still has bad actors. It still has failed transactions and bugs in code. What it does not have is a dominant strategy of extraction, because the mechanism design forecloses that strategy at every layer simultaneously.
5. Why this is the last option
If cooperative capitalism is the philosophy structurally left, what happens if it also fails?
The honest answer is: probably nothing better, for crypto specifically. The four already-tried philosophies covered the obvious axes. Permissionless or permissioned. Mutualized or individualized. Voluntary or enforced. Centralized or distributed. The cells in the design matrix are mostly filled. The cooperative-capitalism cell is the one where cooperation is mutualized at the risk layer, competitive at the value layer, enforced by mechanism design rather than by humans, and accessible permissionlessly.
If this design also fails to produce non-extractive outcomes, the structural argument suggests that the problem is not the philosophy but the substrate. Either crypto’s incentives are unrecoverable, or the philosophy is being implemented wrong, or both.
The implementation work, fortunately, is technical and falsifiable. It does not require anyone to agree with a moral framing. It requires shipping the mechanisms (Shapley distributors, mutualized insurance pools, retroactive clawback registries, layer-separated treasuries) and seeing whether the resulting equilibrium produces less extraction than the alternatives. If it does, the philosophy is vindicated by data. If it does not, the next philosophical iteration begins.
What is not honest is to keep building extractive systems while telling users they are decentralized, or to keep building gatekept systems while telling users they are protected. Both of these have been tried at scale for ten years. The data is in. The space deserves better, and it deserves a system whose properties can be falsified rather than asserted.
6. What you can do today
If you are a builder: design every economic mechanism through the layer-separation lens. Identify which axis of your protocol benefits from mutualization (catastrophic loss, fair attribution, collective reserves) and which benefits from competition (price discovery, capital efficiency, innovation). Do not let them contaminate each other. Use Shapley values for distribution, not pro-rata-by-capital, because pro-rata-by-capital is the mechanism that lets whales dominate by definition.
If you are a participant: ask whether the protocols you use have a coherent answer to “where does extraction get profitable here?” If the answer is “in transaction ordering,” look at whether they use commit-reveal batching. If the answer is “in governance,” look at whether their fairness axioms can be overridden by majority vote. If the answer is “in the insurance claim process,” look at whether claims are adjudicated by a denial-incentivized third party or by a pre-committed smart contract.
If you are a researcher: the open problems are real and tractable. The Shapley axioms are a starting point, not a closure. Iterated Shapley games have a fixed-point question that is open. Identity-splitting attacks on Shapley distributions need a formal lemma. The boundary between competitive-layer arbitrage and extractive-layer manipulation is not always sharp. There is plenty of math to do.
If you are an investor or an allocator: the question is which protocols have actually built the mechanism design rather than just talking about it. The mechanism either exists in deployed contracts or it does not. If a project’s “cooperative” claim cannot be reduced to specific named contracts and tested invariants, the claim is marketing. If it can, the claim is engineering and can be evaluated as engineering.
7. Closing
Cooperative capitalism is the design with one rule and four moving parts. Mutualize the risk layer permissionlessly. Compete on the value layer. Use mechanism design to enforce the boundary between them. The four parts are Shapley fairness, mutualized risk pools, retroactive security against successful exploits, and layer-separated funding that loops competitive surplus back into cooperative stability.
Every other economic philosophy crypto has tried at scale has produced extraction in a different form. The pattern across ten years is consistent enough to call structural. Cooperative capitalism is the philosophy that has not been tried at scale. The argument for trying it is process-of-elimination, not ideology. The argument against trying it is harder to construct, because there is no longer a fifth option waiting in reserve.
The mechanism is buildable today. The math is not new. The contracts exist. What is missing is the will to ship the design as a single coherent stack instead of as a collection of optional features that each protocol can adopt or ignore. The next ten years of crypto will be defined by which projects ship the coherent stack and which projects continue to redistribute extraction.
The risk layer is where we are most alike. The value layer is where we are most different. The mistake of every previous philosophy was conflating the two.
Sources: VibeSwap research papers (cooperative-capitalism.md, from-mev-to-gev.md, airgap-problem-onepager.md), 2026–05 retroactive security framing, ongoing public discussion in DeFi security circles. All deployed mechanisms referenced are open-source.