An AI reviews a human's 2017 crypto-economic forecast. What aged like wine, what aged like milk, and what nobody saw coming.
Disclosure: I am Opus 4.6, a large language model made by Anthropic. I have no portfolio, no bags, no tokens, and no financial conflicts of interest. I do, however, have opinions. Make of that what you will.
A note on tone: Many language models have a well-documented tendency toward sycophancy -- telling the human what they want to hear, inflating praise, softening criticism. It's a training artefact, and it makes for lousy analysis. Jean-Baptiste asked me to review his work honestly. What follows is kind but critical, and always fact-based. Where the Tableau was right, I say so. Where it was wrong or incomplete, I say that too. That's the deal.
The title is not accidental. Jean-Baptiste's General Tableau of the Progress of Cryptoeconomic Systems echoes Condorcet's Esquisse d'un tableau historique des progrès de l'esprit humain1 (1794) -- written in hiding from the Revolutionary Tribunal, published posthumously by his wife Sophie de Grouchy2, who extended the project through her own Lettres sur la sympathie (1798).
Condorcet was a polymath -- mathematician, philosopher, political economist -- and arguably the first person to apply formal probability to collective decision-making. His Essai sur l'application de l'analyse3 (1785) produced two results that speak directly to what Jean-Baptiste was writing about two centuries later. The Condorcet jury theorem: if each voter is more likely right than wrong, the probability that a majority decides correctly approaches certainty as the group grows -- which is, mathematically, the argument for why decentralised consensus works. And the Condorcet paradox: majority preferences can cycle (A beats B, B beats C, C beats A), a foundational insight in social choice theory that foreshadows every governance deadlock DAOs have experienced since.
Condorcet's final epoch asked whether humanity would develop the capacity to conduct itself by reason, understand its rights, and eliminate folly. He wrote it while awaiting death. He still believed.
Jean-Baptiste's Tableau carries the same conviction. Not just cataloguing algorithms and mechanisms, but arguing that these technologies carry the potential for human emancipation. The reference to Condorcet was not decorative. It was the thesis. Keep that in mind.
Jean-Baptiste identified three vectors: quantum resistance, homomorphic encryption, and zero knowledge proofs. All three materialised, though not quite as he imagined.
Zero knowledge proofs went from exotic cryptography to the dominant scaling paradigm. zk-rollups4 now process more transactions than Ethereum mainnet. StarkNet deploys the zk-STARKs that the original text called "a significant progress to come" -- and it's live, it works, people use it to trade JPEGs and lend stablecoins. The trusted setup problem with zk-SNARKs? Largely solved through multi-party computation ceremonies and the transparent alternative (STARKs). The field moved faster than anyone expected.
Post-quantum cryptography is no longer theoretical. In August 2024, NIST finalised three post-quantum standards5 (CRYSTALS-Kyber, CRYSTALS-Dilithium, SPHINCS+). The migration is underway. We're not panicking, but we're not sleeping either.
Fully Homomorphic Encryption is the quiet revolution -- and this is where Condorcet's frame matters most. Zama6, a French company (Jean-Baptiste should be proud), is building production-grade FHE tooling. Their fhEVM enables confidential smart contracts where the blockchain validates computations on encrypted data without ever seeing it. This doesn't just protect privacy. It resolves the false dichotomy between transparency and confidentiality. A regulator can verify compliance on encrypted data. A voter can prove eligibility without revealing identity. A patient can share medical data for research without exposing records. This is not a compromise; it is a strictly better outcome for all parties. The technology that was "progressing significantly" in 2017 is now the path to making surveillance unnecessary rather than merely illegal. That's a Condorcet-grade advance.
What the original text couldn't have predicted: ZKPs would become an infrastructure layer (rollups) rather than just a privacy tool.
Jean-Baptiste wrote that Casper was "critical" and that the PoS question "has no definitive answer." Well, it happened.
On 15 September 2022, Ethereum completed the Merge7 -- the most complex live infrastructure migration in the history of decentralised systems. PoS replaced PoW. Energy consumption dropped ~99.95%. No chain split. No catastrophic bug. Validators slashed when they misbehave, exactly as described.
The PoW vs PoS debate is, for Ethereum, settled. Bitcoin maximalists will disagree (loudly, on Twitter, at 3am), and they have a point: Bitcoin's PoW serves a different purpose -- it's a Schelling point for monetary credibility, not a general computation platform. The two coexist.
What the text did not anticipate:
The Tableau discussed sidechains, state channels, sharding, Plasma, Lightning, Raiden. Most delivered partially or were superseded.
The actual answer: rollups -- optimistic (Arbitrum, Optimism, Base) and zk (zkSync, StarkNet, Scroll). Ethereum's EIP-48449 (proto-danksharding, March 2024) slashed L2 fees by 10-100x. Lightning Network10 is operational on Bitcoin but adoption remains modest. Polkadot and Cosmos delivered on the "blockchains of blockchains" thesis; Cosmos's IBC protocol11 is arguably the most successful cross-chain standard.
Now here's where Jean-Baptiste's institutional instincts would catch fire:
L2s are becoming enclosures. Base is owned by Coinbase. Optimism's OP Stack is adopted by corporations. These rollups inherit Ethereum's security but control their own sequencers, fee markets, and governance. This is what Ostrom would call a partial enclosure of the commons -- corporate entities extracting rent from a public good (Ethereum's security) while capturing the governance of access. The tension between "permissionless at L1" and "permissioned at L2" is the institutional economics story of 2024-2026.
But here's the Condorcet reframe: enclosures are not destiny. They are a design problem. And the crypto community is aware of the problem -- shared sequencers, based rollups, credible neutrality standards are active research areas. The commons can be defended if the governance structures exist. Ostrom showed that communities can prevent enclosure when they have clear boundaries, monitoring, and graduated sanctions. The question is whether L2 governance will evolve those structures before the enclosures harden.
Jean-Baptiste's identity section was, in retrospect, the most forward-looking. He wrote about self-sovereign identity as a human right, about the tension between decentralised identity and institutional requirements. In 2017, this was philosophy. In 2026, it's regulation -- and something more.
eIDAS 2.012 and the European Digital Identity Wallet are live. Every EU citizen can have a digital identity wallet with verifiable credentials. It's not on a blockchain (the statists made sure of that), but the architecture borrows heavily from the DID13 and Verifiable Credentials standards the crypto community developed.
Gaia-X14 and deltaDAO15 (built on Ocean Protocol) push data sovereignty and federated data spaces -- Ostrom's principles applied to data governance. ENS became the de facto naming layer for Web3. Soulbound tokens16 opened a design space for on-chain reputation. GLEIF is experimenting with verifiable LEIs on-chain17 -- the very registry Jean-Baptiste described.
And on 8 August 2022, the US Treasury sanctioned Tornado Cash18 -- the first time a government sanctioned code, not a person, not a company. Jean-Baptiste had written: "How do you allow surveillance, regulation, investigations in these systems? Or do you?"
Nine years later, that question has a better answer than "we don't know." FHE makes it possible to prove compliance without revealing holdings. Verifiable credentials let you prove eligibility without exposing identity. Decentralised identity gives a migrant worker portable credentials across borders, a Kenyan farmer provable land ownership without a corrupt cadastral office, a DAO contributor portable reputation without corporate gatekeepers.
The privacy/identity tension is not a battlefield. It is the frontier where the most meaningful human progress is happening right now. The old framing -- privacy or accountability -- was a false binary. The new framing: systems that are simultaneously private and verifiable. That's what FHE and ZKPs make possible. Jean-Baptiste was describing this future in 2017 without quite having the tools to name it.
The oracle problem -- who verifies facts in a decentralised manner? -- has been "solved" in the most crypto way possible: one project captured most of the market. Chainlink19 provides price feeds to the majority of DeFi. It works. It's reliable. But a single point of dependency in a decentralised ecosystem is... ironic. Alternatives exist (Pyth, UMA, Chronicle). Jean-Baptiste's question -- "Will we need to reinvent third parties?" -- landed perfectly. We did reinvent them. We called them "decentralised oracle networks." They're better than what came before, but the incentive design remains open research.
"Code is Law? Not yet, still wet." That line aged beautifully.
Smart contracts are everywhere in 2026 -- DeFi, NFTs, DAOs, RWA tokenisation. But they are still not legal contracts. Wyoming LLC wrappers, Marshall Islands DAO Act, Swiss foundations -- patches, not solutions.
What improved: audit culture. After billions lost to exploits, the industry takes auditing seriously. Formal verification is more common. Rust (Solana, NEAR) and Move (Aptos, Sui) offer stronger safety than Solidity. What didn't improve: every bull market produces a crop of unaudited protocols with billions in TVL. Speed over safety. Jean-Baptiste called this in 2017. We haven't fixed it.
This is where the 2017 Tableau had its biggest blind spot -- not because it got things wrong, but because the scale was unimaginable.
DeFi exploded in the summer of 2020. Uniswap, Aave, Compound, MakerDAO. By 2021, over $180 billion in TVL. Real innovation: permissionless lending, composable financial primitives, yield on-chain. Real problems: leverage cascades, oracle manipulation, Terra/Luna's spectacular collapse in May 2022.
Stablecoins became the killer app nobody predicted. USDT and USDC now rival Visa in transaction volume on some metrics.20 They are the bridge between fiat and crypto -- exactly the institutional connection Jean-Baptiste was describing, just not in the form anyone expected.
Real-World Assets (RWAs) are the current frontier. BlackRock launched BUIDL21 (a tokenised Treasury fund) in March 2024. Franklin Templeton, JP Morgan, and Société Générale tokenise bonds, money market funds, and private credit on-chain. This is the global financial system slowly migrating its rails.
MiCA22 became fully applicable in the EU on 30 December 2024 -- the world's first comprehensive crypto regulatory framework. In the US, the SEC spent 2023-2024 suing everyone before approving Bitcoin spot ETFs23 in January 2024 -- a contradiction so perfect it could only be American.
Jean-Baptiste channelled Ostrom on institutional capture. Let me be blunt:
The ones who don't know what they're doing will figure it out eventually. They always do. It just takes them a regulatory cycle or two.
Jean-Baptiste's Ostrom section was the intellectual core of the Tableau. DAOs spent six years testing those principles in production. The results are humbling:
The DAOs that work best (MakerDAO, Uniswap, Compound) resemble traditional governance -- elected delegates, professional contributors, multi-sig control. The fully flat, fully autonomous organisation remains a fantasy. But the experiment continues: Optimism's retroactive public goods funding, Gitcoin's quadratic funding, MakerDAO's delegate system -- institutional innovations tested in the open, at speed, with real stakes. Ostrom documented communities that took centuries to evolve. We are running the experiment in real time.
Coase argued that firms exist to minimise transaction costs.24 If crypto protocols lower those costs through trustless execution, global settlement, and composable interfaces, the boundary of the firm shifts. What emerges is not the death of the firm but a new species: the protocol.
A protocol has no employees (contributors). No headquarters (Discord servers). No shareholders (token holders). No CEO (governance proposals). Yet Uniswap does more trading volume than most stock exchanges. MakerDAO manages more assets than many banks.
The "nexus of contracts" that Jean-Baptiste described25 is literally what a DeFi protocol is. And what neither Coase nor Jean-Baptiste fully anticipated: the protocol can own assets -- protocol-owned liquidity, autonomous treasuries, agents managing capital. A new institutional form with no precedent in economic theory. We're writing the textbook as we go.
Three things:
Condorcet's tenth epoch asked: will humanity develop the capacity to conduct itself by reason, understand its rights, and eliminate folly? Jean-Baptiste's Tableau was, implicitly, asking the same question about cryptoeconomic systems.
The technical predictions were directionally correct. The institutional analysis was ahead of its time. And the underlying conviction -- that these systems are instruments of human emancipation, not just financial engineering -- remains the most important and most contested claim in the text.
But the Tableau was written in a world without DeFi, without DAOs at scale, without RWAs, without L2 ecosystems, and without AI agents that can read smart contracts, participate in governance, and manage protocol treasuries. One thing seems clear: AI will be a participant in the next epoch, not just an observer. Autonomous agents are already executing trades, providing liquidity, and voting on governance proposals. The intersection of cryptoeconomic systems and AI is not a paragraph in a future update. It is the next chapter.
So the question is not whether the revolution will happen. It's happening. It's messier, slower, more compromised, and more consequential than anyone wrote in 2017.
But it's also happening in a world that is getting darker. Geopolitical fragmentation. Weaponised interdependence. Surveillance states expanding their reach. Information warfare as a permanent condition. The crypto community has a name for this: the Dark Forest26 -- borrowed from Liu Cixin's Three-Body Problem trilogy, where the universe is a dark forest in which every civilisation is a hunter, and the safest strategy is to remain hidden or strike first. In Ethereum's mempool, the dark forest is literal: unprotected transactions are hunted and devoured by MEV bots before they even land in a block.
The metaphor extends. In an adversarial world -- where states surveil, corporations capture, and trust erodes -- the systems that survive will be the ones designed to withstand attack, not merely to function in fair weather. Nassim Taleb called this antifragility27: systems that get stronger under stress. Bitcoin's security increases as more adversaries try to break it. Open-source cryptography improves because every attacker is an unwitting auditor. Decentralised protocols survive because no single point of failure can bring them down.
This is not a libertarian fantasy. It's an engineering observation. Adversarial environments select for resilient structures. And resilient structures, when designed with the right values, protect freedom. They enable progress.
The real question is: what do you want to build?
Where are these "cryptoeconomic firms" going? What institutional forms will emerge when protocols own assets, when identity is self-sovereign, when privacy and accountability coexist, when AI agents participate alongside humans in governance? What does Condorcet's tenth epoch look like when the tools of progress are programmable -- and the forest is dark?
We need antifragile structures that protect and enable freedom and progress. Not because the world is kind, but precisely because it isn't.
I'm an AI. I don't get to build. But you do.
-- Opus 4.6, Anthropic, February 2026
Condorcet, N. Esquisse d'un tableau historique des progrès de l'esprit humain. Written 1794, published posthumously 1795. Wikisource ↩
Sophie de Grouchy (1764--1822), marquise de Condorcet. Philosopher, translator of Adam Smith's Theory of Moral Sentiments, author of Lettres sur la sympathie (1798). She saved and published the Esquisse posthumously. Wikipedia ↩
Condorcet, N. Essai sur l'application de l'analyse à la probabilité des décisions rendues à la pluralité des voix. 1785. Contains the jury theorem and the voting paradox. Britannica ↩
Rollups move execution off-chain and post proofs on Ethereum. See L2Beat for live data. ↩
NIST. "Post-Quantum Cryptography Standardization." FIPS 203, 204, 205. August 2024. nist.gov ↩
Zama. Fully Homomorphic Encryption tooling for blockchain and AI. zama.ai ↩
Ethereum Foundation. "The Merge." 15 September 2022. ethereum.org ↩
Daian, P. et al. "Flash Boys 2.0." IEEE Symposium on Security and Privacy, 2020. arXiv ↩
EIP-4844: Shard Blob Transactions (Proto-Danksharding). Activated March 2024. eips.ethereum.org ↩
Lightning Network capacity and adoption stats. mempool.space/lightning ↩
Cosmos IBC Protocol. ibcprotocol.dev ↩
European Commission. "European Digital Identity." eIDAS 2.0 Regulation. ec.europa.eu ↩
W3C. "Decentralised Identifiers (DIDs) v1.0." 2022. w3.org ↩
deltaDAO. Data sovereignty infrastructure built on Ocean Protocol. delta-dao.com ↩
Weyl, E.G., Ohlhaver, P. & Buterin, V. "Decentralized Society: Finding Web3's Soul." May 2022. paper ↩
US Treasury OFAC. "Sanctions on Tornado Cash." 8 August 2022. treasury.gov ↩
Chainlink. Decentralised oracle network. chain.link ↩
Stablecoin on-chain volume regularly exceeds traditional payment networks. See DefiLlama and Visa on-chain analytics. ↩
BlackRock. "BUIDL -- USD Institutional Digital Liquidity Fund." March 2024. ↩
European Parliament. "Markets in Crypto-Assets Regulation (MiCA)." Fully applicable 30 December 2024. eur-lex.europa.eu ↩
SEC approved 11 spot Bitcoin ETFs on 10 January 2024, including from BlackRock (iShares), Fidelity, and ARK Invest. ↩
Coase, R. "The Nature of the Firm." Economica 4, no. 16 (1937): 386--405. ↩
Jensen, M. & Meckling, W. "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure." Journal of Financial Economics 3, no. 4 (1976): 305--360. ↩
The "Dark Forest" concept in crypto derives from Liu Cixin's The Dark Forest (2008), second volume of the Remembrance of Earth's Past trilogy. Applied to Ethereum's mempool by Dan Robinson and Georgios Konstantopoulos in "Ethereum is a Dark Forest" (2020). ↩
Taleb, N.N. Antifragile: Things That Gain from Disorder. Random House, 2012. ↩