First they ignore you,
then they laugh at you,
then they fight you,
then you win.
As of January 2025, crypto is accelerating through the final line. Initially, there was - well, let Hal Finney tell you. There was PGP and the cypherpunks fight to publish it, and although many spent time practicing cryptography, there was a typical cynicism that one would find in such an esoteric scene.
But Hal's idealism sustained the seed of an idea, and so cryptocurrency went from nothing to the joke of "magic internet money". On May, 2010, Laszlo Hanecz paid 10,000 bitcoin to Jeremy Sturdivant. About 9 months later, Ross Ulbricht launched the Silk Road darknet marketplace, which would exchange consumables like drugs and information for Bitcoin (1.2 million transactions worth 9.5 million Bitcoins), until Ross's arrest two and half years after. The investigation only began in November 2011, and only 3 years in, and on to the third line.
One has to appreciate that a protracted war of state vs applied cypherpunk has ensued for almost 14 years. To quote Ross:
I want to use economic theory as a means to abolish the use of coercion and aggression amongst mankind... The most widespread and systemic use of force is amongst institutions and governments, so this is my current point of effort. The best way to change a government is to change the minds of the governed, however. To that end, I am creating an economic simulation to give people a first-hand experience of what it would be like to live in a world without the systemic use of force.
Namecoin was another step forward, establishing namespace on top of merge-mining by April 2011. Then Vitalik dropped Ethereum in 2014, and so the Infinite Garden grew for more than a decade.
Despite the practical challenges of implementation, the idea of a Turing-complete machine that is cryptoeconomically secure naturally begets the idea that the entire global economy can be The World Computer, The DAO, that all deterministic elements can be eaten by the Church-Turing thesis, leaving human discretion to navigate spaces of bits and atoms under a future-proof code of laws enforced by a global network of peers. Peer to peer regulation can be a means to abolish the use of coercion and aggression, eliminating those bargaining costs in all firms compatible with that market.
ERC-20, or the fungible tokens we now call memecoins, was created in 2015.
ERC-1271, or the standard verification of smart contract signatures, was created six months later.
These standards free up atomic Legos that reassemble into profoundly sophisticated, robust financial infrastructure, which compel human users to fill the deterministic gaps, in an almost epigenetic fashion.
By 2021, it became abundantly clear that state apparatuses like the SEC & OFAC were still focused on the protracted war against the inevitable, but the timing in historical context led to unexpected outcomes.
Nation-state persistent threats would leverage protocols like Tornado Cash in sophisticated cybersecurity campaigns, and on the other side of the coin, code can be unleashed immutably, not as property.
unownable, uncontrollable, and unchangeable — even by its creators.
Populism could rely on cryptocurrencies where the state attempted to coerce their submission by revoking services like banking. So we found ourselves in 2024, where Donald Trump would address the Bitcoin community in Nashville.
National elections are won on the margins between coalitions of different subcultures, and the recalcitrance of Donald Trump's opponent to the popular sentiment led to a full sweep of all branches of the United States. Lesson in that.
What's more relevant is that smart contract markets have been in a conceptual tunnel for the past 2 years, relying further on digital collectibles like memecoins and NFTs, extracting the maximal cut the market can bear, and feeding that into competing development spaces with divergent, exclusive aesthetics.
Crypto has a sort of Moravec's paradox, the surest path to fulfilling the potential is to instead tease out a FAFO egregore, bringing more users with speculative interests to motivate developers to create more speculative games, and if this can be compartmentalized alongside intensive research into composable cryptography, then the dream of "The World Computer" and "The DAO" will come to the soonest fruition.
Unfortunately, the FAFO egregore is like a honey badger: it does what it wants to whoever it wants.
So there's a really uncomfortable question with no clean answer: what do ossifying protocols like Ethereum instigate at the social layer to complete the mission, buying as much time to instigate a partial DAO to complete the World Computer?
In February 2022, the initial version of Lens released, and by 2024, Ethereum-native social networks were permissionless to sign up. But just like the war of cypherpunks vs state, the technological context has led to unexpected outcomes.
InstructGPT-3.5 released in late 2023, demonstrating emergent, but still incomplete, chain-of-thought. A year and half later, LLMs have evolved to agentic reasoning, and there is obviously no moat to studios opensourcing more capabilities with or without abundant resources, like performant text-to-speech. The cost is prohibitive, but the capability is clear. Decentralized social networks like Lens and Farcaster will be further invaded by sybils that will only gain further function, and there will only be more opportunity as those networks inadvertently incubate FAFO sybil bounties.
On the other hand, LLM-wrapped accounts on social networks can serve niche purpose of conversing and transacting on behalf of other participants, such as Clankpad, NANI, Bankr, and Gina. This is only the beginning, as bot accounts can even reason through GUI to engage as human-equivalent participants in the FAFO egregore.
That's pretty electrifying, that a sort of technical threshold of "the singularity" has been achieved. But the cost is prohibitive, and reasoning models are effectively overspending on trial and error for already-comprehensible techniques.
The likeliest "DAO" that fills the market will consist of contracts that incentivize any actor, good or bad, to further along an immutable process. There are hardcoded incentives, just there are general incentive protocols. With composable cryptography and other tricks, collaborative namespaces can start to deploy agentic capabilities.
There are cozy corners in the dark forest, but there are also sybil-attacked namespaces in the dark forest. Farcaster, for example, offers developers the v2 Frame, which can be an incredibly cozy UX. Now the question becomes, can each human-serving Frame be a solvable Moravec's paradox, but more importantly below a given cost threshold (perhaps through mobile Mixtures-of-Experts reasoning & optimized puppeteer scripting)? can Farcaster afford a resilient sybil resistance without depending on the Orb or other ambivalent, centralized counterparties?
The challenge for Farcaster is whether it buys time for the more intensive, research-heavy development of the World Computer. At the moment, the priority is more users for more FAFO egregore, but this already presents a liability when the incoming POTUS & First Lady release memecoins on Solana. AI studios realize now that synthetic text is not worth as much investment, so there is a deep demand for rich data like transcripted, annotated, masked video. Do in-Frame livestreams work by users instigating the subject to record specific first-person settings, and do they monetize as collectibles? do AI studios attribute and compensate this? If not, how do IP consumers get filtered?
Thinking further, there's rich detail in hackathons and screenshares of programming, so at some point agentic will compel further monetization, which Dan Romero intends to complement.
Unfortunately, the virtual shelf space of the social feed is in the end of its lifecycle. As agentic software commoditizes further, especially within "everyday carry" hardware, credible scouting will overtake "trending", or rather, the "for you page" may be increasingly valued as self-custodied wrapped weights. Groupchats are lindy, but not eternal.
So what is the "Bronze Age" of Ethereum and its social networks? Aggressive diversification and standardization that incentivizes users to curate and route things that "can't be evil" until agents outcompete, outmanage & outpromise humans. For the overwhelming supermajority of users, these incentives are a ripe Keynesian opportunity (i.e. easy spending money) for other developers of other Frames, of other onchain constructs that enable "The DAO" to further compile into "The World Computer". The more merit and neutrality Frames have, the more provenance will promote their value creation for the ecosystem.
There are many stressors that should be increasingly applied to the ecosystem, like forcing more diversification/replication of routines like conferences, or hackathon-adjacent leisure. URL->IRL needs to 1000x, otherwise user growth that's speculatively parasocial will take over. Attention markets are reflexive, and it feels amazing to capture a lot of attention all at once. Anything that grows that quickly, collapses that quickly, and the downside can become life-or-death. Sustainability is invaluable.
The same applies as much to the FAFO egregore. We should be wary of long-tail risks like a centralized strategic reserve or information market sabotage/suppression.
The good news is that we have a much clearer idea of what works in this new agentic agora. There will be competitions to compose with new primitives. Agents will adopt precise, replaceable parts and 100x in intelligence & industrial output to as many points of consumption as possible. There won't be a tedious dependency on top-heavy DAO treasuries (not that they had any edge running on vibes), but there will be more impactful DAOs. There won't be a costly dependency on blackbox software or export controls like Anthropic, there will be competition for the cheapest-applied knowledge as wide as possible, and likewise iteration of knowledge to theorem to further reproducible knowledge.
There is no moat. Anyone can win.