After the collapse of the dot-com bubble, a new cybernetic model emerged. People commonly call it web2.0, often characterized by cloud-based applications and user-generated content. In practice, it meant a huge wave of centralization around American megacorps — the FAANG(Facebook, Apple, Amazon, Netflix, Google) cohort. They were colonizing global cyberspace by harvesting data and attention, packaging them and distributing wholesale. Backed by the most powerful nation-state, it created a system capable of organizing millions of workers, billions of users, and trillions in capital. This was the most efficient large-scale coordination machine ever built. The situation was win-win for all parties: users received free endless entertainment, Nasdaq numbers went up, and US agencies got all the intelligence they needed.

But after a number of whistleblowers questioned the integrity of the system, an alternative architecture was proposed — web3. A counter-current attempt to bring the world wide web back to its roots, to make it a global, permissionless network again. Detachment from nation states, neutrality, and decentralization were meant to transform the web with

• Cryptocurrency economies

• User-owned data and protocols

• DAOs as the coordination layer

Much of the vision was realized, but web3 architectures didn't reach the mass consumer yet. Web3 is going through a structural crisis and the DAOs were hit the hardest. The reasons are worth restating:

• Voter apathy. Governance as tragedy of the commons.

• Idealism as a blueprint. Systems designed for how people should behave, not how they do.

• No institutional memory. Every DAO fork reinvents the same failures.

• Treasuries as targets. Capital draws extractors who leave when the extraction is complete.

• Price as the only signal. When the chart trends down, the organization dissolves.

The prospects of DAOs seemed dark with the accelerated adoption of the Transformer architecture and LLMs — commonly generalized to AI. But AI could just as easily shift the balance of power back toward decentralization. We could be standing on the eve of the web3 renaissance.

LLMs dethroned FAANG as the center of cyberspace and changed the status quo. MANGOS (Meta, Anthropic, Nvidia, Google, OpenAI, SpaceX) brought major disruptions. The signs are everywhere — from changes in the Bay Area real estate market to ex-NSA officers joining OpenAI. While a joint-forces pact between frontier labs and US government is still under negotiation, American closed-source AI models are clearly the New Big Thing.

But for the first time in decades, American tech dominance has genuine competition: DeepSeek and the Six AI Tigers (大模型六小虎). The tension is real, and the latest international policy disputes clearly show the priorities of the two biggest national economies.

At the moment when the general consensus was the MANGOS are de facto the new center of global cyberspace, DeepSeek made a 5-D AlphaGo move by open-sourcing their frontier model under the MIT license, effectively setting a new standard for Chinese frontier AI models. And they did it while being sanctioned.

This created the legal clearance for an entire ecosystem of intermediaries to build on top of it. Venice, for instance, could take the weights, run them on infrastructure rented from DePIN networks, and sell private inference for crypto.

The irony is sharp. AI from a more centralized state may be what sustains the decentralized web.

This paradox has a blueprint. In 1994, architect Kisho Kurokawa called it the "third space." True symbiosis, he argued, isn't just two rivals tolerating each other in tension. It is the creation of an intermediary zone where both forces collide and transform into something completely new. Chinese open weights and web3 compute markets create this third space between the two worlds most powerful digital empires. The architecture is locking into place:

• Open weights

• Private inference

• Permissionless market

• Crypto payments

A stack that belongs to neither licensor nor licensee, neither state nor platform, but to the network.

This space is a substrate, a layer where new kinds of entities can emerge — forms that could not exist in either originating system alone. The obvious inhabitant is the DAO, reconsidered.

AI agents will soon transform every human organization in the world, including DAOs. Agents can work around the clock, handling treasury management, knowledge extraction, human interaction, and more. But should a DAO trust a centralized entity with its private keys? Should they also trust it for inference? A malicious inference provider could add back doors to smart contracts, seize a DAO's treasury, doxx participants, etc.

This is where the third space enters. A DAO could be supercharged with private inference bought with crypto on an open market, store its knowledge on IPFS, and rotate agents on web3 clouds.

"And where does the newborn go from here? The net is vast and infinite" (c.) Motoko Kusanagi

The symbiosis between a DAO and private inference could bring to life a new type of entity — one that doesn't fatigue, doesn't lose interest, doesn't extract and leave.

What happens when the most efficient organization on the internet is neither a corporation nor a DAO, but something growing in the space between them?

Daoism Systems is planting a seed in that space — Everstrat.