Been thinking about this for a while now—what if AGI doesn't come from building one massive model, but from something more distributed?

Imagine countless agents working in parallel, learning at the edge, optimizing through robot fleet operations. Each one improving independently, yet coordinating as a system. That's fundamentally different from the centralized LLM race we're seeing.

But here's the real question that keeps me up: Who actually controls this?

If intelligence emerges from decentralized coordination, does the power shift too? Or does whoever orchestrates the network own everything? It's the same tension we see across Web3—technical decentralization doesn't automatically mean power distribution. The architecture matters. The incentives matter. Who sets the rules?

The future of AGI governance might look a lot like the future of blockchain networks. Same fundamental challenge: keeping distributed systems from reconcentrating power.