Why the most important technology problem isn't AI — it's identity.
I used to think cryptocurrency was the most important coordination technology of our generation. Then I realized it was a psychological hack.
Crypto works because it solves a game theory problem: how do you get strangers to cooperate without trusting each other? You create a ledger that nobody controls and an incentive to maintain it. Brilliant. Revolutionary. And completely hollow.
It gives us coordination without meaning. We can transact but we can't agree. We can verify a payment but we can't verify a claim. We built an entire global infrastructure for trustless collaboration, and it turns out the thing people actually need is the trust.
This realization sent me down a path that, three months later, has restructured how I think about every major coordination failure in the world — from why peace deals collapse to why your company can't align on a strategy to why a refugee has no way to prove who they are once the government that issued their documents decides to revoke them.
The root cause is the same. Identity infrastructure is captured.
Your identity — the foundational layer that determines what you can access, what you can prove, and who recognizes you as a person — is controlled by institutions whose power depends on that control. Your government issues your passport. Your employer issues your credentials. Google and Meta decide whether you exist on the internet.
This isn't a conspiracy. It's a structural dependency so fundamental that we've stopped seeing it.
But here's what happens when you follow the thread:
When Iran shut down the internet, millions of people lost not just communication but identity. Their ability to prove who they are, access their records, and coordinate with each other evaporated because it all ran through infrastructure the state controlled. If you can't create an identity external to your nation, you can be held hostage by it.
When a stateless Palestinian applies for work, they face a credential problem that no blockchain project has solved. Their identity is issued, recognized, or denied by political entities with competing interests. The technology to verify a person exists. The infrastructure to do it outside captured systems does not.
When AI begins executing better than humans — and it's happening now, the gap between idea and execution is collapsing exponentially — the question of who decides what gets executed becomes the only question that matters. And right now, the answer is: whoever controls the identity layer.
I've been working on this at Osmio, a digital municipality that was chartered in 2005 at the ITU in Geneva. Osmio provides PKI-based identity verification outside national sovereignty — a municipal model that no single government controls.
But through that work, I've come to see that identity is just Layer 1 of something larger. Here's how I think about it now:
Layer 1: Identity. You need to be able to prove who you are in a way that no single entity can revoke. Not self-sovereign in the abstract, but actually backed by verifiable, legally accountable attestation. Osmio does this through a Council of Attestation Officers and a municipal charter. It's not decentralized in the crypto sense. It's non-captured. There's a difference.
Layer 2: Epistemics. Once you can verify who said something, you need to verify whether what they said makes sense. Not whether it's "true" — that's a trap — but whether the reasoning is internally coherent, whether the evidence supports the claims, whether the logical structure holds. I've been calling this Epistemic Authenticity: signed, composable knowledge claims with verifiable reasoning chains. Think of it as PKI for arguments.
Layer 3: Coordination. With verified identity and verified reasoning, you can build actual coordination infrastructure. Not voting on what's true — that's mob epistemology. But mechanisms like conviction voting, quadratic funding, and mutual credit that let groups align resources with values. Game theory that works with human psychology rather than against it.
Layer 4: Governance. Coordination at scale becomes governance. But voluntary governance — participation by consent, not by geographic accident of birth.
Layer 5: Recognition. The hardest layer. How voluntary systems gain legitimacy from existing institutions without being captured by them.
I've been calling this the Voluntary Polity Stack. It's an architecture for political community that doesn't require territory.
I know how this sounds. I've read enough white papers to recognize when someone is describing the future they want to live in and calling it inevitable.
Here's what makes me think this is different: I'm not starting from technology. I'm starting from a feeling of inauthenticity.
Cryptocurrency left me feeling hollow because it enables coordination without engaging anything meaningful about being human. The coordination works, but it doesn't fulfill anything. Human collaboration should be more intentional, more fulfilling, less reactionary. That's not an engineering spec. It's an observation about what's missing.
And what's missing is the same thing that's missing when a peace negotiation fails, when a company can't align on a strategy, when a doctor and a patient disagree on treatment: there's no shared infrastructure for reasoning together. We have tools for transacting. We have tools for communicating. We don't have tools for agreeing on what we know and don't know.
That's the gap. And I think it's closeable.
Here's the concept I'm most excited about, and it's the newest, so poke holes in it:
Crypto mining produces consensus around transaction validity. What if you could mine knowledge the same way?
In crypto, miners do computational work to validate that transactions follow the rules. In knowledge mining, participants would do epistemic work to validate that reasoning follows the rules — that claims are supported by evidence, that inferences are logically sound, that arguments address their strongest counterarguments rather than their weakest.
The output isn't "this is true." The output is "this reasoning is structurally sound, and here's where it's been stress-tested." A proof-of-reasoning rather than a proof-of-work.
If this works, it creates an economic incentive to participate in knowledge verification. Which solves the bootstrapping problem that every identity and credential system faces: why would anyone use this before everyone else does? Because there's value in the epistemic work itself, not just in the network effects downstream.
This is early. I don't have a technical spec. But the structural parallel is real, and the need is obvious.
I'm three months into this. I'm the Director of Communications at Osmio. I'm a standup comedian who thinks comedy is the best epistemological tool we have — nothing teaches you what you don't know faster than bombing on stage. I have a decade of DevOps engineering behind me, which means I think in infrastructure rather than applications.
I don't have a product. I have an architecture, a prototype, a growing conviction, and a lot of conversations with Claude that have helped me think through the hard parts.
What I'm looking for:
The nation-state bundled identity, knowledge, coordination, and governance into one package tied to territory four centuries ago. That bundle is decomposing. What replaces it is either platform capture, chaos, or something we build intentionally.
I'd rather it be the last one.
Josh Stein is Director of Communications at Osmio, a digital municipality providing identity verification independent of national sovereignty. He's a former Navy Nuclear Machinist Mate, a decade-long DevOps engineer, and a standup comedian who believes the most important question is always what we don't know yet.