Alex Blania on Worldcoin's US launch: 12M verified users, racing to build proof-of-human before AGI makes the internet untrustworthy

Summary

Alex Blania, co-founder and CEO of Tools for Humanity, makes the case that proof-of-human infrastructure needs to exist at internet scale before agentic AI makes it impossible to tell who — or what — you are interacting with online.

The company, started roughly five and a half years ago with Sam Altman, built its thesis on a single uncomfortable premise: software-based identity checks will all eventually break down as AI improves. Online reputation graphs, government ID systems, phone cameras — Blania argues each of these fails either because AI can spoof them or because they lack the global, anonymous, privacy-preserving properties the problem demands. The conclusion they reached early on was that a physical hardware device, distributed globally, was the only durable answer. That device is the Orb, which scans a user's iris to generate a verified proof-of-human credential stored locally on the user's phone — not on a central server.

Scale and US launch

Worldcoin now has 25–26 million total users in its World App and 12 million verified users across South Korea, Japan, Europe, and Latin America. Tools for Humanity employs around 400–500 people. The company launched in the US two weeks before this conversation, though Blania describes the domestic operation as still small and expects several months to ramp properly. The delay was deliberate: with Altman as co-founder, the project attracted regulatory scrutiny, and the team chose to wait for clearer SEC posture before entering the US market.

Why crypto

The token structure solves a bootstrapping problem. A proof-of-human network is worthless at subscale — the same dynamic that made early PayPal dependent on its costly referral program. Blania's argument is that a crypto network, if it works, can be valued in the billions, which means the protocol can afford to distribute meaningful ownership to early verifiers via the WLD token before the utility is fully there. Every verified user receives a share of the network. That's both an incentive to join early and, eventually, a business model for the network itself rather than just for the company.

Commercial traction

Blania says the network is moving from token incentives toward real utility. Match Group — the parent of Tinder and Hinge — has announced a partnership to bring proof-of-human verification to its dating platforms. Razer, a major gaming hardware manufacturer, is integrating World verification into its webcams to authenticate live video streams, using a signed face image stored on the user's device to confirm the person on camera is not a deepfake. Blania also says that in just the last six months, CEOs of large consumer platforms have been reaching out proactively — and that the current bottleneck is deployment capacity, not demand.

The bot problem

Blania references a University of Zurich study in which AI systems were used to respond to Reddit comments and measurably shift the opinions of entire subreddits. He treats this as a preview, not a peak. His estimate is that current bot activity represents roughly 1% of what will exist within 24 months, with the arrival of a capable open-source agent as the likely inflection point. His prediction is that well over 90% of internet content will eventually be AI-generated, forcing a fundamental inversion: the default assumption shifts from trusting what you see to distrusting it unless verified.

Centaur model and agent delegation

Blania is relaxed about AI-assisted human interaction — someone using an AI writing tool while dating online, for instance — framing it as a natural extension of how people will operate. The harder problem is scale asymmetry: one human deploying thousands of AI accounts simultaneously. His proposed solution is agent delegation, where a verified human credential can authorize one agent to act on their behalf, but cannot be used to spin up thousands.

AGI timeline and system durability

Blania puts true superintelligence — the kind that could synthesize a fake human iris or emit multispectral wavelengths to fool the Orb — at least 10 years out, and possibly further before society would permit it to interact with the physical world. His working assumption is that the Orb-based system remains technically durable through the most critical transition window. He also acknowledges the adversarial dynamic openly: as AI gets better at spoofing biometrics, defensive systems will improve in parallel, and the side with more compute and resources will build stronger defenses. He frames the net outcome as likely positive.

On government-issued digital identity as an alternative, Blania warns that mandatory government ID for internet access creates surveillance infrastructure that could track every page visit — a trade-off he thinks parts of the world will make badly. His stated ambition is for World to occupy the same trust position Signal holds in messaging: anonymous, high-trust, and privacy-preserving by construction.

With inbound demand outpacing deployment capacity and the US market just opening, the immediate pressure is operational — getting enough Orbs deployed fast enough to matter before agentic AI makes the verification problem substantially harder to solve.