Why I'm Advising 10zeros

I spent the last few months with a spreadsheet of career options rated by impact, joy, flexibility, and money. A neurotech company. A drug lab solving suffering. An AI product studio. A $25B foundation. And a visa SaaS company called 10zeros.

The visa company won. Here’s why.

Back to building

For the last few years, I’ve been CEO’ing a research lab and co-founding a venture firm. Important work, but I noticed something: I wasn’t building anymore. I was orchestrating. Writing emails, managing relationships, setting strategy — all necessary, all slowly pulling me away from the thing that made me good at any of this in the first place.

My first freelance work was as a concept artist for games. I ran a graphic design studio. I was art director of an indie game studio. I founded Apart Research and built it for four years. Every time I’ve been most alive, I’ve been making things.

10zeros is a return to that. I’m advising on product design and development — full-stack, hands-on, shipping things that users touch. The others lead on CEO and CTO responsibilities while I get to obsess over the interface, the onboarding, the moment a founder realizes their visa is going to be fine.

The coordination thesis

The deeper reason is philosophical.

We are handing our civilization’s hardest problems to AI systems because we’ve given up on humans solving them together. Every AI lab CEO has some version of the same belief: “We can’t coordinate well enough, so we have to build the good AGI and hope.” I think that’s a failure of imagination.

Humanity’s actual superpower is coordination among genetically unrelated individuals. Nothing in the universe, before or after us, has been better at this. Constitutional conventions, abolition movements, the Meiji Restoration — these were feats of organizational intelligence that no single genius could have achieved alone.

The problem is that we can’t program organizations. We can’t run experiments on how teams work together at the speed we run experiments on neural networks. The cycle time of improving human coordination is too slow compared to the rate at which AI is improving.

10zeros is a bet that this can change. Through software that actually touches how people work together every day.

Why visas?

Visa compliance requires real-time visibility into pay, roles, locations, work content, and relationships across an entire organization. To do visas well in software, you end up building the most accurate system of record about a company’s people that has ever existed. And once you have that data flowing, you can build things that no HRIS has ever built — because you’re starting from a foundation of genuine, verified truth about what’s happening with real people.

The practical impact is immediate. When I ran Seldon Lab, one of the biggest value-adds was getting people visas fast through the 10zeros founders. Almost every person I talk to in the AI safety ecosystem is limited by their visa quality. A streamlined process saves 2-3 months per person. Even conservatively, that’s hundreds of person-months recovered per year in a field that is running out of time.

One of our customers, Manifund, has described their 10zeros subscription as one of the most impactful things they can do. Removing a bottleneck on talent concentration has compounding effects.

What I’m actually doing

10zeros has an AI agent called Tenzy that calls every team member for ten minutes a week. Those conversations surface compliance issues, benefits confusion, brewing retention problems — things that traditional HR systems only catch when someone remembers to file a report. Combined with deep integrations across company systems, Tenzy builds a picture of what’s actually going on with your people.

The vision is a system of record that is proactive. One that talks to the employee. Where the hardest parts of running a company — immigration, compliance, payroll, trust — are handled by software that actually keeps up.

I’m designing and building parts of this. The kind of work where I get to care about whether the onboarding feels magical, whether the homepage inspires, whether the product is something I’d be proud to show my 15-year-old self.

The real reason

If I’m honest, the decision came down to something simple.

I listed all the options and asked: which version of myself is most alive in six months? The answer was the one who shipped a product people love, reconnected with design and code, and helped build something that makes the world’s fastest teams work better together.

The thesis that human coordination can be engineered, measured, and improved at the speed of software is either wrong or one of the most important ideas of the decade. I’d like to find out which.