Rational actors

Throughout college, the work I enjoyed most was coding side projects and studying microeconomics. Microeconomics rests on an assumption that’s rarely questioned—that humans are rational actors who make the decisions with the highest expected utility.

But that’s not always true: we buy Advil instead of generic ibuprofen, we sign up for free trials and forget to cancel them. Imagine if every person had an agent acting in their best interest, they could make better buying decisions and win back time to do more important things. At Natural, we’re building the infrastructure that lets those agents actually transact.

The break

I joined Natural as a software engineer in August. Before that, I went through YC with a sales tech company, then took a three month break to figure out what was next. I was a bit jaded, slightly burnt out, and pretty confused about what I wanted. I spent the break building an AI vocal coach to teach people how to sing. It was effectively an AI tutor: it did real work for people, coaching them through something a human would usually charge for. I kept thinking: what if the AI vocal coach charged the user like a real human vocal coach?

While unemployed but still paying San Francisco rent, I turned my spending into an optimization game. I used AI to find the highest value products for everything, from high-quality generic protein powder to the softest bedsheets for the price. I’m not an incredibly frugal person, but there were certain purchases like trash bags and laundry detergent that I simply didn’t want to be making. There was a small set of purchases I wanted to complete control over and a long tail I was happy to hand off to something that would optimize for the best product at the best price.

Finding Natural

When I started looking for what was next, I wanted three things: stability, a place that felt like home, and a hard open-ended problem. I went looking for startups working on interesting marketplace and economics problems. After some digging, I found Natural and emailed Kahlil. During our first conversation we talked about running Instagram ads - LTV, CAC, and how you could "graph the demand curve" for a product. It felt like Natural was working on two things I loved the most: building and economics.

The onsite

The first time I met Walt, we spent most of the time lightly sparring over the company thesis. I loved that he engaged seriously with my slightly uninformed ideas instead of brushing them off. The onsite was four days of the same energy: he would hand me a messy, open-ended problem that I needed to reason about from first principles. The one that stuck was how agents get access to your card and check out on your behalf in a way that’s legal and compliant.

While working through these questions, I kept noticing how much financial infrastructure this small team had built in under a year, and simultaneously how close knit the team was—friends who even ate dinner together every night. It was the home I’d been looking for, attached to the problem I’d been looking for.

What I’m working on

I’m now working on the infrastructure that lets agents check out on your behalf with a card. It’s a huge regulatory and compliance problem, completely greenfield, and still somewhat unsolved. It also fills in the gaps from that summer of optimizing my spending. Now not only can an AI tell me which generic protein powder was the best value, but it can buy it for me too.

The purchases we keep

The future I imagine is one where purchases start routing through something like a lawyer that acts on your behalf and makes sure the decision is actually rational. The generic ibuprofen, the forgotten free trials, the whole long tail of products where the right answer is just a fact—all handled.

Every dollar an agent saves you, and every hour you don’t spend comparing detergents, is money and time you get back for the purchases that are actually fun—the ones you make just because they feel right. The expensive concert tickets or the nicer bottle of wine.

Economics assumes we’re rational actors, which we’re not. But now we get to build them—and get back to making decisions that matter.

If this sounds like your kind of challenge, we’re hiring: /careers.