Episode #2 — CTRL-LABS

Solving Billion-User Problems With CTRL-labs

Illustrations by Joao Fazenda

At 19 years old, instead of going to college, Thomas Reardon went to Microsoft, where he built Internet Explorer and helped architect the world wide web. After that, he made an unusual move: He left tech and went back to school, where he studied classics and had his “ego utterly obliterated” while completing a PhD in neuroscience. (“That's actually a good exercise,” he says of the ego destruction. “I highly recommend it to other founders.”) Today, all of Reardon’s multidisciplinary experience dovetails into CTRL-labs, where he and a team of machine learning scientists are designing the next wave of paradigm-shifting technology: Gadgets that let humans control computers—with their minds. 

Genentech, but for neuroscience

I got to ride Microsoft from the beginning, when it was a 4,000-person company, to the world-dominating company that it was by the end of the 1990s. I got to bring to life something that went from zero to a billion users. That's a once-in-a-lifetime thing. It got me focused on billion-user outcomes; if the problem isn't a billion-user problem, I'm not interested in it. A lot of tech startups right now are based on a business model of economic arbitrage, where there's no real new product—it's just a different way of calling a cab or renting a house. There’s another way to go about this. I wanted to know: What would be a truly disruptive technology that could move people from an existing paradigm to something totally new?

There was this tremendous amount of neuroscience technology that had been developed over the last 50 years, but it was commercially unexploited. It felt like Genentech in 1975. Thirty years of crazy, academic science went into understanding DNA, but it took years before a company created a commercial model to exploit that science. And that had never happened in neuroscience. Science has built up a phenomenal understanding of the brain, and created a phenomenal number of tools for interrogating the brain, but nobody has exploited it. I felt like we had an opportunity to be Genentech, but for neuroscience and neurotechnology, and I expected an explosion of companies to follow us. There's been a boomlet of that, but it's not quite the explosion I expected.

“What would be a truly disruptive technology that could move people from an existing technology paradigm to something totally new?”

Every b-school businessman I've ever seen give a presentation starts out with some version of the pitch, we're going to do a slight update of an old technology, and it's going to be a breakout. If you look at our start, the arbitrage opportunity would’ve been in medical biotech, where everything is relative to healthcare outcomes. If somebody's lost a limb, or someone’s damaged something in the spinal cord, we can help them renew their lives. But that’s a million people, maybe 10 million. I was trying to flip that on its head and say, no, we're going after the mass market—the way to exploit this is to do something much bigger than a translation of past technologies. We turned down opportunities to take money that would’ve made us into a biotech company rather than, ultimately, a consumer tech company—which is what I wanted to be.

With CTRL-labs, Thomas Reardon set out to answer a big question: Can you take existing neuroscience research, and use it to create paradigm-shifting consumer tech?
“Like splitting the atom”

We started as a team of neuroscientists, with no ambition to do hardware whatsoever. We thought we’d have some loose partnership with another company that did gadgets. But we very quickly realized we had to care about hardware, after one of our founders brought up the analogy of the Wright brothers. They succeeded, when other novice aeronauts and airplane-makers failed, because they created wind tunnels. They could model planes and prototype things under controlled conditions, before going out to live conditions.

For the first couple of years, we just did wind tunnels. I studied motor neuroscience, which is the half of the brain that controls your muscles. We had an idea of how we could work with motor and neural signals, repurposing them so a person could control machines rather than just their own body. But these were just textbook-level principles. In the neuroscience world, to figure this out, we’d put sensors on somebody, collect hours of data, then go off and crunch that data for weeks. Instead, we said the real breakthroughs are going to happen when our wind tunnel—in this case, our algorithms—can operate in real time, and we can drive feedback to somebody to alter their neural output. We needed that closed loop system. 

“We very quickly realized we had to care about hardware, after one of our founders brought up the analogy of the Wright brothers.”

If I could show our decks from the first few board meetings, you'd see us saying again and again, "Okay, we're going to be in this invention-and-discovery phase for a little while, holding our breath, doing these wind tunnels," which I think was probably terrifying for our investors because they were like, "Well, where are the customers?” And we were doing the complete opposite of that. Our goal was to get to the neural signal, not the muscular signal, because at that point we can completely change the rules of how people control machines. You can literally move beyond your five fingers to a paradigm of 20 imaginary fingers that you can still control, with the same excellence and fidelity with which you control your natural five fingers. 

When we got to single-neuron resolution, we had the ability to actually see the electrical activity from a single neuron. As neuroscientists, that was crucial—we hit a level of atomic information where we can't reduce it any further. It took many years to get to that point, but we're unbound now. We can listen to a single neuron, and see all the messages from that neuron. What specific product does that equal? I can't tell you. But I can tell you it is like splitting the atom, and you just have to imagine a little bit what happens after that.

An early prototype of the CTRL-lab armband, which turns brain signals into computer interactions. 
Crazy science joins big tech

Multiple large tech companies were interested in buying us, and there was some bidding. The real decision wasn’t who to sell to, it was whether we should stay on our own or not. This was an incredibly difficult decision because our original plan was to raise a new round of $150 million, at a very high evaluation. It was very, very attractive to stay independent. Even now, when I look back on it, if we had stayed independent we'd be fine. At this time, in 2019, we were still deep in technology development, but we would’ve had a massive mattress of capital to fall asleep on each night. 

“What's the reward I can get for this team of machine learning scientists and believers who all are in high demand?”

The number one argument for selling was people, on both sides. I was attracted to the group that we'd be working with. It's an off-the-charts strong team; a crazy science factory. But for my own people, I just looked out and said, "Let me just do an expectation maximization here. What's the reward I can get for this team of machine learning scientists and believers who all are in high demand?" Should I wait and get them a reward in five years, or should we take this and get a big reward for the whole team right now? I got my own lottery ticket multiple times over at Microsoft. Selling was about the team—a bunch of people who took a big personal economic risk on me and on this mission. I felt like we could get the same outcomes on our own, or inside of big tech. So we’re continuing the mission inside of a bigger company.

Reardon (center) with his ace team of neuroscientists and machine learning scientists in CTRL-labs’ New York office

CTRL Labs: A platform for turning brainwaves into computer interactions
Founded: 2015
Initial Partnership: Led the Series A in 2017
Exit: Acquired by Facebook in 2019