Decoding the Human Nervous System: In conversation with Thomas Reardon, CEO, CTRL-labs

In its largest acquisition since Oculus in 2014, Facebook just announced last night it acquired CTRL-labs, a 4 year old startup based in New York, for a reported $500M-$1B.

Coincidentally, CTRL-labs CEO, Thomas Reardon (who goes by Reardon) was our guest at Data Driven NYC just a couple of weeks ago. Reardon is a particularly compelling entrepreneur, and this was a fascinating fireside chat, where we dove into machine learning, neuroscience, VR and all sorts of cool topics.

CTRL-labs builds what it calls “neural interface technology”: algorithms that decode the activity of individual motor neurons and turns that into control over machines, thereby completely redefining the interaction between humans and machines. Because the technology captures your intentions without requiring any physical movement, you can do things that you could never do by moving, and you can start “imaging experiences where you would have 20 fingers… or 8 arms or legs”.

The video (below) is well worth a watch in its entirety, including the audience Q&A at the end, and I’ve jotted down a few notes as well, for a quick review.

The iPhone is a “Trump-level” disaster:

• Took us in the wrong direction: instead of empowering us, became an instrument of “human enslavement”, and the Apple Watch is even worse
• In that paradigm, the device “inputs to me, but I’m not outputting to it”
• Users are less creative, instead adapt to the device, “correct the autocorrector”

How it works:

• Core technology is surface electromyography
• One key breakthrough is to get to the individual level of single neuron
• Uses a lot of deep learning and sophisticated signal processing
• The only thing your brain does is turn muscles on and off… humans are incredibly good at using their brain to leverage muscles in a dynamic way, adapting to specific circumstances each single time – requires an immense amount of computation, without creating a feeling of cognitive load. What if you can just use that same mechanism to control machines?
• “Decoding the human nervous system is the mother of all machine learning problems”, significantly harder than predicting the weather.

Applications and business aspects:

• Just started shipping the developer kit, after 3+ years of development
• Thousands of developers on the waiting list
• Our own goal of course is “this is how you do everything”. There isn’t anything this is not good for, but as a startup we have to pick our initial targets
• Immersive computing (AR, VR) are obvious targets because there are no good controls
• Pervasive computing (IoT to the Apple Watch) is another interesting area – would offer universal control over time
• Robotics: was a surprise to us but lots of inbound
• In terms of go to market strategy, we are very developer focused (except for text where we may be develop applications)
• We interview every single developer applicant to understand what they want to build

Reardon’s background:

• Left school at 15, couldn’t afford college, so went to work leveraging his coding skills… Started a company at 19 that was acquired, left the company to join Microsoft
• In 1994, started the Internet Explorer initiative at Microsoft
• Helped found the W3C, was involved in the set up of the initial web standards, HTML, XML, CSS etc.
• Left Microsoft after 10 years, then went to college to study Greek and Latin and then did a PhD in neuroscience at Columbia