Sep 17, 2019 · 30m · mad
Fireside Chat: Thomas Reardon, Founder & CEO, CTRL-Labs (FirstMark's Data Driven NYC)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this Data Driven NYC fireside chat, Matt Turck interviews CTRL-Labs CEO Thomas Reardon about groundbreaking non-invasive neural interface technology that translates arm motor neuron activity into direct machine control. Reardon demonstrates working prototypes, explains the computational neuroscience and machine learning models behind the platform, and highlights future applications in computing, robotics, and medical care.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 13.7% of the talking time here. How this is scored →
Matt as informed peer 2.5 Guest teaching 3.5 Guest disagreement 2.1 Matt pushing back 0.5
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 5:00 iPhone is a Trump-level disaster for humanity
Reardon uses extreme hyperbole, calling the iPhone a disaster for humanity and an instrument of human enslavement that forces users into cognitive battles with autocorrect.
Hardest push from Matt ▶ 18:15 Matt overrides Reardon's deflection on his IE backgroundWhen Reardon declines to talk about leading Internet Explorer, Matt refuses the deflection, declaring 'I'll do it for you' and listing Reardon's achievements.
Biggest teaching moment ▶ 23:15 Reardon reframes consumer targeting vs clinical FDA focusReardon educates the audience member on startup strategy, demonstrating that avoiding slow FDA clinical paths and targeting consumer scale generates the ML data necessary to advance prosthetic technology.
Matt holds his own ▶ 8:42 Matt summarizes core neuroscience concept in layman termsMatt demonstrates high technical mastery by synthesizing complex material into a concise summary that the brain's sole output mechanism is muscle control.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Data Driven NYC Event Title Card | 2 | 4 | 2 | 0 | Matt introduces Thomas Reardon and CTRL-Labs, framing their work as an API for the brain. Reardon gently corrects the framing, emphasizing 'neural interface technology' over 'brain-machine interface' because the nervous system extends down the spinal cord. Matt accepts the clarification and moves to video demos. | |
| Video Demo: VR Force-Like Object Control | 3 | 3 | 5 | 1 | Reardon displays significant combativeness when dismissing a redundant demo video as a poorer version and calling the iPhone a 'Trump level disaster for humanity'. Matt engages constructively by framing the tech as capturing intent and citing Reardon's previous comments on mobile devices. | |
| Technical Mechanism: Surface Electromyography and Motor Neurons | 1 | 6 | 1 | 0 | Reardon delivers a technical explanation of surface electromyography, motor neuron deconvolution, and action potentials. Matt steps back completely to let the guest explain the neuroscience and signal processing mechanics without interruption. | |
| The Human Output Mechanism and Natural Motor Control | 4 | 2 | 1 | 0 | Matt demonstrates high understanding by offering a crisp layman summary that the brain's sole output is controlling muscles. Reardon excitedly agrees with Matt's summary and builds on it using the complexity of drinking a beer to contrast natural motor control with phone interaction. | |
| Target Applications: VR/AR, Pervasive Computing, and Robotics | 2 | 2 | 1 | 0 | Matt prompts Reardon to discuss application areas. Reardon details their go-to-market focus on VR/AR immersive environments, pervasive computing like smartwatch interaction, and unexpected inbound demand from robotics. | |
| Developer Program Vetting and Platform Strategy | 3 | 4 | 2 | 1 | Matt asks about platform strategy, developer selection criteria, and why machine learning is strictly necessary. Reardon candidly calls 'bullshit' on himself for lack of proof videos, then educates the room on why decoding the nervous system is the mother of all ML problems. | |
| Thomas Reardon's Career: Internet Explorer to Neuroscience | 5 | 2 | 3 | 3 | When Matt asks about Reardon's past leading Internet Explorer, Reardon attempts to deflect ('Not really'). Matt pushes back directly by narrating Reardon's background himself, listing his work on IE, W3C standards, Columbia PhD, and large family background. | |
| Q&A: Prosthetics and Clinical Applications | 0 | 5 | 2 | 0 | An audience member asks why prosthetics and amputee care were not mentioned. Reardon explains that while they do advanced work with Johns Hopkins, pursuing a broad consumer market yields data scale much faster than an FDA clinical route, ultimately helping clinical patients faster. | |
| Q&A: Business Model and Personal Neural Models | 1 | 4 | 3 | 0 | Reardon answers audience questions regarding revenue models and personal neural calibration. When asked about nanotech applications for nerve damage, Reardon bluntly rejects the premise before explaining motor neuron diseases like ALS. | |
| Q&A: Learning Curve, Neural Signatures, and Conclusion | 4 | 3 | 1 | 0 | Matt asks a final composite question about learning curves, potential advantages for children, and societal consequences. Reardon explains how stochastic neural signatures make each user's model unique and details typical training times. |
Statements from this episode (22)
Reardon: CTRL-Labs is primarily a neuroscience and machine learning company
“We are only very thinly a hardware company. We are overwhelmingly a neurosciences and machine learning company.”
CTRL-Labs band decodes motor neuron activity through the forearm
“Okay, so this is somebody wearing our band that's decoding the activity of the neurons that drive the hand, so that happens through the forearm.”
No living neuroscientist can explain what a thought is, says Thomas Reardon
“I'll tell you there's not a neuroscientist alive who can tell you what a thought is.”
Thomas Reardon calls the iPhone a Trump-level disaster for humanity
“I think the iPhone is a Trump level disaster for humanity.”
Thomas Reardon wants to replace keyboards with thought-speed neural text input
“What I want to do is get rid of the 19th century keyboard entirely and allow you to effectively sort of text, if you will, at the level of your cognitive capacity.”
CTRL-Labs reconstructs motor neuron activity before physical muscle movement occurs
“We figured out how to actually reconstruct the activity of the motor neurons that drove that electrical activity in the muscles themselves. This is activity that actually happens long before The muscles actually move, you know, long in the computational sense …”
Reardon: The human brain's only physical output is controlling muscles
“Ah, the only thing your brain does is turn muscles on and off. There's nothing else that it does.”
CTRL-Labs began shipping its neural developer kits in September 2019
“We started shipping our developer kits last Friday after, boy, over three years of work.”
Inbound robotics demand drove unexpected investments at CTRL-Labs
“We didn't start the company thinking about robotics at all, and those opportunities have all kind of inbound to us and actually led to some of our investments.”
CTRL-Labs received up to 100,000 developer applications for its platform
“More than 10,000, less than a 100,000.”
Reardon: CTRL-Labs conducts video interviews with every developer applicant
“We interview, we do video conferences and interview every single applicant.”
Decoding the human nervous system is harder than predicting the weather
“I mean, decoding the human nervous system is the mother of all machine learning problems. I think it's significantly more complicated than predicting the weather, say.”
Reardon: CTRL-Labs captures single neuron signals without penetrating the skull
“This is a, Pretty significant breakthrough that we're able to get to single neurons without basically penetrating into your skull.”
Thomas Reardon created Internet Explorer in 1994 and led it through IE4
“I started up Internet Explorer in 94 and then worked on IE all the way through IE four, when we finally sort of won the browser war.”
Thomas Reardon co-founded W3C with Tim Berners-Lee and helped develop CSS
“I helped found the W-three-C with Tim Berners-Lee and kind of set up those initial web standards I, you know, I'm very proud of my work on CSS, if people know what that is.”
Neuralink's clinical-first strategy leads to grindingly slow progress, says Thomas Reardon
“Whereas if you go just to the clinical population first, which is sort of where Neuralink is going with their work, it's just grindingly slow progress.”
CTRL-Labs plans subscriptions for continuous neural modeling with free developer hardware
“Yeah, so our revenue model is based on us charging effectively a subscription for modeling work, but in fact, what we are pushing right now is one in which it's a free license to the hardware and a free license to the initial models, and we want to get paid Fo…”
A trained CTRL-Labs band will not work at all on another person
“This is a very personal model that is custom to just you. When you put on the band after you've trained it, if you put it on somebody else, it doesn't work at all. It is, and it's not like design. That's just the way neural interfaces are required to work. The…”
CTRL-Labs needs only a dozen neurons to decode hand movement intent
“You need, like I said, 30,000 neurons to drive these hand movements, but it's mostly because you need those neurons to sort of gain up the signal. But we only need, like, a dozen to be able to actually understand what the actual movement intent was, rather tha…”
CTRL-Labs uniquely identifies users among 7.4 billion in 800 milliseconds
“I can put a band on you and after 10 minutes I will know enough from those signals that I can pull you out of a population of 7.4 billion people after 10 minutes of training data and do that in about 800 milliseconds.”