What a Decade of Building Prosthetic Hands Taught Me About Touch

The first robotic hand I built was made out of LEGO, fishing wire, and surgical tubing. I was 14, living in a small town in southwest Colorado, and I didn't really know what I was doing. I learned most of it from YouTube and by emailing people who knew more than me.

It worked well enough to take third place at the Colorado state science fair. That wasn't the important part of that day though.

At the fair I met a girl who was about seven years old. She was wearing a prosthetic arm that cost her family $80,000. It could open and close, and that was about it. She was going to outgrow it in a couple of years, and then her family would have to do it all over again.

I remember thinking that what I had built on my bedroom floor, with way less money, wasn't that far off from what she was wearing. That stuck with me, and honestly it's the reason everything since then happened.

From science fair to company

The next few years moved fast. I got a 3D printer for my 16th birthday and started printing hands instead of building them out of LEGO. The cost of a working arm came down to a few hundred dollars. In 2012 that arm took second place in engineering at the International Science and Engineering Fair. In 2013 I brought it to the White House Science Fair and got to shake President Obama's hand, which is still a little surreal to say. That summer I interned at NASA working on Robonaut.

In 2014, at 18, I started Unlimited Tomorrow with seed funding from Tony Robbins. I skipped college. The goal was simple to say and hard to do: make advanced prosthetic arms that are personal, functional, and affordable.

That turned into TrueLimb, a custom 3D-printed bionic arm that we fit remotely. People scanned their limb at home with a scanner we shipped them, and we designed the arm around their body and even matched their skin tone. Along the way we worked with Microsoft on remote fitting, and with HP and Siemens on how to make and design the arms at scale.

I spent most of a decade on that problem. It taught me a lot about hands, and most of it wasn't what I expected.

Moving a hand isn't the same as having one

When you build a prosthetic hand, almost all of the early effort goes into making it move. More grips, more speed, better control. That's what gets demoed, and that's what I focused on too.

But a hand isn't only an output device. Your own hand is constantly telling you things. How hard you're squeezing a paper cup. Whether a bolt has actually seated. When to stop gripping before you crush something. You don't think about any of this because it happens automatically, through your skin and through the tendons in your fingers.

Take that feedback away and everything gets harder. Without a sense of grip force, you have to watch your hand to know what it's doing, and fine tasks turn into guesswork.

That's the lesson I keep coming back to. Control gets the attention, but feedback is half of what a hand is.

VR has the same problem

When I started spending real time in VR, I noticed the same gap from the other direction. Headsets have gotten really good at hand tracking. You can see your hand, move every finger, and reach for things. But when you close your hand around a virtual tool, nothing happens. Your fingers go straight through it.

So people adapt the same way someone with a basic prosthetic does. They watch their hands. They learn the visual cue instead of the feel. For a lot of training, that's a problem, because the feel is the skill. A technician learning to torque a fitting or seat a connector needs to know what "right" feels like, not just what it looks like.

That's where TactraFlex came from. In 2024 we started taking the tendon-driven systems we'd developed for prosthetic hands and running them the other way. Instead of the motors moving a hand, they resist your fingers, so when you grab something in VR, your fingers actually stop.

The things prosthetics taught me that apply directly

Weight decides whether something gets used. We fought for every gram on our prosthetic arms. Momo Sutton was nine when we fit her, through our work with Microsoft, with an arm that weighed under a pound. If a device is heavy and uncomfortable, people take it off, no matter how good the tech inside is. TactraFlex weighs under 150 grams with force feedback built in. That number came from that lesson.

Setup is part of the product. With TrueLimb, we shipped a scanner to people's homes so they didn't have to travel to a clinic for every fitting. The best technology in the world doesn't matter if it's a pain to get started. TactraFlex uses the headset's own hand tracking, so there are no base stations or external trackers to set up. A trainer should be able to hand someone a pair of gloves and get going.

Price is a design decision. That $80,000 arm is the reason I started all of this. In enterprise hardware it's normal to hide pricing behind a sales call. We publish ours: $7,500 per pair. I'd rather a team know right away whether we're a fit.

The person using it is the expert. The most useful feedback I ever got on a prosthetic arm came from the people wearing them, not from engineers. Same thing with training. The trainees and the people running the program will tell you very quickly what works and what doesn't.

Where we are now

This year our force feedback patent was issued, and we launched TactraFlex. The glove combines three-channel tendon-driven force feedback with 22 vibrotactile zones, works with native hand tracking, and has a Unity SDK built on OpenXR so teams can add haptics to the training they already have.

It's a different product for a different customer than TrueLimb, but it's really the same work. I've spent most of my life trying to understand what hands do and how to give some of that back to people. First it was people who had lost a hand. Now it's everyone who trains with their hands in VR.

If you're running hands-on training and want to see what it feels like, reach out. I'd love to hear what you're working on.