In recent years, the two-legged humanoid robot has brought bold visions from industry leaders like Tesla or Figure.AI of highly intelligent robot servants unlocking limitless economic potential.
A little-known startup out in Austin says not so fast: The wheeled robot is fine for most conditions.
Diligent Robotics, an Austin-based startup founded in 2017, builds robots for the healthcare space. Its wheeled robot, Moxi, can automate tasks in the hospital, such as delivering lab samples or medical supplies. These are routine but essential chores that Vivian Chu, co-founder of Diligent Robotics, says take precious time away from human staff and patients.
“I do think that over time, you’re going to very quickly see that the wheel form factor — like two arms, a head, and a wheel — is probably going to be a form factor that ends up covering the majority of what you need,” Chu told Business Insider in an interview.
For Moxi, don’t think of a humanoid robot installed with the latest ChatGPT model, carrying out free-flowing dialogue. Chu likes to think of Moxi more as the “minimum viable humanoid” whittled down to focus on specific tasks.
That doesn’t mean Moxi is simple to build or any less useful.
With more than 90 robots deployed across 25+ hospitals, Chu told BI that Diligent’s Moxi robots have carried out over 1.2 million supply deliveries. The company estimates that that equates to about half a million hours of human time saved.
Business Insider spoke with Chu about how Diligent Robotics is meeting hospitals’ needs, where the technology needs improvements, and why humanoid robots are a ways away from entering our households.
This interview has been edited for length and clarity.
How did you get into robotics?
I grew up in the Bay Area with classic engineering parents. In my senior year at Berkeley, I took my first intro to robotics class, where I got to program a robot. I thought, “Damn it, I want to learn more.”
So I went to Penn for a master’s and then to Georgia Tech for a PhD. That’s where I met my cofounder, Andrea Thomaz, who was my PhD advisor. She’s a leading world expert on how to put robots and people together.
When I met her, we were both at the stage of life where we thought, “Robots are cool. We write our papers about them, we do one lab experiment, and then we just go.” But we wanted robots to actually make a difference, like we said in our papers.
We decided that the technology was starting to get ready. When Andrea first started our lab, a robot cost half a million dollars. By 2015, 2016, you could build the same robot for $80,000 to $100,00.
Even now, the cost has dramatically gone down. That’s when we decided to start Diligent.
Why hospitals?
We did about 100 to 150 interviews across the market. Really quickly, we homed in on healthcare.
In the early days, we shadowed nurses and staff with clipboards and stopwatches to see where robots could make a difference. Even before COVID, we found that nurses and clinicians spend up to 30% of their time just fetching and gathering supplies, and less than 30% on direct patient care because they’re so busy with logistical tasks.
Any time returned to a staff member meant better patient outcomes. We saw that and thought, “This is where robots should be. This is how you think about robots augmenting people as opposed to replacing them.”
How did you decide on Moxi’s design?
When we were fundraising early in 2017, investors would ask us, “Why do you have an arm on your robot?”
For us, it was a no-brainer. If you don’t, hospitals would have to modify every single door and elevator. Retrofitting a hospital could cost a million dollars and could take a year.
With an arm, Moxi can just push the elevator button directly.
The head is important too. We’re putting robots in a place where people have never interacted with a robot. Things like eye contact, or if I’m looking at my cup, you know I’m going to drink my coffee. A head means the robot can communicate intent. Our main sensor for manipulation is in the head, so every time it looks, it’s actively perceiving and communicating intent.
As for the wheels, hospitals are ADA-compliant. They have ramps and automated doors. Going with a wheel form factor made sense. It is statistically stable, safe, and perfect for an environment with elevators and ramps. If something happens and you power down, you can just roll the robot to the side.
How do hospitals measure success with Moxi?
Time and steps saved are probably the first basic metrics.
One of our hospital systems had teleboxes — these are bedside monitors that monitor a patient’s vital signs and have to be sent around constantly. After we installed Moxi, now every single telemetry box in three massive buildings is sent through the robot. There are now zero people who run the boxes.
Overall, Moxi has made over 1.2 million deliveries. Based on average distance and tasks, we estimate that staff would’ve saved around half a million hours of time walking equipment, lab samples, or medication around.
Do you think humanoid, legged robots are coming soon?
Over time, you’re going to see that the form factor with a head, two arms, and wheels is probably going to cover the majority of tasks. There is an interesting segment where you do need legs — disaster recovery is one that comes to mind, or homes where you have stairs.
It comes down to the workflow. What is the pain point? What can robots be doing 24/7 that truly gives people their time back?
It can be very premature to talk about legged robots because the thing that a warehouse needs might not require legs.
A legged form factor in homes — I see it pretty far, far away.
How have investor conversations changed in light of recent focus on bipedal humanoid robots?
In the beginning, investors would ask, “Why do you have an arm?” Now they ask, “Why don’t you have two arms?” And there are tasks that will require a second arm — that’s on our future roadmap.
There is some education you have to do with investors — do you go all in on general-purpose, or focus on a specific segment? With all deep tech, you just have to find the right investors who understand the thesis.
Any broader advancements in robots you’re looking out for?
Battery technology. Our robots last about 16 hours total with smart charging in between. As we scale compute, power becomes the bottleneck. I think some of these humanoid companies only have a 90-minute runtime. Then there’s safety: Hospitals care a lot about the safety of your platform, like whether it could catch on fire.
That’s what I’d love to see innovation in — safer, longer-lasting batteries.
What’s next for Diligent?
More dexterity is part of the future product lines that we’re thinking about. Hospitals want us to do more than transport — things like assembling case carts or handling more supplies. Over time, having other form factors like two arms will make sense. Dialogue is also something we’re exploring.
We’re also working on integrating Moxi directly with hospital records. So instead of someone tapping on an app, a patient admission could automatically trigger a task for the robot.
Read the full article here
