How Boston Dynamics' robots from went from YouTube stars to actual tools for helping doctors treat coronavirus patients and enforcing social distancing

• Boston Dynamics' lifelike robots have become famous for their almost uncanny, realistic movements in recent years.
• After years of hype mostly generated by their massively popular YouTube videos, Boston Dynamics began licensing its four-legged Spot robot last year.
• Now, Spot is being used in a hospital in Boston to help doctors remotely communicate with COVID-19 patients, and in Singapore to help enforce social distancing.
• Boston Dynamics founder Marc Raibert spoke with Business Insider about Spot's ambitions in telemedicine, how Spot became Internet famous, and whether robots like Spot will ever surface in the home.
• Visit Business Insider's homepage for more stories.

Spot, Boston Dynamics' four-legged dog-like machine, may just be the most famous robot of all time.

Ever since the Massachussetts-based robotics company posted its first video to YouTube more than 10 years ago, a clip that shows how a quadruped robot called "BigDog" was able to keep its balance after being kicked, the company's robots have gained a fervent following.

Love them or hate them — some have called them "terrifying", while others have applauded their parkour skills — they've certainly attracted an audience, and more than 1.7 million subscribers on YouTube.

That first video from 2009 has 1.5 million views, while Boston Dynamics' most recent video of its humanoid Atlas robot has more than 9.7 million views.

Marc Raibert, Boston Dynamics' founder, chairman, and former CEO, says its robots were never meant to be famous. In fact, it happened by accident.

"At the beginning, we had no idea," Raibert said to Business Insider. "A couple of early videos, they actually kind of leaked onto YouTube, and we saw that our customers and sponsors had a very strong, positive response."

Over the past year, Boston Dynamics' Spot robot has established more of a presence offscreen and outside the lab as well. Boston Dynamics began leasing out its Spot robots back in September, and since then the robots have been used to patrol oil rigs, assist law enforcement, and help on construction sites.

Most recently, Spot has gotten an even bigger job: helping doctors communicate with COVID-19 patients remotely at the Brigham and Women's Hospital in Boston. Raibert believes there's a broader opportunity for robots like Spot to assist with telemedicine, and hopes the robot will soon be able to help take vitals from patients in the near future.

"The big picture is, you can have a situation where it's not good to have direct human contact with whatever is being contacted, and the robots are great for that," he said.

Business Insider recently spoke with Raibert and Michael Perry, Boston Dynamics' vice president of business development, to discuss how the company has adapted to working remotely during the COVID-19 pandemic, how Spot is being used in hospitals, and whether there will ever be robots in our homes unloading our dishwashers.

The interview below has been lightly edited for length and clarity.

On robotics after the pandemic

Lisa Eadicicco, Business Insider: How do you think Boston Dynamics will change after the coronavirus? How do you think the robotics and technology industries overall will change after the virus, and how do you think the world will change?

Marc Raibert, chairman of Boston Dynamics: Well, I think there's sort of an immediate today, tomorrow and the long-term-future answers to your question. Immediately, we've had to make a transition from working heavily in our laboratories, where we have robots and testing and basically rely on the facilities being the center that all the work revolves around, to a more distributed operation. We have successfully kept our operation going by having people with robots at home and able to work on them.

I think we were very skeptical about being able to work from home and being distributed, and clearly the success and making that work is going to carry forward and more people will continue to do that, probably permanently. I don't think we're going to permanently become a distributed virtual company like some are. We're still going to value getting together both to see each other, but also to work with the machines.

On the robotics side, for some people it's becoming clear that having robots that can free people from directly engaging in dull, dirty, and dangerous work is a thing that would be valuable to the degree that robotics companies can achieve it.

One of the projects we've been working on we've been calling Dr. Spot. We're working with the Brigham and Women's hospital here in Boston. We've had robots there that are doing tele-interaction with potential patients. We started to monitor just the front end, just a little bit of monitoring their vital signs remote with cameras mounted on the robot. And we're working on doing more and more that way.

Eadicicco: How long do you think it'll be before Spot can measure vitals more broadly? And how long do you think it'll be until that capability is out of development?

Michael Perry, vice president of business development at Boston Dynamics: So we actually deployed our remote vitals assessment version of Spot to Brigham and Women's Hospital, so they just started testing it. The current capabilities [include] being able to do assessment of body temperature, the surface temperature, and you can make some inferences on core body temperature from surface temperature. And then you can also look at respiratory rate as well as pulse.

The fourth thing that's really critical for this application is looking at the volume of oxygen in the blood. And that's still something that the entire field is looking at how you can do that without contact. So we've got some leads on algorithms there, but we've not put them into production yet.

Eadicicco: Are you looking to expand Spot to other hospitals as well during the pandemic?

Raibert: We've gotten a lot of inbound inquiries about it. We don't have it deployed anywhere else yet, Michael and his team are in conversations with lots of potential users.

Perry: So the good news about hospitals is that they're very meticulous about how they adopt new technology into their workflow. Doctors and their nurses are saying [they] need this immediately, but it also has to go to the internal review board to go through their legal team and things like that. So we're in that process with a few other hospitals. But, you know, to be candid, we're probably going to be past the peak by the time that more hospitals actually start deploying it.

On making robots useful

Eadicicco: Before this happened, a lot of the headlines were about how Spot was being used in a lot of different applications like law enforcement, inspecting oil rigs, and things like that. What has the virus taught you about the future of how robots like Spot can be used elsewhere? Are there any new industries or applications that you're looking to move into that maybe you weren't considering before the pandemic?

Raibert: Well, it's always been our ambition to embrace a broad set of uses for Spot. We called it a platform and still do. So it's always been our intention to get the robot out there in a form where the capability was accessible to a wide variety of people, including people who are going to develop actual use cases on top of the platform. So we're not surprised, and we've had, even from day one, let's say for the last six or eight months, we've had a broad set of kinds of uses.

Perry: The key story is about safety. You know, where the robot is primarily providing value by taking a person out of a dangerous situation. On a construction site that's somewhat true, but the main story for those customers previously was about process improvement and really clear [return on investment] calculations about how much we're going to decrease the overall project costs by doing better monitoring.

But it's also a function of building a just-in-case model so that there's limited disruption to their business flow. So we've got construction site managers who are managing 30 projects across the Northeast, and they're sitting in their homes unable to access those sites. And even if work is continuing progress on those sites, wouldn't it be great if they could dial into a robot and just walk around the construction site and see how things are going? Or better document the progress on the job site?

The safety story has become more top of mind for more industries at a faster rate than it was, I would say, six months ago. But now what we're seeing across all industry segments, including entertainment and theme parks, is how to limit the amount of people that have to be in a given environment at a given time. And that's obviously top of mind right now during this crisis.

How Boston Dynamics' robots became internet famous

Eadicicco: You've released videos of your robots over the years that have become pretty popular. I haven't really seen that from like any other robotics companies. What has that been like? And did you ever expect them to catch on the way that they have over the years?

Raibert: At the beginning we had no idea. We didn't say, 'Okay, how are we going to do marketing without a marketing budget?' A couple of early videos, they actually kind of leaked onto YouTube. And we saw that our customers and sponsors had a very strong positive response. And so we decided, well, we'll put out more.

But in terms of the style, which I think is responsible for some of the reaction, it's always been our goal to just show what the robot is doing. You know, not do a lot of rigmarole with boosting it up. Obviously we edit to show usually the best stuff, but we try to be honest in them. We don't show something that we only did once and couldn't be done before or after.

Eadicicco: So the first video resulted from a leak?

Raibert: It wasn't exactly a leak. We put the video on our website and then someone took it and put it on YouTube and it got a big response. And so we said, 'Oh, okay, I guess we better have our own YouTube site.' We'd always made videos even for many years before YouTube. Part of our DNA was to show progress. If you write papers and do graphs about robots, you get a totally different impression than if you see that moving in the world.

On when you'll see robots in the home, doing everything

Eadicicco: We've talked a lot about all the different industries that Spot has been present in, but there have also been these videos over the years of Spot helping unload the dishwasher and things like that. I'd love to hear more about whether you think there's ever going to be a place for a robot like Spot in the home and what that would look like. It's not really something that has broken into the consumer space yet, even though companies like Amazon are rumored to be working on it.

Raibert: We absolutely believe in the long run, robots will be in people's homes helping to take care of elderly relatives who now require outsiders to come in and help them do routine things like clean up or empty the dishwasher. I think you've seen our humanoid robot Atlas doing gymnastics and other kinds of athletic activity. Those projects have a very long, long-term focus on being able to build a very capable set of skills and abilities that could be used for that.

But I think it's going to be some time before you see robots broadly in our homes. Mostly because people are still smarter than robots. People are still more dexterous and more mobile than robots. And so you need a more structured use case in order to justify them today. So are we going to go there? You bet. Are we working on it? Part of what we do is focused on that long-term progress that will enable us to use them very broadly that way. But are the current robots we're building going to be used that way? We're not doing that ourselves, and it's a tough problem to solve.

Eadicicco: How far away do you think that is?

Raibert: People like us are always saying five or eight years, something like those timeframes. I'm not saying that there won't be something that looks a little bit more like an appliance in people's homes that also has some robot characteristics. But it will be very narrow focused probably if it's going to be successful, and also if it's going to be low enough cost.

Perry: If you look at most of the industrial environments that we've been talking about or even in the home, there are robots available. But they're scoped to a very narrow set of tasks that are constrained to a very narrow set of environments.

Raibert: I dream about the day when we'll have robots that have a very significant mobility perception and manipulation capability to do things in your home the way anybody could.

Perry: One of the hard things that we have, at least in talking to customers, is to disabuse them that Spot is "Rosey the Robot" — that with one robot, you can do every single thing. But ultimately that's kind of the dream robotics scenario.

Eadicicco: What do you think will be like the biggest breakthrough in robotics and like the next five years?

Raibert: I don't really believe in breakthroughs the way the media likes to talk about it. There's so many different pieces to robots, all of which are making progress that even if you had a "breakthrough" in how well the camera worked, you'd still have 10 other systems. There's the power in the robot, and getting better batteries could really help. But that wouldn't just solve robotics.

There's the perception of the cameras. Cloud-based machine learning has really advanced visual perception, but when you have robots moving in an unknown world, dynamically interacting with things, there's still a lot of progress that can and will be made there that's needed to reach the dream we were talking about a little bit earlier.