Next session with one of the top engineers in the field, who is developing all sorts of extraordinary machines that can do new functions in the workplace. We are going to demonstrate that as well. So, please welcome to the stage marc raibert, the founder and ceo of boston dynamics. He is going to be interviewed by a columnist at the wall street journal. [applause] im not sure i belong here, but you most certainly do. Marc it is great to be here. We have a late guest. One more guest. We would like to welcome them. Spot, wherever you are. Marc come on in, spot. Robot, walkinge in from the back. People sitting close to the aisles should keep their hands in their laps please. [laughter] bite. Spot does not you must be sensational at parties with Something Like this. Marc at thanksgiving, we had one at home. It was great fun. Spot is a very exciting robot for us. We have been shipping it for about two months. He uses the cameras for navigating up the stairs, things like that. One of the cool things about a legged robot is that it can drive, it can drive sideways, it can turn in place. I think it already did that. This is a really small stage for doing this. And it can pitch and roll the body. That is really important. Im going to give you all an assignment for lunch. Sit at the table and keep your shoulders still and try to reach her water glass and you are going to learn how short and really incapable your arms are by themselves, but if you have your arm in a body that can travel, then you really have expanded the workspace dramatically to make the are more useful. This is a mobile manipulation system. Although danielle is driving today, its got autonomy functions that let it travel around the space that is mapped. There are some cool coordination that a robot like this can do, which is my favorite thing, called chicken head mode. [laughter] marc you notice it is Holding Still while its body moves around it. We all have these capabilities, too. It really lets us think about the task we are doing with our hands even though our bodies are part of the equation. Coordinatingy for and allowing things like that. How long has spot been developing . You did not do this in two weeks. All in . Marc about 10 years ago, we developed the quadra pad with darpa funding. Big dog. Marc that was big dog. I used to be a professor where we had some running quadra pads in my lab also funded by darpa, but many years earlier. Between big dog and this, we went through about four generations of machines, each going at things a little different way. Big dog had a gasoline engine. This one has batteries. That was hydraulic, this is electric. It did not have as much perception as this robot. There has been lots of evolution. There is lots of evolution to go. If i want to get one of these, can you put me in one of these today . Marc about 45 days. 45 days. Marc we have been selling them into an early adopter program, where the customers, we have been trying to pick customers that have verticals we think have potential for lots of growth. We have been selling into construction sites, data acquisition, which is really a burgeoning field. Where people mount their own sensors and collect either 3d data or 360 imagery or things like that, so they can monitor progress of a construction site. We have been also working with gas and oil. A refinery, most of the refineries we have seen so far were built years ago. They are not digitized. They are really hard to get around. There is really no other robot besides one like this that can get up the catwalks and go to all the places to do methane sensing, read gauges, and things like that. What is the benefit to Something Like this versus a drone . Marc drones are great. They have a lot of access, but they are short operation time. I think 30 minutes is the max. Some of which has to be getting there and back. Very small payload. This can carry about 25 pounds payload. This by the way is designed as a platform. There is what looks like a roof rack on the back where you can mount your own hardware. It has a programming ati. Can i get it in red . Marc we actually have christmas skins. We have changing lights and stuff. If you want, i think i have it on my cell phone, the video. We are going to take spot out in the lobby after the talk, so if you want to get some handson experience, if that photographer could get out of the way [laughter] marc we will have it out there for half an hour so you could drive it around or watch people take selfies. Goodbye, spot. I dont know how we follow that. Have a seat. Marc i have been telling people, its like an organ grinder and a monkey. The question is who is the organ grinder and who is the monkey . [laughter] lets talk pricing. The word i heard was you can get me a spot for about the price of may be midrange car. Is that accurate . Marc that is accurate. Andt now, we are leasing doing deals depending on how many they want and how long and what kind of support they get. We are not announcing a price. About the price of a midsized car. Not a ceos car. [laughter] in mid managers car. Everybody here drives a ugo. You mentioned at the top that this is an exciting time for your company. You have been known in the world is kind of this dream factory. Now, you are out in the world. Feedback. Tting the what are you learning about your product out there now that people finally have them . It has been fantastic. In addition to providing feedback lets us improve the product, we are experimenting to find out where the market is. We are not a normal company that finds a market first. We are robot people first, for better or worse. Having this platform that can be adapted to lots of different verticals is an opportunity to see what is out there. We sort of thought that gas and oil, public safety, which means police, bomb squad type stuff, construction, entertainment. We have some pretty big entertainment customers that we are working with. Without that those would be the big things. We are really looking forward to seeing what comes in that weve never thought of. Wait a second. Entertainment . You have done something with cirque du soleil. You are getting into showbiz . Marc we are going to probably get into showbiz. [laughter] marc we did a project with cirque du soleil. I think the video is online if you want to find it. Weve got the robots doing a little bit of dancing. We do have some customers who are looking at other applications, not exactly theatrical, but not too far off. Can i buy a stuffed spot at walmart and a couple of years . Marc probably not from us. [laughter] marc i think we are in a lego contest. Lego has markups of different products, than they get people to vote. I think the voting on ours is not so high yet. I think we have a video here coming up of a couple other products that are in development with boston dynamics. Marc this is our lamborghini project. In addition to making robots that can be productized. This is a robot that lets us do r d on the boundaries of what is possible. We think of this as i race car. Really expensive, kind of finicky, but the highest possible performance. We have been using it to develop new techniques. Some of the things we have done have trickled down to more practical designs that are in our pipeline. This is both a hardware and a software project. As all of our work is. The intimate connection between hardware and software is really an important part of what we do. This is spot doing marc this is spot being tested. Some people think we are being mean to the robot. I promise you, the robot has no feelings. It is not angry. Thate really showing off to get a robot to be useful, you dont want it to do exactly what the plan situation is, you wanted to be able to tolerate various kinds of disturbances because that is what the real world is like. [laughter] marc if you can get the robot to have those skills to deal with the unexpected, it is a lot easier to get higher level ai communicating with it and telling it what to do to get practical work done. A drone cannot do that. Marc probably most cant. [laughter] i saw a phrase you said once that you were talking about making automation for places not designed for automation. Marc you know, the mobility of these robots lets them go anywhere a person or animal could go. They are not quite there yet, but that is sort of the vision. Uneven terrain. As i said, and oil refinery is a place that has really hard to get to places. A person can do it. Climbing stairways and walkways and railings and things like that. Your normal robots cant do that. So, that is where we are focused. The robot that we saw in the , there is another one called logistics. Can you describe that . Marc spot i call out today robot because we have productized it and we are shipping it. Our tomorrow robot is one that is about a year or 18 months away from being a product. It is a logistics robot. Spot is generalpurpose and is designed to be able to do lots of Different Things defined by customers. This one does exactly one thing. It picks up a box, travels with it, and puts it down, and it uses Computer Vision to do that. It can take the box out of a stack, it can go to the back of a truck and unload a stack. It is designed for moving boxes in a logistics world. Boxesare about a trillion shipped every year around the world. It is a huge market. Even though you have heard a lot about automation, for instance at amazon and its competitors, that is mostly in one narrow area which is picking items out of cases and putting them into orders in order to get shipped to you. All around that, both at amazon and altogether Retail Operations , there is taking boxes off trucks, sorting them into pallets, going from single skew pallets to multiskew pallets, than putting things back in trucks. There is also retail opportunities for this. That is our tomorrow activity, where we are building a specialpurpose robot to do that. It will be about a year or 18 months. We are doing physical testing with the people you would imagine. Atlas is the future. Marc atlas is our experimental basically, i think robots are cool today, but they are not really what we dream of. I really think robots could be physically as adept or better than people at all the range of things we can do. We are working toward that. Hosting conferences, maybe . Marc mostly physical things. [laughter] you resisted the notion out there of robots being scary. And the guy read somewhere that you said, it is people who are scary, it is not the robots themselves. Marc we build the robots. I guess it is our fault, we build a robot that has a shape of a human and if it starts to behave realistically enough, then an audience will look at it and attributed with all the characteristics of a person. , wanting having an ego power anger. Marc anger, having to deal with ethical issues like the Previous Panel was talking about, but those are nowhere in the technology we are developing nor most robot builders. It is a totally separate thing. There are people out there making parity videos of robots. Those must make you nuts. Marc some of them cause some pain. We have worked to make our keep our we use to robots only for the purpose of showing off how good they work. Then we started poking them with a hockey stick for the same thing, to show how they can respond to uncertainty. A lot of people took that as us being mean to the robot. We stopped doing that sort of thing and found other ways to test the robots. Some of these parodies are showing robots shooting back at people and kicking them. Lets jump the questions from the audience on this. Let me start with my own. It is interesting to me to watch spot come up on stage. There is a driver still. There is somebody directing it. Where does Machine Learning enter into this, Artificial Intelligence . Where the robot can be given a task by a control center and then walk out into the Oil Installation for the construction site and execute something without a driver . Marc let me give two answers. In terms of spot, although today we were driving it because there are people close by, it has modes that are more autonomous. I think autonomy is at the root of your question. Autonomy is a very tall, multilayered cake. I submit that even people are rarely fully autonomous. The context of the other people who give them instruction or who they have to interact with. There are plans for a lot of things they do. We are good at solving problems and we can make up a plan for ourselves, its true. Anyway, our robots have various levels of autonomy. Spot can be given an assignment to travel around a complicated environment and take pictures at specific locations and then send them to a centralized place, where they are are all collated to be able to do the kind of Construction Management i was talking about. You are probably asking, when can they be told to go to an unmapped place and then decide, i dont have a map, i will have to map it, then once they have a map, figure out the readings. We are not working on that, but i think other people in ai could be and we have designed our platform to be something that can interact with the cloud, that could have other peoples ai interacting with it and it would be possible to build more autonomy. We are actually talking to some customers, but also developers who are working on more advanced higher level ai for that. I would like to say you can break ai into the intelligence into the athletic part and the scholarly part. The athletic part lets me move where imlay a sport maneuvering with respect to another person. It is realtime, very immediate, local. Scholarly ai are things like we are talking about planning your trip to the airport, making sure youre going to get there on time given the traffic and all that. Questions from the audience. We talked a little bit earlier about the utility industry. Many industries are focused on automation, digitization, the future. How should we look at that from your perspective . When we see the acceleration of the ability to certainly move to automation, those kinds of activities, what do you cs the future in terms of robotics and how quickly you will move ahead . Many of us are looking at how to deal with that in the future. Marc i think the kind of issues that come out at facilities like yours and we are working with national grid, probably a similar kind of company, where we are doing some inspections. Some inspections were people just cant go because it is too dangerous. Others, people could do, but there are various tradeoffs. Frome turning the corner being a Pure Research company to being a product company. The things that are right in our facebook reliability, cost face are reliability, cost, functionality. Aiont think for physical that things are happening in the eye like they seem to have happened for the internet. But they are absolutely making good progress. I think one of the considerations is you need to get in the water in order to learn what is involved in your side. , youch as we can provide have to have knowledgeable people who are ready to receive. That might involve having an Innovation Group or doing your own research projects. Even if you are planning on just accepting our technology as the solution. I think that is step one. In that process. Other questions . Yes, right here. Washington institute. Of do you address the issue ethical Artificial Intelligence . Some of the pioneers in technology have expressed their concern that what was a very pure desire on their part to be able to create things like the internet have turned into vehicles for ill. Im wondering how you address that in the work you are doing. Thoughtthink my first is that any new technology, i dont think robots and ai are that much different than other technologies, cars, airplanes, lasers, computers. They offer opportunities and they offer risks. Decide whathave to level of balance you want between taking advantage of the opportunities versus dealing with the potential risks. I dont agree that robotics has advanced in the ill application that you sort of mentioned. I think hollywood portrays them that way and in our minds, if you want to movies when you were a kid, you cant help but have that as a starting point for a lot of people. But i think it is a pretty big distortion from what actual robots that People Like Us and others maker doing or might be used for. Nearkill level is nowhere the levels portrayed by hollywood. I think to a large degree we are fighting those images. More quick one question. Right here. Just a quick i dont need to be cheeky. Robots in think about the manufacturing process of your robots . Marc you know, we are definitely a ways off from doing that. That obviously robots are used to do manufacturing. I have to say as a side comment that we still see a line between traditional manufacturing robots, which are pretty much turning into a commodity these days, very low cost very specific function, many Different Companies making the same thing and robots that have some amount of ai sensing and real time adjustment to their environments, like the one going through the door when someone is hassling it. That is a different world. I think that automation can be used to manufacture robots. We are working at a smaller scale then would make that possible right now. Is weow, a related thing are using contract manufacturers to build our robots. I think at first we thought we would be able to find contract manufacturers that would overnight translate our designs into something they were manufacturing. I think that might be true for Circuit Boards and iot devices that have a couple buttons and a screen, but for robots that have complex articulated mechanisms, most companies manufacturing those things are doing it themselves. That is not a small problem. We are going to have to leave it there. We have spot outfront. During the break a little bit later on, you can test drive. Right . Marc yes. Thank you very much, both of you. [captions Copyright National cable satellite corp. 2019] [captioning performed by the national captioning institute, which is responsible for its caption content and accuracy. Visit ncicap. Org] washington journal live every day. Coming up wednesday morning, as part of authors week, the author of the price we pay talking about efforts to fix the u. S. Health care system. Be sure to watch live at 7 00 eastern wednesday morning. Join the discussion. Watch authors week all this week starting at 8 00 a. M. Some of our look at featured programming this holiday week on cspan. On Christmas Day at 10 00 a. M. Eastern, this years white house decorations with first lady melania trump. Plus, he look back at previous years by former first ladys hillary clinton, laura bush, and michelle obama. At 12 30 eastern, a discussion about Global Technology issues at the manhattan institute. , john0 p. M. Eastern miller in the history of journalism and fake news at the liberty forum. On thursday at 5 45 eastern, a joint Economic Committee hearing on the high cost of raising a family. Eastern, occupational licensing requirements at the federalist society. Watch this holiday week on cspan. University of washington history professor Margaret Ohara discusses her book the code Silicon Valley and the remaking of america. You have the space race. You have the militaryindustrial complex. That becomes the foundation for this entrepreneurial flywheel of and privatereation Wealth Creation and an industry that is kind of considering itself an industry that bills itself on its own, that government has become almost invisible to many of the people who are in Silicon Valley, who are the creators of these companies, these technologies. They think there is not a role, but actually there is. It is kind of a government out of sight. Sunday night at 8 00 p. M. Eastern on cspans q a. Veterans who served in iraq, afghanistan, and vietnam discuss the complexities of war. This is hosted by the nixon library, chapman university, and the Orange County navy league. Good evening. I want to thank you again for coming out tonight. Just over four years ago, i came to the university from the History Department at west point. To direct a brandnew graduate program conceded by jennifer keene. We started from the presupposition based on comments from senior military officers and policymars