The Robot Report Podcast
00:23:11 · Mike Oitzman, Steve Crowe
[00:23:11] S1: Mhmm. [00:23:12] S0: Right? And I believe the number is probably pretty slim as in how a cobot is actually interacting with a human often and doing whatever task that they are carried out with doing. Think it's probably very slim. I think there's other benefits that cobots have had over the years rather than actually collaboratively working with people. But this really points to, and I think Aaron had a picture of this in his newsletter this morning, is any robot can be unsafe. [00:23:42] S1: Mhmm. [00:23:43] S0: Right? Just... And his picture, and I've seen different iterations of this picture many times over the years, is if you stuck a knife on the end of a quote unquote collaborative robot arm, that robot arm is now very unsafe. Right? [00:23:58] S1: Could still slice you. [00:23:59] S0: They can still slice you. So, it was kind of a misnomer. And, again, these standards, man, they take forever to update. [00:24:07] S1: Yep.
00:24:08 · Steve Crowe
[00:24:08] S0: There's just so much that goes into this. There's so many stakeholders and players and people that they have to reach out to and talk to. And I can't imagine being a part of, you know, these standards committees. These people are saints who do this stuff. Roberta Nelson Shea, she recently won an Engelberger Award. So she's done a lot of that stuff. We're talking with A3 next week about this new standard here to learn more about the changes and the big changes and how this will affect or impact things going forward. But I'm also curious. I think I know the answer to this. I'm gonna hope... Hopefully talk to Universal Robots about this as well. You know, they are the leader of, leading developer of collaborative robot arms, But they've always taken a very honest, safe approach to all of their deployments. So I'm not really sure how much this is going to impact them. I don't think it's really going to change anything that they do. But just getting rid of like, does this term collaborative robot, does it actually go away? You know, I think that's kind of the interesting part of it. When you hear the word cobot, I think universal robots is synonymous with that. Yep. What does that do to their marketing capabilities? Do they change things up? I don't know. I'm kind of interested in hearing what they have to say about that. But yeah, we're going dive into this more next week. As again, we're talking to A3 and some other folks about this. So stay tuned on that. But, is this the end of a cobot? I don't know. I don't think so. But it's fun. It's fun to think about anyway. But, that's it for the news of the week. And again, we have you folks covered with our network of websites. If you want to stay up to date on what's going on in the global robotics industry, make sure you bookmark the Robot Report, Automated Warehouse, sign up for our weekly newsletters, follow us on social media, and then you'll never miss a beat. So, moving on here before we get to the featured interview quickly, just plugging a webinar that's coming up this Tuesday, September 16 at noon Eastern, where Mike's going to be, leading a panel that looks at robotics for small and mid sized manufacturers. We're going to be diving in to really learn how robotics can help these manufacturers boost productivity, overcome labor shortages and improve their bottom lines.
00:26:22 · Steve Crowe, Mike Oitzman
He's joined by a great panel of folks who, are leading the charge on these fronts. So Saman Farid, he's the founder and CEO of Formic James Taylor, the CCO of On Robot, and good friend Eric Nieves, the CEO and co founder of Plus One are all going to be joining Mike, again on Tuesday, September 16 at noon Eastern. This is available for free. And they're going to be diving into all sorts of topics and learning about where all of the stuff is going. So if you play in this space or just want to learn about what's going on with robotics for manufacturers, please sign up for this webinar. You can do that by going to therobotreport.com/webinars. We'll make sure to include a link here in the show notes. But, well, there's a lot of other webinars that you can see as well. But Mike, sounds like it's going be a [00:27:08] S1: great panel. It is. Yeah. Looking forward to it. Great. Great to have those gentlemen on this webinar. And so be sure to sign up. Now, Steve, let's see. You... The interview this week is with Scott Kuindersma from Boston Dynamics. And and you and... As you said, you and Gene had a chance to talk to Scott last week. What'd you guys talk about?
00:27:27 · Steve Crowe
[00:27:27] S0: Yeah. So this was... We're supposed to do this probably a week ago. And, I unfortunately had to reschedule the interview with Scott. So I appreciate his flexibility on this. But we... This was just a complete reaction to the video they released of the Atlas humanoid. I think they called it the Spot Workshop where Atlas was taking pieces of the the quadruped robots out of these bins and putting them into different bins. It was legs and faceplates and all sorts of things. And it was just every time they release a video of Atlas, I think it just captivates the robotics world. Right? It's like, how are they doing this? You know, it was completely autonomous. They played it at one x speed. Atlas is adapting to changes in the environments in the environment, which we've seen these guys do, a number of times over the years, but there's somebody, you know, using a hockey stick to shut the lid to the bins and moving the lids and putting pieces of the Spot quadruped on the floor. So it's it's trying to see how adept is Atlas at reacting to changes in its environment in real time, which we all know really for all of these robot form factors to be more useful going forward, they need to be adaptive and reactionary and no environment is going to be perfect. And so we just really unpacked how are they doing this? So we talk a lot about large behaviour models. These are, you know, techniques that are invented by Toyota Research Institute, which Boston Dynamics has partnered with on this. So we talk about what are LBMs? How do they work? What is the data pipeline process? What is the setup process to train these LBMs? What are some of the limitations of them right now? Like what does Boston Dynamics need more of? Do they need more compute? Do they need more data to help these LBMs work better than they're already working? We talk about the future of Atlas. When does the stuff with Hyundai and the tests with Hyundai, which owns Boston Dynamics? When does that start? So we dive into everything with Scott Kuindersma. Again, he's the VP of Research at Boston Dynamics. Great guy, great speaker. So we dive into all sorts of things about Atlas and humanoids. It's a great interview. So, take a listen. Alright. We talked about this, what, a couple of weeks ago, I think, when this...
00:29:51 · Steve Crowe, Scott Kuindersma
When Boston Dynamics released this video of its new Atlas humanoid doing some work with spot parts, picking up legs and faceplates and putting them in one bin and adding them to another bin. Fascinating video, over a million views on YouTube at this point. So fortunate enough to be joined by Scott Kuindersma. He's the VP of robotics research at Boston Dynamics. He's been there for almost seven and a half years now. He's obviously seen some fascinating things. So he's going to help us unpack as much as he can large behavior models and how they're helping Atlas learn new skills and generalize to new environments. So Scott, thanks so much for being there. How are doing, man? [00:30:34] S4: I'm doing well. Thanks for having me. [00:30:36] S0: Yeah, and all the pleasure is ours. Open invitation anytime you guys want to come on the podcast. We're here for you. So the video that we're trying to break down, was called the Spot Workshop, I guess the task that you guys were undertaking here. And I always find it fascinating, I shouldn't, maybe there's a very simple reason that you guys pick this specific task. I always love when robots are working with other robots or picking pieces or parts of other robots. So again, you guys are using Atlas to pick parts of the Spot quadruped. Why this task? Why was this important enough to undertake and film? Explain the backstory as much as you can.
00:31:17 · Scott Kuindersma, Steve Crowe
[00:31:17] S4: Yeah, for sure. So oftentimes when we pick tasks to work on to develop new algorithms, new ideas at Boston Dynamics, we're usually trying to isolate some features, like some some behavioral features or some technical properties of those tasks. And then the semantic content of those tasks can just sort of be whatever. And sometimes we'll decide to make it fun or sometimes make it, like, on brand, like, with the Spot parts. So here, what we are trying to do is pick tasks that involved doing bimanual manipulation over the entire robot's workspace, so being able to pick things up from very low, put them up high, and move around with environmental constraints. So shelves that are close to the robot or bins that have to be pulled between the legs. And reason for that was the strategy that we're using to program the robot here is a demonstration based programming, right? So we're using teleoperation as a way to demonstrate examples of the behavior that we want the robot to perform. And then we're using basically machine learning to take that data and turn it into a model that can perform many of the tasks that we demonstrated. And one of the things we're interested in is, of course, we've seen a lot of people using these types of techniques in the last couple of years. And a lot of that's focused on kind of largely sort of tabletop style manipulation. And [00:32:49] S0: one of