{"version": "1.0.0", "segments": [{"startTime": 0.33, "endTime": 11.369, "body": "Over time, I expect robots to be that intelligent using the data from simulation, so the robot will start to generate its own mission based on the knowledge and the information it can get out of the digital twin.", "speaker": "Péter Fankhauser"}, {"startTime": 24.93, "endTime": 30.229, "body": "Welcome to a brand new episode of our podcast, Human and AI Mind Machine and the great in ascent.", "speaker": "Speaker 2"}, {"startTime": 31.37, "endTime": 41.36, "body": "Thanks for tuning in to our geeky podcast where we discuss the fascinating field of AI and machine learning, corporate craziness, passion for technology, and the role of humans.", "speaker": "Speaker 3"}, {"startTime": 41.65, "endTime": 44.549, "body": "We are Uli and Aburi, your host for this episode.", "speaker": "Speaker 3"}, {"startTime": 44.95, "endTime": 53.56, "body": "And now let's dive into a topic that has been always quite emotionally discussed, uh, specifically in the general public, I guess, along the hype of and buzz on AI.", "speaker": "Speaker 2"}, {"startTime": 53.83, "endTime": 55.56, "body": "Let's talk about autonomous robots.", "speaker": "Speaker 2"}, {"startTime": 55.79, "endTime": 61.549, "body": "Market size, autonomous mobile robots has been estimated in 2023 to be valued of 1%.", "speaker": "Speaker 2"}, {"startTime": 62.009, "endTime": 73.889, "body": "$8 billion US dollar and expected to grow by a kegger as we call it, right, and a gross rate of 17%, so reaching a market valuation of $4.1 billion in 2028.", "speaker": "Speaker 2"}, {"startTime": 73.97, "endTime": 78.669, "body": "So that's big market, big money and quite some numbers.", "speaker": "Speaker 2"}, {"startTime": 80.029, "endTime": 82.629, "body": "Yeah, AMRs are certainly super exciting.", "speaker": "Speaker 3"}, {"startTime": 82.79, "endTime": 86.669, "body": "Um, so let's dive deeper into the field of AI and robotics.", "speaker": "Speaker 3"}, {"startTime": 86.79, "endTime": 93.269, "body": "And Oli, you know what, we, we got just the right guest, an expert in the field with us today.", "speaker": "Speaker 3"}, {"startTime": 93.349, "endTime": 99.769, "body": "So get ready to meet Peter Frankhauser, whose vision is to create a workforce of autonomous robots.", "speaker": "Speaker 3"}, {"startTime": 100.11, "endTime": 100.949, "body": "Peter is driven.", "speaker": "Speaker 3"}, {"startTime": 101.035, "endTime": 107.254, "body": "By a passion for technological advancement and a desire to create a positive impact for society.", "speaker": "Speaker 3"}, {"startTime": 107.555, "endTime": 119.985, "body": "Peter developed perception and motion planning algorithms for legged robot robots in challenging environments at ETH Zurich and is now the co-founder and CEO of Enibotics.", "speaker": "Speaker 3"}, {"startTime": 120.345, "endTime": 122.055, "body": "Peter, it's so great to have you.", "speaker": "Speaker 3"}, {"startTime": 122.714, "endTime": 125.775, "body": "Thank you so much, both Julie and Aubrey, pleasure to be here today.", "speaker": "Péter Fankhauser"}, {"startTime": 126.98, "endTime": 133.529, "body": "So I guess to keep things fresh and fun and also appeal to the Gen Z's in the audience,", "speaker": "Speaker 3"}, {"startTime": 133.97, "endTime": 141.47, "body": "let's in the audience, you know, press press, press, right, if you're Gen Z in the audience.", "speaker": "Speaker 2"}, {"startTime": 142.399, "endTime": 143.13, "body": "No, for sure.", "speaker": "Speaker 3"}, {"startTime": 143.32, "endTime": 146.91, "body": "So let's kick things off with some hashtag magic maybe.", "speaker": "Speaker 3"}, {"startTime": 147.279, "endTime": 152.74, "body": "So, Peter, what three hashtags would you use to sum up your awesomeness?", "speaker": "Speaker 3"}, {"startTime": 154.199, "endTime": 154.809, "body": "Appreciate it.", "speaker": "Péter Fankhauser"}, {"startTime": 154.88, "endTime": 162.96, "body": "So, I think that the three pillars, uh, that I right now focus on, number 1, entrepreneurship, building a company.", "speaker": "Péter Fankhauser"}, {"startTime": 163.039, "endTime": 168.759, "body": "I love creating things, building new things, um, you know, just being out there and, and see what can be done.", "speaker": "Péter Fankhauser"}, {"startTime": 168.839, "endTime": 173.479, "body": "And so, This is a life passion for me, uh, building antibiotics together with my colleagues.", "speaker": "Péter Fankhauser"}, {"startTime": 174.27, "endTime": 176.07, "body": "Hashtag 2 would be technology.", "speaker": "Péter Fankhauser"}, {"startTime": 176.27, "endTime": 176.86, "body": "I am a geek.", "speaker": "Péter Fankhauser"}, {"startTime": 176.949, "endTime": 179.649, "body": "You mentioned in the beginning, it's a podcast for geek, love that.", "speaker": "Péter Fankhauser"}, {"startTime": 179.99, "endTime": 183.059, "body": "Um, so robotics is in the center of what I do.", "speaker": "Péter Fankhauser"}, {"startTime": 183.429, "endTime": 189.029, "body": "I did my master's studying robotics, my PhD and have, have been following that for a very long time.", "speaker": "Péter Fankhauser"}, {"startTime": 189.5, "endTime": 191.35, "body": "Some people even call me a robot.", "speaker": "Péter Fankhauser"}, {"startTime": 191.47, "endTime": 194.229, "body": "I'm very logical and data driven, try to be that way.", "speaker": "Péter Fankhauser"}, {"startTime": 194.309, "endTime": 196.589, "body": "So certainly there's that, that technology sense.", "speaker": "Péter Fankhauser"}, {"startTime": 196.63, "endTime": 200.69, "body": "And, and lastly, um, I really strive to create hashtag impact.", "speaker": "Péter Fankhauser"}, {"startTime": 201.07, "endTime": 205.47, "body": "So for me, just technology for technology's sake, it's You know, not the main topic.", "speaker": "Péter Fankhauser"}, {"startTime": 205.509, "endTime": 210.88, "body": "It's about creating impact for people, the environment, essentially making the world a better place.", "speaker": "Péter Fankhauser"}, {"startTime": 210.96, "endTime": 217.279, "body": "So for me, really concretely putting a smile on people's faces is what Uh, yeah, gets me up in the morning.", "speaker": "Péter Fankhauser"}, {"startTime": 217.36, "endTime": 218.009, "body": "This is what I do.", "speaker": "Péter Fankhauser"}, {"startTime": 218.32, "endTime": 224.399, "body": "I already have a smile on my face and to you because that's, that's pretty, Peter, great to meet you.", "speaker": "Speaker 2"}, {"startTime": 224.6, "endTime": 235.039, "body": "So there's a saying, and I, I also resonates somehow with my wives and my kids at home, they say like, you know, I want an AI to do my laundry and dishes, not my writing or any creative task, right?", "speaker": "Speaker 2"}, {"startTime": 235.119, "endTime": 235.929, "body": "So can you help us?", "speaker": "Speaker 2"}, {"startTime": 236.039, "endTime": 241.58, "body": "So when will my shirts be folded by an affordable robots robotic systems, right?", "speaker": "Speaker 2"}, {"startTime": 242.58, "endTime": 249.22, "body": "But if I had a penny for every time that question was asked, let me tell you a little bit of an anecdote.", "speaker": "Péter Fankhauser"}, {"startTime": 250.009, "endTime": 255.419, "body": "Interestingly, we, you know, we sometimes underestimate what a robot can or should do or how hard it is.", "speaker": "Péter Fankhauser"}, {"startTime": 255.5, "endTime": 264.94, "body": "So the home, the home is, especially if you have a home with kids, it's one of the hardest challenges in an environment for a robot to navigate.", "speaker": "Péter Fankhauser"}, {"startTime": 265.519, "endTime": 271.2, "body": "What seemingly seems very easy for a human is often, you know, very difficult for a robot.", "speaker": "Péter Fankhauser"}, {"startTime": 271.519, "endTime": 281.119, "body": "So until we have a, you know, a cheap versatile robot that can not only do your laundry but dishes, all of that, I'm not sure if that's, you know, the right format, um, but it.", "speaker": "Péter Fankhauser"}, {"startTime": 281.234, "endTime": 283.225, "body": "Certainly will take many, many years until we get there.", "speaker": "Péter Fankhauser"}, {"startTime": 283.554, "endTime": 285.875, "body": "However, robots are already here, right?", "speaker": "Péter Fankhauser"}, {"startTime": 286.265, "endTime": 291.834, "body": "Sometimes we forget that in manufacturing, if you own a car, it's been built by a robot.", "speaker": "Péter Fankhauser"}, {"startTime": 292.105, "endTime": 305.595, "body": "If you get a parcel from Amazon, they have 750,000 robots moving your parcels around, and now robotics coming to the next level, you know, industrial work, wherever you have people doing a lot of work in the power, in mining, these kind of environments.", "speaker": "Péter Fankhauser"}, {"startTime": 305.665, "endTime": 307.015, "body": "So robotics is here.", "speaker": "Péter Fankhauser"}, {"startTime": 307.48, "endTime": 315.6, "body": "Not in our home yet or, you know, doing your, your, your laundry, um, but I think it's a very exciting journey as we observe robotics entering more and more environments.", "speaker": "Péter Fankhauser"}, {"startTime": 317.029, "endTime": 323.049, "body": "And why did you decide to go into the field of industry where you're staying um in the in the home market?", "speaker": "Speaker 3"}, {"startTime": 325.279, "endTime": 347.67, "body": "So what we, where we came from is I joined a group of researchers at ETI Zurich in Switzerland where our vision is to take robots out in complex, complex environments, and for us this meant we build robots with four legs because they can move anywhere where people can walk, meaning they can climb over stairs, steps, move into narrow environments, move in all directions.", "speaker": "Péter Fankhauser"}, {"startTime": 347.92, "endTime": 352.079, "body": "There's this fascinating form factor that allows robots to conquer new environments.", "speaker": "Péter Fankhauser"}, {"startTime": 352.559, "endTime": 365.44, "body": "And what happened is that quite soon once we published that research work on YouTube videos that operators of facilities contacted us and say, look, what you have here is very interesting because we need the eyes and ears inside of our facilities.", "speaker": "Péter Fankhauser"}, {"startTime": 365.76, "endTime": 371.64, "body": "We need to send people in to observe if things are productive and safe, but we don't have enough people to go around.", "speaker": "Péter Fankhauser"}, {"startTime": 371.709, "endTime": 375.38, "body": "It's very hard to find those young people who can and want to do those types of jobs.", "speaker": "Péter Fankhauser"}, {"startTime": 375.72, "endTime": 378.559, "body": "It's often very remote and dangerous situations.", "speaker": "Péter Fankhauser"}, {"startTime": 378.79, "endTime": 381.989, "body": "And actually people are not great at constantly doing the same things.", "speaker": "Péter Fankhauser"}, {"startTime": 382.149, "endTime": 385.739, "body": "Our eyes and ears and our brains cannot process that large amount of information.", "speaker": "Péter Fankhauser"}, {"startTime": 385.79, "endTime": 387.329, "body": "We cannot pick up all the signals.", "speaker": "Péter Fankhauser"}, {"startTime": 387.67, "endTime": 392.209, "body": "And so we had the technology and we also really fell in love then with that problem.", "speaker": "Péter Fankhauser"}, {"startTime": 392.35, "endTime": 405.109, "body": "I spent the first couple of months and years at these facilities, learning the ins and outs, and I think it's a match made in heaven to help the facility run the facilities, you know, more sustainably and productively while helping the people that are there.", "speaker": "Péter Fankhauser"}, {"startTime": 405.869, "endTime": 410.01, "body": "And having you in this podcast today also feels like a match made in heaven.", "speaker": "Speaker 3"}, {"startTime": 410.35, "endTime": 426.079, "body": "And you just um touched upon uh what antibiotics does and is, but um we also read that you were just ranked as one of the top 5 hottest companies in the field of robotics and AI by Uh, HP in 2024.", "speaker": "Speaker 3"}, {"startTime": 426.179, "endTime": 427.88, "body": "So congratulations on that.", "speaker": "Speaker 3"}, {"startTime": 428.269, "endTime": 431.299, "body": "Can you give us maybe an elevator pitch of antibiotics?", "speaker": "Speaker 3"}, {"startTime": 431.38, "endTime": 432.869, "body": "So what's really your mission?", "speaker": "Speaker 3"}, {"startTime": 432.98, "endTime": 436.929, "body": "What's the unique challenges that you're tackling in the field of robotics?", "speaker": "Speaker 3"}, {"startTime": 437.14, "endTime": 439.559, "body": "So we are really curious to hear more about that.", "speaker": "Speaker 3"}, {"startTime": 440.51, "endTime": 441.35, "body": "Absolutely.", "speaker": "Péter Fankhauser"}, {"startTime": 442.119, "endTime": 448.399, "body": "You know, companies in the future and today will not only hire people, but they will hire robots to get the job done.", "speaker": "Péter Fankhauser"}, {"startTime": 448.839, "endTime": 453.92, "body": "So our vision is to bring robotics into industries that so far could not profit of robotics.", "speaker": "Péter Fankhauser"}, {"startTime": 453.989, "endTime": 454.82, "body": "So we work.", "speaker": "Péter Fankhauser"}, {"startTime": 455.309, "endTime": 461.149, "body": "Essentially to create the workforce of autonomous robots to help the critical infrastructure.", "speaker": "Péter Fankhauser"}, {"startTime": 461.269, "endTime": 473.589, "body": "Imagine power, utilities, sewage systems, construction sites, oil and gas, offshore, wind farms, solar farms, nuclear industries, right, all of those facilities, which requires tremendous amount of manual work.", "speaker": "Péter Fankhauser"}, {"startTime": 473.67, "endTime": 475.339, "body": "You have constant presence of people.", "speaker": "Péter Fankhauser"}, {"startTime": 475.67, "endTime": 480.549, "body": "We help those industries to run their facilities more safely, more productively.", "speaker": "Péter Fankhauser"}, {"startTime": 481.029, "endTime": 486.51, "body": "More sustainably and how we do that is by extending the human workforce with a robotic workforce.", "speaker": "Péter Fankhauser"}, {"startTime": 486.869, "endTime": 495.89, "body": "So we build today four-legged robots that are fully autonomous, meaning that you can install them once and they can do the tasks multiple times a day by themselves.", "speaker": "Péter Fankhauser"}, {"startTime": 496.35, "endTime": 497.64, "body": "What they do is collect data.", "speaker": "Péter Fankhauser"}, {"startTime": 497.79, "endTime": 498.739, "body": "This is step number 1.", "speaker": "Péter Fankhauser"}, {"startTime": 499.07, "endTime": 500.66, "body": "They use high-end sensors.", "speaker": "Péter Fankhauser"}, {"startTime": 500.779, "endTime": 504.07, "body": "Imagine cameras, thermal camera, acoustic sensors.", "speaker": "Péter Fankhauser"}, {"startTime": 504.309, "endTime": 507.75, "body": "So it's often their sense of seeing, feeling, hearing.", "speaker": "Péter Fankhauser"}, {"startTime": 507.989, "endTime": 509.549, "body": "And we replicate that in a robot.", "speaker": "Péter Fankhauser"}, {"startTime": 509.709, "endTime": 512.78, "body": "Now the robot can do that, you know, for many, many hours without getting tired.", "speaker": "Péter Fankhauser"}, {"startTime": 513.03, "endTime": 514.349, "body": "It can see more than the human.", "speaker": "Péter Fankhauser"}, {"startTime": 514.39, "endTime": 516.02, "body": "It can see, you know, hotspots.", "speaker": "Péter Fankhauser"}, {"startTime": 516.26, "endTime": 517.669, "body": "It can hear sounds that we don't.", "speaker": "Péter Fankhauser"}, {"startTime": 517.75, "endTime": 524.469, "body": "It can smell gasses that we don't, and it can then combine that data and immediately report on any anomaly.", "speaker": "Péter Fankhauser"}, {"startTime": 524.539, "endTime": 534.63, "body": "So the trick here is how quickly can you find a problem in your facility, report that back such that the human can then look at the information and to take the right action.", "speaker": "Péter Fankhauser"}, {"startTime": 535.039, "endTime": 537.679, "body": "That reduces the maintenance work you need to do.", "speaker": "Péter Fankhauser"}, {"startTime": 537.799, "endTime": 545.44, "body": "You can stop unplanned downtime and you save a lot of energy by finding leakages early and eventually you can avoid accidents.", "speaker": "Péter Fankhauser"}, {"startTime": 546.669, "endTime": 547.669, "body": "Super cool actually.", "speaker": "Speaker 2"}, {"startTime": 547.95, "endTime": 551.349, "body": "So my favorite digital ethics dude is Floridi.", "speaker": "Speaker 2"}, {"startTime": 551.469, "endTime": 568.01, "body": "I don't know, Italian, you know, UK centric ethics, and he says like, you know, in the world of AI we actually cheo fans, so we are not living in a world of, you know, overlaying AI intelligence, let's say also robotics, but we're actually separating, we're fencing it out, right?", "speaker": "Speaker 2"}, {"startTime": 568.34, "endTime": 572.59, "body": "Um, that means if you go in the manufacturing side on a on a factory, right, there there's.", "speaker": "Speaker 2"}, {"startTime": 572.7, "endTime": 583.07, "body": "Robotic system there and moves vehicles and, and, or, you know, parts and pieces, tons of, you know, of weight, but it's actually separated and fenced out there.", "speaker": "Speaker 2"}, {"startTime": 583.229, "endTime": 595.19, "body": "It seems to be your approach is somehow a bit of a different angle, like, you know, overcoming the, let's say the clear separation that this is where I place the field and here is where the human plays in the field there.", "speaker": "Speaker 2"}, {"startTime": 595.4, "endTime": 598.729, "body": "What's your stake on that on these kinds of separations?", "speaker": "Speaker 2"}, {"startTime": 599.979, "endTime": 607.719, "body": "Totally makes sense in the beginning to look at it separately because these are still machines that carry a lot of energy, so safety, you know, is a key aspect.", "speaker": "Péter Fankhauser"}, {"startTime": 608.179, "endTime": 620.26, "body": "Now in the industries that we work in and imagine, you know, multiple floors, metal, but it's still outdoors, but the stairs in between, it's a very, you know, condense and complex environment.", "speaker": "Péter Fankhauser"}, {"startTime": 620.76, "endTime": 628.849, "body": "There's going to be humans and so to be able to send in robots into that environment they need to be human friendly and the humans also need to be taught how to deal around robots.", "speaker": "Péter Fankhauser"}, {"startTime": 628.929, "endTime": 631.369, "body": "So it's, it's meeting of those two work groups.", "speaker": "Péter Fankhauser"}, {"startTime": 631.65, "endTime": 638.289, "body": "So yes, they will be sharing safely the same environment, but they're not going to do the same tasks.", "speaker": "Péter Fankhauser"}, {"startTime": 638.405, "endTime": 644.005, "body": "It's still about making sure that, you know, what are humans really good at and separating that from what are robots really good at.", "speaker": "Péter Fankhauser"}, {"startTime": 644.044, "endTime": 645.724, "body": "It's not about, you know, replicating that.", "speaker": "Péter Fankhauser"}, {"startTime": 645.844, "endTime": 649.424, "body": "So human is creative, it's intuitive, and it's flexible.", "speaker": "Péter Fankhauser"}, {"startTime": 649.645, "endTime": 650.145, "body": "We're much smarter.", "speaker": "Péter Fankhauser"}, {"startTime": 650.205, "endTime": 653.484, "body": "We can adapt to new challenges, and we're also individual.", "speaker": "Péter Fankhauser"}, {"startTime": 653.594, "endTime": 655.965, "body": "Every person brings in, you know, their strength to the work.", "speaker": "Péter Fankhauser"}, {"startTime": 656.33, "endTime": 658.76, "body": "A robot, on the other hand, is very predictable, right?", "speaker": "Péter Fankhauser"}, {"startTime": 658.809, "endTime": 660.609, "body": "It's very logical, you know why it's doing it.", "speaker": "Péter Fankhauser"}, {"startTime": 660.809, "endTime": 662.28, "body": "It's much more precise than human.", "speaker": "Péter Fankhauser"}, {"startTime": 662.45, "endTime": 669.02, "body": "It's also consistent and repeatable, meaning that if you take one robot out, you replace it with the same model, it'll do the same thing, right?", "speaker": "Péter Fankhauser"}, {"startTime": 669.33, "endTime": 680.7, "body": "So it's about understanding how, you know, what are the capabilities of humans and robots and Putting them in the right, uh, kind of at the right tasks, but yes, they can share the workspace and coexist and collaborate with each other.", "speaker": "Péter Fankhauser"}, {"startTime": 681.789, "endTime": 686.909, "body": "So speaking of USP, maybe, you know, what makes antibiotics so different from its competitor, right?", "speaker": "Speaker 2"}, {"startTime": 686.95, "endTime": 691.21, "body": "Or is there a specific aspect, achievement, milestones, you know, you're super proud of?", "speaker": "Speaker 2"}, {"startTime": 692.179, "endTime": 702.369, "body": "Yes, I think fundamentally, and when we came out of the university, what we saw is the biggest risk of such a company is that you continue what you're doing and just, you know, continue researching.", "speaker": "Péter Fankhauser"}, {"startTime": 702.619, "endTime": 703.14, "body": "That's fascinating.", "speaker": "Péter Fankhauser"}, {"startTime": 703.219, "endTime": 706.94, "body": "I would love to do that, but it doesn't lead to that impact that I mentioned in the beginning.", "speaker": "Péter Fankhauser"}, {"startTime": 707.26, "endTime": 718.099, "body": "So for us it was very important to from the beginning, be very customer centric and understand the market, and that has led us down a path of very, you know, interesting, you know, unique value propositions.", "speaker": "Péter Fankhauser"}, {"startTime": 718.369, "endTime": 725.19, "body": "And so for example, we were the first ones to really deploy very focused, these four-legged systems for industrial inspection.", "speaker": "Péter Fankhauser"}, {"startTime": 725.21, "endTime": 727.32, "body": "That's a market we we we uncovered first.", "speaker": "Péter Fankhauser"}, {"startTime": 727.809, "endTime": 742.409, "body": "We didn't build very specific robots, so our robots are water and dustproof because we know, we understand that these systems are going to be used for a long time outdoors in rainy environment, in corrosive environments, offshore saltwater, in dusty environments.", "speaker": "Péter Fankhauser"}, {"startTime": 742.84, "endTime": 745.83, "body": "So we built the architecture to cater to that industry.", "speaker": "Péter Fankhauser"}, {"startTime": 746.349, "endTime": 748.099, "body": "Then quite soon we learned around.", "speaker": "Péter Fankhauser"}, {"startTime": 748.669, "endTime": 756.53, "body": "The, you know, the needs in the chemicals industry, in the oil and gas industry where there's hazardous substances, explosive substance, substances, so you're not allowed to ignite a spark.", "speaker": "Péter Fankhauser"}, {"startTime": 756.989, "endTime": 766.119, "body": "So that's why we're the world's first and only company building an X-proof called an explosion proof robot for these environments that, that is, you know, of the legged form factor.", "speaker": "Péter Fankhauser"}, {"startTime": 766.71, "endTime": 771.789, "body": "And then, you know, there's there's more topics which make us unique, such as, you know, our autonomy is extremely powerful.", "speaker": "Péter Fankhauser"}, {"startTime": 771.909, "endTime": 780.09, "body": "We believe that a robotic fleet can only work when autonomy is at the core, and one supervisor can look at dozens of robots at once.", "speaker": "Péter Fankhauser"}, {"startTime": 780.39, "endTime": 785.19, "body": "And so how we then also integrate with digital twins and simulations has been a pioneering topic for us.", "speaker": "Péter Fankhauser"}, {"startTime": 786.63, "endTime": 791.679, "body": "I wonder because what you're up to is such a huge, huge topic.", "speaker": "Speaker 3"}, {"startTime": 791.799, "endTime": 801.239, "body": "I guess it probably takes a village to, to really um build on that vision and continue like making those robots better and better every day.", "speaker": "Speaker 3"}, {"startTime": 801.599, "endTime": 808.7, "body": "So, what role do partnerships and your ecosystem actually play in the development of antibiotics?", "speaker": "Speaker 3"}, {"startTime": 809.0, "endTime": 812.219, "body": "And what role do other partners play in that?", "speaker": "Speaker 3"}, {"startTime": 813.119, "endTime": 813.84, "body": "Absolutely.", "speaker": "Péter Fankhauser"}, {"startTime": 814.03, "endTime": 819.119, "body": "Robotics is very complex, so we have to focus on the things we're really good at.", "speaker": "Péter Fankhauser"}, {"startTime": 819.2, "endTime": 822.359, "body": "So if something somebody else can do, we shouldn't do it, right?", "speaker": "Péter Fankhauser"}, {"startTime": 822.4, "endTime": 823.539, "body": "So this is where we partner.", "speaker": "Péter Fankhauser"}, {"startTime": 824.039, "endTime": 834.7, "body": "Um, so, for example, on certain components, we're collaborating with Maxzone, that's a Swiss motor company, because we built our own motors in the beginning, but then quite soon you realize there's others.", "speaker": "Péter Fankhauser"}, {"startTime": 834.799, "endTime": 841.0, "body": "We bring in certain IP but in a collaboration, we can build a specific motor that's really optimized for our, for our robot.", "speaker": "Péter Fankhauser"}, {"startTime": 841.679, "endTime": 850.14, "body": "Then simulation tools, right, that we work closely with Nvidia to, you know, get ever closer to bridging the gap between the reality and simulations.", "speaker": "Péter Fankhauser"}, {"startTime": 850.719, "endTime": 858.039, "body": "It's with, you know, AWS for all the data storage, cloud architecture, these topics, and then it's certainly, you know, with partners like Siemens.", "speaker": "Péter Fankhauser"}, {"startTime": 858.489, "endTime": 860.78, "body": "It's about to go to market integrations.", "speaker": "Péter Fankhauser"}, {"startTime": 861.14, "endTime": 865.099, "body": "It's the, you know, the the domain expertise that we don't have.", "speaker": "Péter Fankhauser"}, {"startTime": 865.34, "endTime": 872.08, "body": "So let's partner with those, you know, excellent partners, those companies, uh, to make what we have even bigger and scale it around the world.", "speaker": "Péter Fankhauser"}, {"startTime": 884.44, "endTime": 893.08, "body": "So if we, if we took, you know, years and century, maybe a couple of more centuries backwards, right, robotics has been always a great field, right?", "speaker": "Speaker 2"}, {"startTime": 893.239, "endTime": 917.0, "body": "And Um, you know, from really early days, and I just learned that, you know, a mathematician architas of Tarentum actually postulated actually the first, you know, a bit of somehow the first piggin, you know, robotic system which was proper by steam, at least on a, you know, on the chart I seemed to be right, and then you have obviously famous metropolis and then you have all the, you know, you know, movies inspired, let's say aspects.", "speaker": "Speaker 2"}, {"startTime": 917.539, "endTime": 937.179, "body": "Maybe, you know, um, spinning out of ETH, right, and it seemed to be the first experiments, the first launch, but also things, the first, you know, tangible piece of robotic smart, whatever smartness was there back in the days in 2009, I think it was called AOF, um, and now your current version seemed to be since 2021.", "speaker": "Speaker 2"}, {"startTime": 937.715, "endTime": 939.145, "body": "Animal Gen D, right?", "speaker": "Speaker 2"}, {"startTime": 939.325, "endTime": 940.784, "body": "Generation D, I guess, right?", "speaker": "Speaker 2"}, {"startTime": 941.085, "endTime": 943.364, "body": "What, what actually has changed, right?", "speaker": "Speaker 2"}, {"startTime": 944.155, "endTime": 946.604, "body": "Can you, can you summarize the first prototype?", "speaker": "Speaker 2"}, {"startTime": 946.645, "endTime": 950.695, "body": "It's like, well, it's still a thing, but it's just a different shell now, I guess not, right?", "speaker": "Speaker 2"}, {"startTime": 951.085, "endTime": 954.195, "body": "I'm happy to take you a little bit on, you know, a spin of the history.", "speaker": "Péter Fankhauser"}, {"startTime": 954.284, "endTime": 957.174, "body": "The first robot that we built, Alof in 2009.", "speaker": "Péter Fankhauser"}, {"startTime": 957.809, "endTime": 960.95, "body": "It was just showing the world that you can build a four-legged system.", "speaker": "Péter Fankhauser"}, {"startTime": 961.169, "endTime": 970.01, "body": "Now we took a very classical component off the shelf, classical industrial motors, gear, electronics, lots and lots of cables, right?", "speaker": "Péter Fankhauser"}, {"startTime": 970.33, "endTime": 978.33, "body": "And we were able to show that the robot could walk on 4 legs, but you have to imagine it a little bit like a walking table, very stiff, no adaptation.", "speaker": "Péter Fankhauser"}, {"startTime": 978.409, "endTime": 985.57, "body": "There was, it, it wasn't, you know, if you look at the dog, it runs around, it climbs, it's much more athletic and dynamic.", "speaker": "Péter Fankhauser"}, {"startTime": 986.239, "endTime": 993.719, "body": "So on the path to what we have today, there certainly have been 3 topics that that, you know, happened through us and through the environment.", "speaker": "Péter Fankhauser"}, {"startTime": 993.789, "endTime": 1001.39, "body": "So number 1 is How to build the actuation, the motor, how do you know that that that actuation system.", "speaker": "Péter Fankhauser"}, {"startTime": 1002.049, "endTime": 1005.01, "body": "What we realized quite soon is if you look at the dog or even humans.", "speaker": "Péter Fankhauser"}, {"startTime": 1005.659, "endTime": 1009.669, "body": "It doesn't matter so much where is your foot placed at.", "speaker": "Péter Fankhauser"}, {"startTime": 1009.859, "endTime": 1013.7, "body": "It's much more, can you feel the reaction to when you step down?", "speaker": "Péter Fankhauser"}, {"startTime": 1013.979, "endTime": 1019.719, "body": "So if you walk on concrete versus if you walk on ice, your foot will slip, but you react very quickly to it and you walk more carefully.", "speaker": "Péter Fankhauser"}, {"startTime": 1020.26, "endTime": 1027.579, "body": "That reaction is determined by the sense of touch and your muscles that are not, that are not stiff, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1027.78, "endTime": 1028.699, "body": "They're elastic.", "speaker": "Péter Fankhauser"}, {"startTime": 1029.418, "endTime": 1032.379, "body": "Because so that it can elongate and be compressed.", "speaker": "Péter Fankhauser"}, {"startTime": 1032.899, "endTime": 1037.379, "body": "So that's systems we replicated on a robot, so we added springs to our motor.", "speaker": "Péter Fankhauser"}, {"startTime": 1037.418, "endTime": 1042.639, "body": "So instead of having a stiff motor in the gear, we add a spring in between that's quite flexible.", "speaker": "Péter Fankhauser"}, {"startTime": 1043.178, "endTime": 1048.178, "body": "Now as the robot walks, that allows you to measure the force because you can measure the deflection of the spring.", "speaker": "Péter Fankhauser"}, {"startTime": 1048.699, "endTime": 1051.609, "body": "It allows you to store energy as well, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1051.78, "endTime": 1063.959, "body": "So if you jump and walk, the energy, the spring can store energy, and it also allows you to walk dynamically because if you have impact of the foot on the ground, it doesn't shake the whole system, but you know, it softens that landing.", "speaker": "Péter Fankhauser"}, {"startTime": 1064.219, "endTime": 1072.819, "body": "So that was one of the key elements, this force control and elasticity in the springs that led to a very, firstly, a dynamic robot that could jump and run.", "speaker": "Péter Fankhauser"}, {"startTime": 1073.719, "endTime": 1079.819, "body": "Now that system is blind then at the moment, so you can do things, but it's only in very controlled environments that you preprogram it for.", "speaker": "Péter Fankhauser"}, {"startTime": 1080.68, "endTime": 1083.28, "body": "The second layer then is the perception.", "speaker": "Péter Fankhauser"}, {"startTime": 1083.359, "endTime": 1092.119, "body": "How does the robot start to see, and there has been a great revolution on making cameras smaller, is on LiDARR systems, you know, cameras.", "speaker": "Péter Fankhauser"}, {"startTime": 1092.319, "endTime": 1097.979, "body": "Today it's quite normal for us to use face ID on our mobile phone, structured lights, stereo camera, these kind of topics.", "speaker": "Péter Fankhauser"}, {"startTime": 1098.28, "endTime": 1099.5, "body": "Um, this is.", "speaker": "Péter Fankhauser"}, {"startTime": 1099.939, "endTime": 1113.959, "body": "We could then use these technologies for the robots to start to see not only pictures, but see in 3D, understand the world, understand where do you place your feet, how do you react to obstacles, and that really allowed the robots to become autonomous.", "speaker": "Péter Fankhauser"}, {"startTime": 1115.239, "endTime": 1118.819, "body": "The third piece that I'm really excited about is the AI.", "speaker": "Péter Fankhauser"}, {"startTime": 1119.119, "endTime": 1122.479, "body": "Now, in a traditional sense, we have been programming the brains of the robot.", "speaker": "Péter Fankhauser"}, {"startTime": 1123.28, "endTime": 1129.959, "body": "But that, you know, you can do it, but you end up at, you know, at the human's creativity to define all the rules on how to do that.", "speaker": "Péter Fankhauser"}, {"startTime": 1130.439, "endTime": 1134.0, "body": "Really, the magic kicked in when we allowed the robots to learn by themselves.", "speaker": "Péter Fankhauser"}, {"startTime": 1134.439, "endTime": 1138.67, "body": "And so they optimize based on a policy on the expected outcome.", "speaker": "Péter Fankhauser"}, {"startTime": 1139.04, "endTime": 1148.839, "body": "The neural network can get better and better, and with the existing hardware we can show that, you know, the systems massively outperform what a human could come in, could come up manually.", "speaker": "Péter Fankhauser"}, {"startTime": 1150.739, "endTime": 1170.68, "body": "Uh, it's super interesting what you just said in one of those layers you were mentioning were around, um, sensors and that they um changed in shape and I also um remember from my trip to San Francisco, like the, like the self-driving cars, like the Waymo cars that uh suddenly you have lighter sensors and they're like so small that they can fit on a car.", "speaker": "Speaker 3"}, {"startTime": 1171.119, "endTime": 1183.839, "body": "And um I wonder what's your perspective on, on the role that innovation has played over the past years to um to make advanced sensor technologies more um accessible.", "speaker": "Speaker 3"}, {"startTime": 1183.88, "endTime": 1184.819, "body": "I also guess.", "speaker": "Speaker 3"}, {"startTime": 1185.229, "endTime": 1186.869, "body": "So what's your view on that?", "speaker": "Speaker 3"}, {"startTime": 1187.0, "endTime": 1189.459, "body": "It's a massive enabler for robotics.", "speaker": "Péter Fankhauser"}, {"startTime": 1189.54, "endTime": 1191.599, "body": "Robotics is always a mix of different disciplines.", "speaker": "Péter Fankhauser"}, {"startTime": 1191.68, "endTime": 1198.54, "body": "So we take the newest technologies in actuation, in sensing, in battery, and compute, all of that, and we combine that to a robot.", "speaker": "Péter Fankhauser"}, {"startTime": 1198.8, "endTime": 1203.56, "body": "So whenever there's something, you know, exciting new technology in the market, it will have an impact on robotics.", "speaker": "Péter Fankhauser"}, {"startTime": 1203.64, "endTime": 1215.512, "body": "So for us, the miniaturization and you know, the commoditization to Certain sense of LDAs, of depth cameras, time of flight sensors, all of that has massively contributed to what robots can do today.", "speaker": "Péter Fankhauser"}, {"startTime": 1215.772, "endTime": 1231.826, "body": "So instead of them being fixed, right, and in a cage, as you mentioned in the beginning, just doing a repetitive A to B task to be safely to, you know, to work in an environment where also there's people, you need to understand whereas a person, you need to react to a changing environment and they These sensors have played a massive role.", "speaker": "Péter Fankhauser"}, {"startTime": 1232.186, "endTime": 1236.586, "body": "So for us it's about, yeah, the combination of different modality of sensors as well.", "speaker": "Péter Fankhauser"}, {"startTime": 1236.666, "endTime": 1238.265, "body": "There's no best sensors for robotics.", "speaker": "Péter Fankhauser"}, {"startTime": 1238.546, "endTime": 1240.656, "body": "So typically you deploy a set of different ones.", "speaker": "Péter Fankhauser"}, {"startTime": 1240.666, "endTime": 1251.765, "body": "We use often depth cameras and LDARs, fuse the data, optimize for its environment and the tasks, make sure it can perform the tasks autonomously but also reliably and safely.", "speaker": "Péter Fankhauser"}, {"startTime": 1252.829, "endTime": 1259.189, "body": "So, um, speaking on, on a bit of an AI and tech stuff, right, G AI is obviously everywhere, right, nowadays.", "speaker": "Speaker 2"}, {"startTime": 1259.3, "endTime": 1271.51, "body": "Everybody talks about, you know, chat GBT and different generative AI aspects, and there's, I, I guess also, you know, from a large language models or foundation models, quite, quite an interesting stances even so it never picked up obviously.", "speaker": "Speaker 2"}, {"startTime": 1271.635, "endTime": 1276.875, "body": "It's picked up in the consumer industry in service, maintenance, sometimes in engineering.", "speaker": "Speaker 2"}, {"startTime": 1277.114, "endTime": 1285.555, "body": "What's your view on, you know, on, on, on such foundational G AI tech stack like Palmy or RT2 and these kinds of stuff?", "speaker": "Speaker 2"}, {"startTime": 1285.625, "endTime": 1287.435, "body": "Is that something you follow up?", "speaker": "Speaker 2"}, {"startTime": 1287.505, "endTime": 1290.535, "body": "Are you contributing actually to open source stuff as well?", "speaker": "Speaker 2"}, {"startTime": 1291.114, "endTime": 1293.834, "body": "So we are contributing to open source topic.", "speaker": "Péter Fankhauser"}, {"startTime": 1294.265, "endTime": 1296.334, "body": "We're in GitHub, so there's a couple of projects there certainly.", "speaker": "Péter Fankhauser"}, {"startTime": 1297.01, "endTime": 1301.16, "body": "Um, now we're also contributing to AI a lot again with partnerships.", "speaker": "Péter Fankhauser"}, {"startTime": 1301.339, "endTime": 1310.939, "body": "Um, our take is we're not in the game just to come up with the greatest AI technologies or we're in the game of taking the best technologies and applying them to our customers.", "speaker": "Péter Fankhauser"}, {"startTime": 1311.14, "endTime": 1317.38, "body": "Now these, you know, large single foundation models, very intriguing philosophic philosophically.", "speaker": "Péter Fankhauser"}, {"startTime": 1317.79, "endTime": 1322.02, "body": "I've not seen yet the results that, you know, that we can deploy them in operations.", "speaker": "Péter Fankhauser"}, {"startTime": 1322.25, "endTime": 1323.93, "body": "So they haven't reached that maturity.", "speaker": "Péter Fankhauser"}, {"startTime": 1324.229, "endTime": 1326.8, "body": "I believe one of the tricky parts here is the data.", "speaker": "Péter Fankhauser"}, {"startTime": 1326.949, "endTime": 1336.219, "body": "So if you really deploy a one foundation model, you need so much data from a physical robot in the relevant environment that probably that doesn't exist yet.", "speaker": "Péter Fankhauser"}, {"startTime": 1336.63, "endTime": 1337.65, "body": "Eventually it will come.", "speaker": "Péter Fankhauser"}, {"startTime": 1338.189, "endTime": 1347.03, "body": "However, our approach here is actually to look at a smart architectural architecture of a, you know, specialized models which are much, much easier to train with less data.", "speaker": "Péter Fankhauser"}, {"startTime": 1347.65, "endTime": 1353.969, "body": "So for us this means our robot's mobility is trained with AI in simulation where we can do that.", "speaker": "Péter Fankhauser"}, {"startTime": 1354.42, "endTime": 1359.939, "body": "We use AI in the navigation to perceive the environment, differentiate between objects and calculate a path.", "speaker": "Péter Fankhauser"}, {"startTime": 1360.26, "endTime": 1362.939, "body": "We use AI a lot obviously in the inspection, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1363.02, "endTime": 1366.03, "body": "How do you turn all those images into valuable insights?", "speaker": "Péter Fankhauser"}, {"startTime": 1366.06, "endTime": 1367.469, "body": "Nobody wants to look at 8 hours of video.", "speaker": "Péter Fankhauser"}, {"startTime": 1367.939, "endTime": 1375.025, "body": "So how do you take frame by frame, look at what What's good, what's bad, you know, detect objects and read out the state of the environment.", "speaker": "Péter Fankhauser"}, {"startTime": 1375.084, "endTime": 1376.925, "body": "So that's the inspection results that we're doing.", "speaker": "Péter Fankhauser"}, {"startTime": 1377.285, "endTime": 1386.505, "body": "For us, this means it's a large architecture, smart architecture of these specialized models, um, then we feed constantly with new data from the field.", "speaker": "Péter Fankhauser"}, {"startTime": 1387.125, "endTime": 1391.285, "body": "Eventually this might lead in a large foundation model, but for the moment this is really the approach to go that we're taking.", "speaker": "Péter Fankhauser"}, {"startTime": 1391.829, "endTime": 1393.229, "body": "And maybe we jump directly, right?", "speaker": "Speaker 2"}, {"startTime": 1393.3, "endTime": 1406.3, "body": "So antibiotics has been leveraging simulation, digital twins, somehow the industrial metaverse, Siemens sometimes also push it uh also with, you know, additional aspects to train AI and robotic systems.", "speaker": "Speaker 2"}, {"startTime": 1406.51, "endTime": 1409.989, "body": "How does this contribute to the training and development of the robotic system?", "speaker": "Speaker 2"}, {"startTime": 1410.15, "endTime": 1412.77, "body": "Can you guide us a bit of, you know, the learning, the performance.", "speaker": "Speaker 2"}, {"startTime": 1412.88, "endTime": 1415.8, "body": "Compare it to traditional models, how does the pipeline work?", "speaker": "Speaker 2"}, {"startTime": 1416.0, "endTime": 1421.52, "body": "You know, to, to robot, um, what, what is, what is, can you give us a bit of insight?", "speaker": "Speaker 2"}, {"startTime": 1421.88, "endTime": 1422.229, "body": "Yes.", "speaker": "Péter Fankhauser"}, {"startTime": 1422.439, "endTime": 1430.979, "body": "So, so for the users who don't know connected deep reinforcement learning, what we do is essentially define and it's very similar as a child learns to walk.", "speaker": "Péter Fankhauser"}, {"startTime": 1431.239, "endTime": 1440.719, "body": "Now, there's a parent that says, hey, come here and the baby tries in the beginning, falls and struggles, but over time it learns from that experience and gets better and better.", "speaker": "Péter Fankhauser"}, {"startTime": 1441.42, "endTime": 1449.28, "body": "For a child this takes multiple months and a lot of crashes, right, and learnings, but eventually get there becomes really good, but it takes a long time.", "speaker": "Péter Fankhauser"}, {"startTime": 1450.14, "endTime": 1460.89, "body": "It's the same system we give our robots a policy, we give it a task and say this is what you optimize for in terms of speed, stability, lift your feet sufficiently, walk over the terrain.", "speaker": "Péter Fankhauser"}, {"startTime": 1460.969, "endTime": 1469.959, "body": "So this is a mathematical rule we give it, and then the robot starts to optimize and converges an artificial neural network over time.", "speaker": "Péter Fankhauser"}, {"startTime": 1470.42, "endTime": 1475.38, "body": "Now theoretically we could do that on the real robot, but again, that would take a lot of robots and a lot of time.", "speaker": "Péter Fankhauser"}, {"startTime": 1475.459, "endTime": 1478.54, "body": "So instead of doing that on a physical robot, we do that in simulation.", "speaker": "Péter Fankhauser"}, {"startTime": 1479.239, "endTime": 1481.599, "body": "There you have to imagine 1000 robots, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1481.609, "endTime": 1489.76, "body": "We can have as many as we want, also accelerated in time, practicing around, then sharing the knowledge among each other in the next round.", "speaker": "Péter Fankhauser"}, {"startTime": 1489.839, "endTime": 1497.8, "body": "So these days with the help of simulation, we can train from scratch an artificial neural network for walking within 30 minutes to 60 minutes.", "speaker": "Péter Fankhauser"}, {"startTime": 1498.91, "endTime": 1499.9, "body": "That's pretty cool.", "speaker": "Speaker 2"}, {"startTime": 1500.03, "endTime": 1500.959, "body": "That's actually pretty cool.", "speaker": "Speaker 2"}, {"startTime": 1501.189, "endTime": 1514.93, "body": "So that means um you have, I don't know, cloud services that trained on individual simulated environments, but on individual capabilities or maybe even broader scenarios like walking.", "speaker": "Speaker 2"}, {"startTime": 1515.39, "endTime": 1528.88, "body": "And then for each customer, you have certain kinds of, you know, specifics, I would say, right, the one who needs to, I don't know, identify corrosions or defects or holes or, you know, from those kinds of use cases, right?", "speaker": "Speaker 2"}, {"startTime": 1528.92, "endTime": 1547.439, "body": "And then you in addition, You know, tie up certain kinds of customization per customer project or is there, hey, it comes already with the, you know, sheer capabilities, you know, that strands to the individual customers and that's why, you know, you scale on customer sites already.", "speaker": "Speaker 2"}, {"startTime": 1548.06, "endTime": 1549.89, "body": "Yeah, you, you touch up on a couple of topics.", "speaker": "Péter Fankhauser"}, {"startTime": 1549.9, "endTime": 1555.42, "body": "Um, in terms of the mobility, what I described, training robot to walk and also navigate.", "speaker": "Péter Fankhauser"}, {"startTime": 1556.15, "endTime": 1557.27, "body": "It's always the same model.", "speaker": "Péter Fankhauser"}, {"startTime": 1557.449, "endTime": 1559.13, "body": "We, you know, there's no customization needed.", "speaker": "Péter Fankhauser"}, {"startTime": 1559.189, "endTime": 1579.619, "body": "The reason is that in stimulation we can have a huge, you know, different types, number of environments that we can replicate essentially everything and, you know, independent of the industry that we work in, typically the stairs, the railings, the steps, the cable, all of these things look exactly the same, independent if you're in a nuclear facility at a water treatment plant.", "speaker": "Péter Fankhauser"}, {"startTime": 1580.03, "endTime": 1582.0, "body": "Looks pretty much, you know, very similar.", "speaker": "Péter Fankhauser"}, {"startTime": 1582.01, "endTime": 1583.949, "body": "So in terms of the mobility and navigation.", "speaker": "Péter Fankhauser"}, {"startTime": 1584.579, "endTime": 1586.55, "body": "Now the second topic you mentioned is inspections.", "speaker": "Péter Fankhauser"}, {"startTime": 1586.76, "endTime": 1600.53, "body": "Also there we find the typical tasks of turning kind of analog gauges, lever positions into digital signals, same hotspots for rotating equipment, pumps, same topics, acoustics, these kind of things are replicate.", "speaker": "Péter Fankhauser"}, {"startTime": 1600.77, "endTime": 1606.349, "body": "But you're right, then there's a long list of things that, you know, um are industry specific.", "speaker": "Péter Fankhauser"}, {"startTime": 1607.03, "endTime": 1610.349, "body": "For example, for the nuclear industry, you you want radiation mapping.", "speaker": "Péter Fankhauser"}, {"startTime": 1610.79, "endTime": 1622.449, "body": "So what we did in the beginning is provide the core inspection capabilities out of the box, the ones that are cross-industry, and then start working with partners that are specialized for specialized industries.", "speaker": "Péter Fankhauser"}, {"startTime": 1622.8, "endTime": 1629.52, "body": "Now the beauty is that the robot has compute on board, it has the relevant sensors often we can carry extra payload to do that.", "speaker": "Péter Fankhauser"}, {"startTime": 1629.88, "endTime": 1636.859, "body": "So we're continuously extending the capabilities of this robotic worker with new, you know, skills, so.", "speaker": "Péter Fankhauser"}, {"startTime": 1637.349, "endTime": 1644.469, "body": "You know, the further we develop, then we can also, you know, the customers can extend the capabilities of the robotic workforce and then across the fleet.", "speaker": "Péter Fankhauser"}, {"startTime": 1644.55, "endTime": 1650.329, "body": "So you know, once we trained the systems, we can give it to all the robots, and that's a magical scaling moment.", "speaker": "Péter Fankhauser"}, {"startTime": 1650.969, "endTime": 1651.31, "body": "Super cool.", "speaker": "Speaker 2"}, {"startTime": 1651.43, "endTime": 1653.089, "body": "How long does it take to train an animal?", "speaker": "Speaker 2"}, {"startTime": 1654.939, "endTime": 1662.959, "body": "So for, for mobility, it's, you know, it's, it's under 1 hour and then for inspection tasks, you typically have a couple of 100 sample images.", "speaker": "Péter Fankhauser"}, {"startTime": 1663.329, "endTime": 1669.14, "body": "Um, it's less about the training than itself once the pipeline is there, it's about collecting the data.", "speaker": "Péter Fankhauser"}, {"startTime": 1669.665, "endTime": 1673.954, "body": "1, but then it's also validating, ensuring, you know, the model really does what it does.", "speaker": "Péter Fankhauser"}, {"startTime": 1674.354, "endTime": 1684.234, "body": "We ship, um, every, you know, roughly 6 to 8 weeks, we ship ship a software update to the entire fleet over the air, so you can expect your robot to get better and better over time.", "speaker": "Péter Fankhauser"}, {"startTime": 1697.31, "endTime": 1707.05, "body": "Can you share a specific examples of case studies where the usage of digital twins and simulation led to significant breakthroughs or improvements in your robotic system?", "speaker": "Speaker 3"}, {"startTime": 1708.01, "endTime": 1721.56, "body": "Mhm Typically how we set up the robot, the autonomy is that we take a tablet, we take the robot, and we walk with the customer, the operator of the facility through the facility and replicate what humans do today.", "speaker": "Péter Fankhauser"}, {"startTime": 1722.17, "endTime": 1725.329, "body": "So you manually control the robot back and forth sideways.", "speaker": "Péter Fankhauser"}, {"startTime": 1725.369, "endTime": 1730.569, "body": "You let it climb the stairs, but then you point the cameras at certain elements and you say, record, record, record.", "speaker": "Péter Fankhauser"}, {"startTime": 1731.13, "endTime": 1742.369, "body": "Within that process, which takes a couple of hours, the robot creates this understanding and its own digital twin of the world and can then, you know, you can edit that and essentially execute those missions.", "speaker": "Péter Fankhauser"}, {"startTime": 1743.449, "endTime": 1748.239, "body": "Now, if a customer has a digital twin already, we can use that for two things.", "speaker": "Péter Fankhauser"}, {"startTime": 1748.369, "endTime": 1764.17, "body": "We can simulate already the mission in advance, so tell the customer how long it's going to take, what's a good viewpoint to look at those elements, and then we can train the robot in that environment, set it up, and then bring it on site and just let it run without that installation process.", "speaker": "Péter Fankhauser"}, {"startTime": 1764.609, "endTime": 1768.52, "body": "Now we had a couple of examples where the customer had really good digital twins which are accurate.", "speaker": "Péter Fankhauser"}, {"startTime": 1768.839, "endTime": 1772.65, "body": "Within the 1st 10 minutes, the robot was already navigating that environment.", "speaker": "Péter Fankhauser"}, {"startTime": 1772.689, "endTime": 1777.15, "body": "It has never seen it, but it was able to leverage all the data available through the digital twin.", "speaker": "Péter Fankhauser"}, {"startTime": 1777.369, "endTime": 1784.51, "body": "And then obviously that's a magical moment where you can truly, you know, massively accelerate the robotics commission process.", "speaker": "Péter Fankhauser"}, {"startTime": 1786.04, "endTime": 1786.859, "body": "That's super cool.", "speaker": "Speaker 2"}, {"startTime": 1787.479, "endTime": 1799.959, "body": "So, if we look a bit of ahead, right, in the role, if you bring it on the next level, let's say in simulation and digital twin technologies, right, in AI and robotics, what future developments or trends do you anticipate in the space?", "speaker": "Speaker 2"}, {"startTime": 1801.359, "endTime": 1803.39, "body": "Yeah, I imagine, you know, the robots.", "speaker": "Péter Fankhauser"}, {"startTime": 1804.8, "endTime": 1810.339, "body": "Will, you know, have seen thousands of facilities over time, so they become also really experts.", "speaker": "Péter Fankhauser"}, {"startTime": 1810.76, "endTime": 1815.839, "body": "Now, in simulation, you'll be then able to have the robot generate emissions by themselves.", "speaker": "Péter Fankhauser"}, {"startTime": 1816.04, "endTime": 1818.479, "body": "Today it's still a human in the loop telling the robot what to do.", "speaker": "Péter Fankhauser"}, {"startTime": 1819.03, "endTime": 1823.75, "body": "Over time, I expect robots to be that intelligent using the data from simulation.", "speaker": "Péter Fankhauser"}, {"startTime": 1823.829, "endTime": 1829.05, "body": "So essentially you open the box, you turn them on, and the robot walks around and tells you, these are the 200 things I found.", "speaker": "Péter Fankhauser"}, {"startTime": 1829.63, "endTime": 1831.27, "body": "Is that correct, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1831.31, "endTime": 1842.829, "body": "And you say, yes, you adapt, you work with the human on its creativity and decision making, connectual understanding, but the robot will start to generate its own mission based on the knowledge and the information you can get out of the digital twin.", "speaker": "Péter Fankhauser"}, {"startTime": 1844.079, "endTime": 1854.239, "body": "This, this notion of autonomy, right, seems to be, you seem to, you, you found your key, let's say key vertical or your key industries, but it seems to be very adaptable, right, to, to others.", "speaker": "Speaker 2"}, {"startTime": 1854.319, "endTime": 1865.719, "body": "Are you, are you keen on taking over the world, other industries, or do you think in, in the other world around actually words around, I think other industries can benefit from your capabilities that you develop, right?", "speaker": "Speaker 2"}, {"startTime": 1865.989, "endTime": 1868.06, "body": "Do you have any, any thoughts on that?", "speaker": "Speaker 2"}, {"startTime": 1868.52, "endTime": 1868.599, "body": "Absolutely.", "speaker": "Péter Fankhauser"}, {"startTime": 1869.199, "endTime": 1872.729, "body": "I mean, I think fundamentally what we're building is extremely versatile.", "speaker": "Péter Fankhauser"}, {"startTime": 1873.109, "endTime": 1879.209, "body": "And we need to be extremely focused as a small company, so this is where we honed in on those core verticals.", "speaker": "Péter Fankhauser"}, {"startTime": 1879.52, "endTime": 1886.66, "body": "Long term, I see tremendous potential in the construction industry, measuring the progress, comparing that against the digital twin.", "speaker": "Péter Fankhauser"}, {"startTime": 1886.75, "endTime": 1891.81, "body": "Are we actually building, you know, the right thing here and detecting deviations super early?", "speaker": "Péter Fankhauser"}, {"startTime": 1892.39, "endTime": 1894.15, "body": "I'm excited about agriculture, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1894.189, "endTime": 1898.53, "body": "There's so much manual work being done, so that's a huge topic.", "speaker": "Péter Fankhauser"}, {"startTime": 1899.02, "endTime": 1900.099, "body": "Cleaning, you know.", "speaker": "Péter Fankhauser"}, {"startTime": 1900.63, "endTime": 1904.739, "body": "Um, switch systems, these kind of utilities extremely interesting.", "speaker": "Péter Fankhauser"}, {"startTime": 1904.839, "endTime": 1913.939, "body": "So I'll I'll believe over time we'll see more and more robots enter these seemingly unroboticized tasks and take over more and more of the work.", "speaker": "Péter Fankhauser"}, {"startTime": 1915.859, "endTime": 1918.03, "body": "So exciting times ahead for sure.", "speaker": "Speaker 3"}, {"startTime": 1918.38, "endTime": 1936.119, "body": "Um, what, what I'm also interested to learn is, um, what's your perspective on the growing emphasis of deploying robots and human facing roles, because you also mentioned earlier that humans and, and the robots are really meant to coexist in the same, same places and spaces.", "speaker": "Speaker 3"}, {"startTime": 1937.349, "endTime": 1941.839, "body": "This is a very important topic that we, you know, super actively work on.", "speaker": "Péter Fankhauser"}, {"startTime": 1943.469, "endTime": 1948.239, "body": "Fundamentally what needs to happen for robots to be accepted, they need to be trustworthy.", "speaker": "Péter Fankhauser"}, {"startTime": 1948.939, "endTime": 1954.579, "body": "Very similar, that's the ground, you know, the ground rules of also human interaction, it's the trust for high performance teams.", "speaker": "Péter Fankhauser"}, {"startTime": 1955.02, "endTime": 1959.599, "body": "And so for us what this means in a robotic system is number one, the robot needs to be transparent.", "speaker": "Péter Fankhauser"}, {"startTime": 1959.699, "endTime": 1961.859, "body": "You need to understand what it's doing, why it's doing.", "speaker": "Péter Fankhauser"}, {"startTime": 1961.939, "endTime": 1963.54, "body": "If it's not working, why is it not working?", "speaker": "Péter Fankhauser"}, {"startTime": 1963.609, "endTime": 1964.54, "body": "You can help it, right?", "speaker": "Péter Fankhauser"}, {"startTime": 1965.38, "endTime": 1966.92, "body": "Number 2, you, you need to.", "speaker": "Péter Fankhauser"}, {"startTime": 1967.88, "endTime": 1972.699, "body": "Be still in control and have the sense that the human, if it doesn't work, you can turn it off.", "speaker": "Péter Fankhauser"}, {"startTime": 1973.199, "endTime": 1978.119, "body": "There's an off button, you can shut it down, so it will not compete with you in terms of DD intelligence.", "speaker": "Péter Fankhauser"}, {"startTime": 1978.199, "endTime": 1982.439, "body": "So being very clear from the beginning how to interact with a robotic system.", "speaker": "Péter Fankhauser"}, {"startTime": 1984.239, "endTime": 2000.0, "body": "And then it's about the social connection, you know, for us this means the robot has a face, it has an eye, it has a little, you know, smile on the face, being an approachable and not creepy, but rather an approachable, nice coworker that you appreciate.", "speaker": "Péter Fankhauser"}, {"startTime": 2000.239, "endTime": 2003.42, "body": "Often in the process of deploying robots or customers give you the name.", "speaker": "Péter Fankhauser"}, {"startTime": 2003.56, "endTime": 2004.839, "body": "It's been very important, right?", "speaker": "Péter Fankhauser"}, {"startTime": 2005.29, "endTime": 2010.41, "body": "They run around and then they have a little competition and say, OK, how do we name this robot, Rosie or whatnot.", "speaker": "Péter Fankhauser"}, {"startTime": 2010.489, "endTime": 2012.53, "body": "They sometimes build a little doghouse for it, right?", "speaker": "Péter Fankhauser"}, {"startTime": 2012.569, "endTime": 2015.26, "body": "That's all part of welcoming your new employee.", "speaker": "Péter Fankhauser"}, {"startTime": 2015.329, "endTime": 2016.609, "body": "That's super, super important.", "speaker": "Péter Fankhauser"}, {"startTime": 2016.729, "endTime": 2032.3, "body": "So often we see a, a very quick transformation from Initial skepticism to like taking selfies, putting it on social media because it's our new coworker, you know, Felix, whatever name they give it, and so that that is a fundamental moment you need to give people time.", "speaker": "Péter Fankhauser"}, {"startTime": 2032.329, "endTime": 2038.459, "body": "You need to take them on the journey and then it's truly, you know, the, yeah, a mix of human and robotic workforce.", "speaker": "Péter Fankhauser"}, {"startTime": 2039.599, "endTime": 2041.349, "body": "Yeah, and this way it also becomes more.", "speaker": "Speaker 3"}, {"startTime": 2042.229, "endTime": 2052.129, "body": "Approachable, I guess, like if humans are also part, like in the, in the process of like giving it a name and welcoming it and probably the adoption, it makes the adoption certainly easier.", "speaker": "Speaker 3"}, {"startTime": 2052.469, "endTime": 2059.81, "body": "You already mentioned uh what verticals you see in the future for, for robotics uh applications.", "speaker": "Speaker 3"}, {"startTime": 2060.03, "endTime": 2065.75, "body": "Um, but generally speaking, what do you see does the future hold for AI and robotics?", "speaker": "Speaker 3"}, {"startTime": 2065.83, "endTime": 2067.419, "body": "What are you most excited about?", "speaker": "Speaker 3"}, {"startTime": 2067.469, "endTime": 2070.57, "body": "And do you have a moonshot um in, in mind?", "speaker": "Speaker 3"}, {"startTime": 2071.399, "endTime": 2075.06, "body": "Yes, so today, we're all about inspection, collecting data.", "speaker": "Péter Fankhauser"}, {"startTime": 2075.479, "endTime": 2082.3, "body": "That's a very important um routine tasks and so everything we do is, is, is based around that.", "speaker": "Péter Fankhauser"}, {"startTime": 2082.719, "endTime": 2084.12, "body": "However, robotics can do much more.", "speaker": "Péter Fankhauser"}, {"startTime": 2084.199, "endTime": 2088.679, "body": "Instead of just perceiving the environment, we can start to interact with the environment, right?", "speaker": "Péter Fankhauser"}, {"startTime": 2089.0, "endTime": 2093.12, "body": "So manipulation is also at the core of our next development work.", "speaker": "Péter Fankhauser"}, {"startTime": 2093.469, "endTime": 2103.56, "body": "Imagine, you know, an arm or multiple arms on top of a robot to do simple, simple things such as pushing a button, turning a lever, taking a sample out of a chemical tank, right?", "speaker": "Péter Fankhauser"}, {"startTime": 2103.87, "endTime": 2105.979, "body": "These kind of tasks are everywhere.", "speaker": "Péter Fankhauser"}, {"startTime": 2106.03, "endTime": 2108.909, "body": "There's still, you know, a tremendous amount of human labor involved in those.", "speaker": "Péter Fankhauser"}, {"startTime": 2109.155, "endTime": 2110.965, "body": "So, that's the next frontier still.", "speaker": "Péter Fankhauser"}, {"startTime": 2111.155, "endTime": 2122.665, "body": "We can show it in the lab, you can find it online, but it's still very difficult to deploy that at scale reliably and safely in the industry, but certainly robotics and AI has the capability and we have the capacity to do so.", "speaker": "Péter Fankhauser"}, {"startTime": 2122.715, "endTime": 2133.754, "body": "So that's for us, a, a key next step um as we build a robots and it's closing really the physic the loop between the digital and the physical world through robots.", "speaker": "Péter Fankhauser"}, {"startTime": 2134.84, "endTime": 2140.959, "body": "And that's what we also love to do at Siemens to connect the digital and physical world.", "speaker": "Speaker 3"}, {"startTime": 2141.3, "endTime": 2152.08, "body": "So Peter, thanks so much uh for your time, for sharing uh what you're passionate about, for, for giving us um some, some insights into what antibiotics is, what you're doing.", "speaker": "Speaker 3"}, {"startTime": 2152.669, "endTime": 2162.36, "body": "And what your, your mission is, and we're almost at the end of this session, but uh before we, we let you go, we would love to play our signature game with you.", "speaker": "Speaker 3"}, {"startTime": 2162.6, "endTime": 2169.31, "body": "It's called Authentic Autocomplete, and what I can reveal up front, it's, it's really, really great fun.", "speaker": "Speaker 3"}, {"startTime": 2169.679, "endTime": 2174.76, "body": "So for the closing, I'd like to give you a couple, couple of sentence starters, and you can finish them.", "speaker": "Speaker 3"}, {"startTime": 2175.5, "endTime": 2177.1, "body": "Does that sound, sound good?", "speaker": "Speaker 3"}, {"startTime": 2177.58, "endTime": 2178.219, "body": "Let's do it.", "speaker": "Speaker 3"}, {"startTime": 2178.699, "endTime": 2179.3, "body": "Let's do it.", "speaker": "Speaker 3"}, {"startTime": 2179.459, "endTime": 2180.679, "body": "So Siemens is.", "speaker": "Speaker 3"}, {"startTime": 2182.469, "endTime": 2187.399, "body": "But Siemens for us is a fantastic partner to bring a robotic fleet, you know, to market.", "speaker": "Péter Fankhauser"}, {"startTime": 2187.6, "endTime": 2195.739, "body": "I think we work together great in integrations and scaling and we leverage the domain expertise of Siemens, so Siemens is a great partner to antibiotics.", "speaker": "Péter Fankhauser"}, {"startTime": 2196.08, "endTime": 2196.62, "body": "Cool.", "speaker": "Speaker 2"}, {"startTime": 2198.83, "endTime": 2202.79, "body": "Robots can surprise you by performing the task of", "speaker": "Speaker 3"}, {"startTime": 2206.129, "endTime": 2210.729, "body": "Which, you know, as I mentioned in the beginning, robots just think differently.", "speaker": "Péter Fankhauser"}, {"startTime": 2211.05, "endTime": 2222.53, "body": "Robots can surprise you by being extremely accurate, smart, and they can show through AI kind of very intuitive behavior that you wouldn't expect, but it's still a robotic systems.", "speaker": "Péter Fankhauser"}, {"startTime": 2222.889, "endTime": 2224.61, "body": "There's a lot of things that robots cannot do.", "speaker": "Péter Fankhauser"}, {"startTime": 2224.77, "endTime": 2230.27, "body": "For example, you know, nobody has shown yet that you can take a robot to take the key out of your pocket.", "speaker": "Péter Fankhauser"}, {"startTime": 2230.53, "endTime": 2236.57, "body": "This is one of the unsolved robotics tasks, and I think it's being, you know, telling the stories of what robots can do and cannot do.", "speaker": "Péter Fankhauser"}, {"startTime": 2236.649, "endTime": 2241.01, "body": "I think that's fundamental to create that category and be transparent around it.", "speaker": "Péter Fankhauser"}, {"startTime": 2242.09, "endTime": 2244.479, "body": "Definitely, and autonomy is.", "speaker": "Speaker 3"}, {"startTime": 2245.86, "endTime": 2248.129, "body": "Autonomy is at the core of our products.", "speaker": "Péter Fankhauser"}, {"startTime": 2248.459, "endTime": 2255.639, "body": "We look at robotics really freed up when you have one robot being able to perform the tasks by itself.", "speaker": "Péter Fankhauser"}, {"startTime": 2255.899, "endTime": 2257.979, "body": "So one supervisor can look at hundreds of robots.", "speaker": "Péter Fankhauser"}, {"startTime": 2258.06, "endTime": 2266.62, "body": "So for me, robotics is not just a remote control car that still the human is in the room, but autonomy is at the core of building an autonomous workforce of robots.", "speaker": "Péter Fankhauser"}, {"startTime": 2268.55, "endTime": 2269.149, "body": "Agreed.", "speaker": "Speaker 3"}, {"startTime": 2269.199, "endTime": 2276.879, "body": "And last but not least, and this is a tough one, if I could invent one rule for everyone in the world to follow, it would be.", "speaker": "Speaker 3"}, {"startTime": 2279.449, "endTime": 2286.879, "body": "If I could invent one rule for everyone to follow, it'd be to be really responsible about the technology you're building.", "speaker": "Péter Fankhauser"}, {"startTime": 2287.409, "endTime": 2292.07, "body": "Obviously the technology we're building, you know, there's always a potential for misuse.", "speaker": "Péter Fankhauser"}, {"startTime": 2292.29, "endTime": 2297.739, "body": "We see this a lot on TV and media with drones being used in warfare.", "speaker": "Péter Fankhauser"}, {"startTime": 2297.979, "endTime": 2303.709, "body": "For us, we take, you know, a big stance on non-weaponization of our systems.", "speaker": "Péter Fankhauser"}, {"startTime": 2303.719, "endTime": 2304.689, "body": "We track the robots.", "speaker": "Péter Fankhauser"}, {"startTime": 2304.739, "endTime": 2305.969, "body": "We know our customers really well.", "speaker": "Péter Fankhauser"}, {"startTime": 2306.199, "endTime": 2314.479, "body": "And this comes, I think, also from, you know, the responsibility that we know we carry also Carry at our company, and I wish that others would follow that rule often as well.", "speaker": "Péter Fankhauser"}, {"startTime": 2316.06, "endTime": 2342.75, "body": "Peter, thanks so much for your time, for being that open, being that clear, I must admit, I'm deeply impressed about, you know, how clear you articulate, you know, the messages and being such a passionate mind which we can see and, and feel throughout the ether of the digital ether, uh, looking forward um to see, you know, the progress and the collaboration and maybe also meeting somewhere, either one of your kids, robots, let's say, or you in person.", "speaker": "Speaker 2"}, {"startTime": 2342.85, "endTime": 2345.159, "body": "Thanks so much for spending time on the Human AI podcast.", "speaker": "Speaker 2"}, {"startTime": 2345.949, "endTime": 2346.469, "body": "Thanks so much.", "speaker": "Péter Fankhauser"}, {"startTime": 2346.55, "endTime": 2347.429, "body": "This was a pleasure.", "speaker": "Péter Fankhauser"}, {"startTime": 2347.59, "endTime": 2348.699, "body": "Thanks, Albert and Joli.", "speaker": "Péter Fankhauser"}, {"startTime": 2348.879, "endTime": 2349.399, "body": "Talk soon.", "speaker": "Péter Fankhauser"}, {"startTime": 2349.939, "endTime": 2356.32, "body": "Folks, geeks and robotic enthusiasts out there, connect with Peter, antibiotics, right, or with us on LinkedIn.", "speaker": "Speaker 2"}, {"startTime": 2356.469, "endTime": 2362.229, "body": "Tell what you like, what we can prove or what kind of use cases you foresee for antibiotics and obviously, stay tuned.", "speaker": "Speaker 2"}, {"startTime": 2362.31, "endTime": 2363.8, "body": "There's so much to come, right?", "speaker": "Speaker 2"}, {"startTime": 2363.939, "endTime": 2368.55, "body": "Stay bold, committed and open-minded and will hear us at the next Siemens A I Lab podcast.", "speaker": "Speaker 2"}, {"startTime": 2368.709, "endTime": 2369.149, "body": "Cheers.", "speaker": "Speaker 2"}], "language": "en-US"}