Episode 110

The Future of Robotics

Watch on YouTube

About this episode

Welcome to the final installment of our three-part robotics series on the Untangling Web3 podcast.

In this episode, we explore how robots and automation are transforming real-world industries—from agriculture and manufacturing to healthcare, humanoid design, and beyond.

As the lines blur between robotics, artificial intelligence (AI), and human labor, we examine how this revolution is already unfolding and what the future might hold.

Key highlights

  • Sector-by-Sector Applications of Robotics. we consider how robots are being used today across key sectors. In manufacturing, robots dominate production lines with precision and 24/7 uptime. In logistics, massive fleets of warehouse bots (like those used by Amazon) are reshaping how goods move from factory to front door. In agriculture, autonomous tractors, drones, and AI-powered sensors are helping boost crop yields while solving labor shortages. And in healthcare, surgical robots, exoskeletons, and even companion bots are supplementing human doctors and care providers.
  • The Rise of Humanoids and AI-Native Robots. One of the most forward-looking discussions explores the shift toward humanoid robots—bipedal machines designed to fit seamlessly into human-built environments. These machines offer a general-purpose solution for industries facing labor shortages, especially in aging societies. The hosts also explore the integration of AI and large language models (LLMs), allowing robots to understand natural language, learn from users, and personalize their tasks. This shift marks a move from rule-based behavior to AI-native agency.
  • Soft Robotics, Swarms, and Ethical Futures. Emerging trends like soft robotics, which mimic human tissue and enable safer interactions, and swarm robotics, where many small bots collaborate to complete complex tasks like environmental clean-up. But with this explosion of capability comes enormous ethical questions. From job displacement to regulation, digital rights for robots, and geopolitical control over robotic infrastructure, the discussion highlights how society must prepare for a radically automated future.

The future of robotics is here—and it’s far more nuanced, powerful, and intertwined with AI than most people realize. Whether it’s building homes, harvesting food, or assisting in surgery, robots are moving from the lab to the real world at scale.

As we navigate this new age of automation, the way we design, regulate, and relate to these machines will define our shared future.

This episode is sponsored by the VeChain foundation. Learn more about VeBetterDAO here: https://vebetterdao.org/

Transcript

Auto-generated from the episode audio.

Welcome to the Untangling Web 3 podcast and also welcome to part three, our first three-part series on robotics. In part one, we explored the origins of robotics from ancient myths and early machines to the rise of science fiction and the first industrial robots. Then in part two, we broke down what makes a robot tick from the sensors, controllers to software and power systems. And now in part three, we're going to be diving into where robots are being used today across industries like healthcare, agriculture, logistics, and many more.

And what the future holds for robotics as it starts to merge with AI, autonomy, and even the ethical and political challenges that are already starting to arise. Jack, how you feeling about this first part three series we've ever done? Yeah, I uh I've really enjoyed it so far. It's been interesting kind of being the uh the sounding board for you as the expert in the room. It's quite it's quite fun role that I'm enjoying playing. So yeah, um I think today is going to be no different.

I think you're going to be, you know, holding my hand through the the world of where robots are actually being used in the future because it's it's definitely not my wheelhouse. But yeah, I've learned a lot so far and really enjoyed it. So yeah, thank you. It's been good fun. Yeah, I'm glad. I hope that there's some listeners who have also enjoyed it and it's not just you and me. Um but I think one of the things about robotics is it's so big. like it's kind of similar to AI and there's so many things that you can talk about but rather than just AI generic robotics is quite applied and each of them unlike AI as well it's quite tangible in how to understand things because they're physical because they're being used because you can see them and understand the impact like quite visually and quite like real in a practical sense I think people it makes a bit more interesting to talk about makes it a bit more interesting to understand but yeah rather

than just talking about what we're going to talk about let's actually talk about it um so yeah like I said we've talked a are about kind of the fundamentals. Um, but you know, they're they aren't just something we build. You know, they're tools that shape how we live and they're being used across every sector. We're seeing them being used in every niche. Um, and I think it's really important to maybe go through some of the key areas and sectors they're being used in right now and really like kind of basically work out how they're being used, what they're being used for, and then maybe we can start to understand what they could be used for in the future.

So I think like the first and probably the most important one to talk about is manufacturing. Like we kind of said in part two and part one, this is really where everything started for robotics. Like when you think about those early giant robotic arms. So like like I said, manufacturing this is really the origin story of robotics and and and the OG of robotics and being applied. Um it still dominates today. like in 2024 2025 there are over 4.3 million industrial robots around the world and most of them in factory lines.

I think like the really important thing to understand about why they're being used here and still dominate today is because they're very controlled environments and very repetitive tasks. So where you want speed, precision, 247 uptime and you also have an environment that isn't super dynamic. So you imagine these arms basically fixed in place. They get given the same output every time. They produce the same output. So get given the same input every time, produce the same output every time. And you want that as quickly and as precisely constantly all day.

That's a manufacturing line, right? That's how you make a car. I'm sure everyone's seen the videos of how car gets cars get built with these arms. And yeah, that's why they're so popular at manufacturing. Yeah, it makes sense, right? when we spoke about the fundamentals of what robotics is and and the various kind of tradeoffs versus a human being, right? I guess ro um I guess manufacturing is maybe the most obvious place where robotics has the the huge advantage over human beings, right? They they literally don't have the same physical limits and when it's to do with physical hard labor or as you said the uptime even like so you can do 24/7 manufacturing.

Yeah, it makes perfect sense that this and again this is like as you said kind of the OG, right, of how robotics have been used. So it's it's been there the longest time. It's probably I I would guess the most mature in this sense because it's been you know I imagine there's been a hyper specialization iteration of these robots over time. Yeah, we spoke about the you know our school trips to Jaguar Landro factory, right? that was, you know, 15 years ago. And so, so you can see how long it's been installed in in these processes already.

And I'm sure it's kind of got only more kind of more sophisticated from there. So, yeah, this this one makes perfect sense to me. I think again, as you said in the intro, it's very tangible. You could everyone I think now growing up would associate uh manufacturing with robotics. It's not human processes anymore on the factory line. It's primarily uh yeah, robot driven, which makes sense. Exactly. I think like and you know what they do isn't just like grip move that kind of stuff like you can imagine them you know if you put a welder on there it can weld parts together it can paint they can obviously assemble as well create packages on production lines I think like some of the interesting trends we're starting to see that is more around the idea the term is like cobots so like collaborative robots so rather than just have these very defined environments where it gets given something it does something it puts it out to the next robot them working in collaboratively with humans in

the loop as Well, so like I can pass up a a piece of, you know, I don't know, a piece of metal or a piece of cardboard or something. It takes it off me. Again, like starting to understand that there is sometimes a need for a dynamic nature. So you can imagine them working in more flexible environments. Um, that's one aspect. Another aspect is basically like I think maybe 10 20 years ago, they're only used in the really big heavy spec environments like the Jaguar Land Rover you said there.

But more and more as they get more affordable and they start to become more intelligent, you can put them in say smaller, more customizable environments as well like smaller manufacturing plants where maybe they do have this degree of understanding what a task is dynamically and being able to act in a dynamic environment. And then yeah, the other thing is like they're starting to auto check. So where is previously it would be you know they produce a million parts a human comes in the loop and we'll check that millionth part to see if it's okay and then they say yes no now we're starting to see that the robots are you know with AI enhanced vision systems and whatnot are starting to be able to selfch check and selfcorrect and be like oh I made a mistake on this or robot 3 made a mistake on this what went wrong there and they're starting to self-learn and self-update which is yeah it's quite it's quite cool to be fair yeah cool a little bit scary like

that's there's a lot to unpack there I think yeah the point you make about um the advancements being in you know not being quite as specific task-wise as they used to be and being a bit more general being able to do the QA itself as well. So like quality assurance where you I guess that would be a classic place where you would have a human in the loop. Now with the advancement of sensor systems you have again back to our previous episode you have more inputs that you can you know give to these robots that they can you know produce more complex outputs like the the self QHX.

Yeah that's really interesting. I think I had hadn't kind of thought of that. I actually was also, as you were speaking, thinking about, you know, what what uh people's perception of manufacturing would be. And I think there was that show on one of the maybe the Discovery Channel about how how it's made, I think it was called, when you'd see all the robots working together. And I think now I'm just imagining if you made that in in 2025, you'll probably see them, you know, doing them doing these things much more efficiently in kind of Yeah.

a much more modern way. But still at the core of it, it's it's essentially robot controlled systems. Right. Exactly. And I think like the big thing there is it makes sense to do that kind of thing. If you say a Jaguar Land Rover says, I'm going to build 400 million of the same car and it'll take me a long time to set all these robots up, but once they're set up, they go. But more and more we're seeing it making them them being more flexible, more configurable, easier to deploy and then they can work on smaller and more custom scales.

So the next one is not too big a leap from like manufacturing. It's the idea of like logistics and warehouses. This is kind of like the new epicenter of robotic growth like you know we basically over the last 20 years seen an absolute e-commerce boom like and with e-commerce boom comes like the idea of like just in time arrival the need for giant warehouses that store you know lots of Amazon products so they can get to users all around suburbia. Um but also we're seeing lots of labor shortages and people don't want to have these jobs necessarily.

They can't fill the you know the growing and you know massive demand that is coming from e-commerce. So like logistics has become one of the fastest if not the fastest growing sector for robotics by like number of devices. So what do they do there? Slightly different to manufacturing. They're still quite controlled environments but it's much more on like the mobile robots navigating through warehouses effectively moving in inventory. I imagine most people have seen like if you haven't I would recommend it seen like the videos of how Amazon robotic warehouses work right now.

kind of have these giant rumbers that will just pick up cardboard boxes and move them from place to place and put them in the delivery zone so that it can be sent directly on a truck. And they have like, you know, automated pickers that grab goods and shells with suction cups. There's loads of actually like really nice little neat tricks that we don't really think of in terms of robotics like soft grippers and vision guided arms and all this kind of stuff as well as like the typical, you know, smart fork lifts and all this.

But yeah, there's like I think Amazon use right now over 750,000 robots across the logistics network. So it it is huge. Yeah, I think it's I think Amazon's a great example here because that's one where I think with the rise of, you know, next day delivery, one-click purchasing, and the the just the speed that people expect now and the efficiency of supply chains for consumer purchases, I think there was there's been a lot of controversy and up upheaval around the use of human labor in these kind of markets and and the production logistics for these things.

So, it's kind of an obvious place where you think, well, that that would be great to resolve that issue by replacing a lot of the the more kind of dodgy labor there with robotics, right? And obviously that you can argue the the human cost and the the jobs replacement, you know, what happens there like we did previously, but it makes sense purely from like a safety and efficiency perspective. Um, and also when we're talking about logistics, I guess does this also stretch all the way to kind of the end of the delivery chain as well, right?

I I remember reading about Amazon thinking about like delivery via drones. I think this maybe been trial and piloted in some places which is again a use of robotics to replace another part of the supply chain. So like with lots of supply chains now you see them being kind of vertically integrated right like Tesla are doing this Amazon are doing this trying to take control of the entire supply chain they need and that would maybe include the the use of robotics in integrating in each of these parts including delivery which is again slightly scary but also you know drone delivery kind of it makes a lot of sense right for especially kind of it's the use of an airspace right where where you currently don't have you know an underused resource essentially, right?

So, I can see how that would take off in the future, but I don't know what the current state of that kind of stuff is. It got legalized in the UK, not in cities, but I think in like some super suburban areas where it makes the most amount of sense, where like not near airports, not near Exactly. not near airports. But you're totally right. There is like obviously we always have to think about the human impact on these things. And the human impact in my mind is like twofold.

One is the loss of jobs, but again, these aren't necessarily jobs that people wanted. Like there's a lot of like vacancies at Amazon. And this was one of the reasons why they started to fill this with robotics. But again like these systems I think some of the stats are locus robotics they're one of the main providers of these they can improve picking speed by two times like double the double as the speed and they reduce the error rates by 30 to 40% in their like logistics process and all of that brings cost down for Amazon.

So like you could say argue oh they're making more profits or you could also argue maybe they're bringing costs down as well which again affects us positively but yeah the warehouse automation market is predicted to hit 40 billion by 2030. This is big money going into this. It's not going away. And all the reasons we just stated, speed, accuracy, adaptability, like 247 uptime, precise tracking. Again, like hive learning. One person makes a mistake, they all learn from it. But also being completely in sync with every single workaround.

If you've got 750,000 humans running around a warehouse, they're going to bump into each other. It's going to be absolute carnage. 750,000 robots. If they're all in completely in sync, they know exactly where they are. they know exactly where every other robot is and they know exactly what path each of those robots are taking. You can really start to sync up and you know move much faster and reduce collisions and all this kind of stuff. So yeah, there's a lot of optimization going in these spaces and I think it makes a lot of sense.

Yeah, for sure. I I think it you're also seeing the kind of increase in complexity as well now which is a marker of as we said the journey of robotics is now you're talking about kind of hives right of many many of these robotic systems working together. We've talked about a kind of more abstract layer version of that with AI agents and them all talking to each other online and paying each other and doing everything for us. But, you know, we have a much more practical tangible version of that already in these in these warehouses, in these logistics and supply chains where yeah, you do have robots interacting with each other.

I like the idea of them not bumping into each other instead of humans. I mean, that's a that's a clearly a pro of them, right? So they don't you you don't have the risk of the the human to human error uh you know posing a safety risk and things like that, right? If two robots bump into each other, no one's going to get too worried. Yeah. You've se I've seen like the the like kind of the theorized and simulated highways of the future where it's all autonomous cars.

You don't need traffic lights or islands or anything like this. Everything just knows where everything is and where it's going if everything just perfectly merges. Yeah, I guess this is what's already happening like the Amazon warehouses. So number three, I think maybe one of the more misunderstood ones or people don't really know what's going on is agriculture. When you think of agriculture, you kind of think typically I imagine like a farmer with a horse like plowing the land and all this kind of stuff. But from horsepower to robot power, even agriculture is really starting to transition massively.

Again, similar to warehouses, like there's, you know, food is important and the growing population needs more and more food. We have more and more demands and like but there's two things that kind of impact this. One is labor shortages, something we're massively seeing in the UK right now. The other is climate stress as well which probably don't typically think of. So we need to kind of do these things in a more efficient and effective way and robots are really stepping in to that niche not just to automate like basic processes but also to optimize those processes as well.

So you can imagine I've seen actually some of these like autonomous tractors basically like a Tesla tractor style that PL sews and fertilizes the land but also back to you know one of my favorite topics and you just mentioned it earlier drones as well can like scan and review crop health you know imagery compare it daytoday do like a time lapse see what's going on but with like multisspectral imaging like UV imaging as well not just like vis visual based um and then also they can start to you know actually deploy um pesticides herbicides whatever you want to call them at a much faster rate 24/7 again be much more reactive than a human can because they've got more information and basically be more impactful and effective across like agriculture as a whole.

Yeah. And I think this is this is a really interesting one because this is where you have a you have a scale a scale challenge right because you know farms especially in the UK I guess you know it's a hot topic at the minute but uh kind of family run farms you don't necessarily have that many people working on them and you have to scale what they're actually doing to these very very large plots of land. So I guess I'm imagining now I hadn't really thought about agriculture before in this in this light but I'm now imagining it's basically the Roomba use case again right it's you have a defined plot of land you need to go around and do something instead of cleaning up and hoovering this time it's you know cutting down crops and and storing them somewhere it makes sense like there really does if given the capabilities of robots now it would make very little sense in the future I guess for there to be a person sat on that tractor

right while while it's doing while all the combine harvest while it's doing its job. It just doesn't make any sense any way when you have that that capability. And yeah, the the the drone thing is really really interesting because again it it clearly tackles that that problem of human scaling up of this challenge and then also you can get more information that you couldn't before from the sky, right? It is a different kind of way of uh inspecting the land as a human being. You have a local view when you're on the ground.

you can see kind of, you know, what's right in front of your eyes and a small patch, but when you take it to drones and things, yeah, you can have them automatically scanning everything, working out where you need to when you need to focus on. So, yeah, really interesting use case. I think like it kind of relates to the AI as like AI topic that we obviously constantly talking about like one of the main benefits is the sheer amount of information like for one person to run a farm and just be random, not randomly, but collecting data in like a a normal manner and be like, "Oh, well, you know, it didn't work so well this year.

what did I do differently? It didn't work so well that year. Like what am I doing differently between these crops? But for robots and drones taking all this information and storing it all and comparing it all, you know, things that we couldn't maybe understand like get in say different degrees of information, not just visual input as well, and starting to optimize, say, okay, how do I optimize the problem? Why is it that this crop over here performed very well, this one didn't? Can I do something different and start to work out what patterns create optimal yield for for crop production?

And like that's one thing like the sheer amount of information all of them plugged into one grid collecting information across farms across countries all this kind of stuff and trying to optimize AI does that very very well. We've seen like some of the systems that are out there right now I think there's one called um CN spray or something like that they reduce pesticide usage by 90% with these like smart tailored pesticide drones. Um, another robot can I think it's like a small robotic drone as well.

Quite sorry, not small, quite a big one that can eliminate over a 100,000 weeds per hour using AI and laser techch. And again, this market, agriculture, robotics is predicted to be 40 billion by 2032. So all of these, you know, in all these sectors, robotics specific sectors are booming right now. And yeah, feeding 9 billion people needs more than just, you know, manpower. We need robot power, I guess. And that's kind of where the sector is going. Yeah. And it makes sense like I guess somewhere especially like in the UK where people talk about food security all the time.

This is a great all the time. All anyone talks about when you go to the pub is food security. Depends where you are in in the world. But yeah, I think um it's a great example of how technology has a really practical solid application to solving that as a big, you know, ongoing challenge and and yeah, I think it makes perfect sense. Um the other one I wanted to talk about because I I think again I don't have much visibility or knowledge for for this but I imagine healthcare is also one of these areas where it's being used right so I mentioned in the last episode about Amazon I don't know if it ever actually got off the ground but there was this pill packing robotic system right for how they they had like an Amazon healthcare thing and it was about how you you know sort medicine into the right buckets and things and get it you know packaged but I imagine robotics is is being used in healthcare, but I also I'm struggling

in my mind to visualize, you know, is that something that can replace surgeries and things or not? Or is it just kind of in the kind of the day-to-day running of a hospital or something? So, yeah, I don't know if you know anything about this that you can enlighten me on. Yeah, I think so. Right now, it's quite supplementary. I think like all the use cases we've talked about so far are very much hidden from public view as in like Amazon and not showing the robots, you know, on the front page.

They're hidden and optimizing in the back. Whereas I guess surgical robots are like it's much more in your face and if you know most patients if you ask them do you want a robot to perform the surgery or this consultant over here they're going to choose the consultant. So I think when we went through the overview of like the history of robotics we talked about Da Vinci like one of these surgical robots and the idea there was basically just to reduce instability. So like I guess like one of the things you can think about is when a human is holding something, it takes strain and effort to hold something in place.

Whereas for a robot, it's basically doesn't exist. Like it can hold that position perfectly well. So imagine if you're operating on someone, you're in there with a scalpel like in their heart or something like that. Like if you're flinching or vibrating or you need to hold that position while you insert a tube or something like that, that's a lot of effort for a surgeon, but for a robot, not so much. So I think right now a lot of these like Da Vinci is probably the biggest company.

They sell these like million pound surgical robots. Um, but they're all supplementary. They don't do anything autonomously necessarily. It's mostly just imagine me. A lot of them have like glove kind of style manipulation. All your fingers are on like little dials in there. It's kind of like a game in a lot of ways and you have the visual cues and everything and you're just using that to manipulate inside the person, but it's still guided by the surgeon. But just gives you like levels of precision. Imagine like I can move I can control the movement of my finger by millimeters.

If I scale that up to like a massive degree where I can control, you know, one millimeter of my finger equals one one millionth of a millimeter on the roadike gives me so much more precision to actually do things. Today's episode is sponsored by the Vchain Foundation. Now, we often talk about how emerging technologies like blockchain have the potential to revolutionize sustainability and Vchain is at the forefront of this green movement with initiatives like Vet Dao. Let's take a closer look. Imagine tracking small actions like recycling at low cost, making them valuable assets.

These verified records build a massive collective impact benefiting people through their own actions. Sound a bit complex, but Vchain made it fun. Explore our apps at vb better.com. Vchain's goal is to power a new era of sustainable solutions. So, if you've got an app idea or just want to earn rewards for small sustainable actions every day, go check out vetero.com to find out more. And now back to the show. Yeah. So this is really interesting actually because I don't think we've maybe touched enough on this kind of class category of of robot in robotics in the previous episodes.

But this is almost like the exoskeleton model for robotics, right? Where it's instead of trying to replace an anthropomorphize uh machines in in a way that completely replaces a human's actions. This is like where they are an extension of a real human behind the scenes, right? So, you know, maybe it's like uh on on the large scale, if you want to visualize this, it's like your Pacific Rim uh you know, robots or your Power Rangers Megazord or something like that. Imagine you're about to have a surgery and someone comes in like a Pacific Rim style exoskeleton.

I'd probably, you know, still be more trusting of that than me to do a surgery. But yeah, this is where you still so because that's what was interesting is like we talked about the tension between the capabilities of robot robots being higher than humans in lots of cases, but there's there's this kind of trust element and and the risk involved that you have a high risk situation like a surgery. People don't necessarily trust a robot to to autonomously do that. So you're removing the autonomy part and you're saying no, we're going to try and mimic the physical uh the physical improvements you get from robotics but still have that human not in the loop but literally driving.

So replacing the brain part I guess that we were talking about before. Yeah, exactly. So that's one of the big things is exoskeletons is a good example of this stuff like I'd say Da Vinci kind of is a version of that with like this this surgical control. I think they've done like over 12 million surgeries globally right now as well which is like so this stuff is booming. and they've been around for like 20 years. But I think a lot of it is also being able to do surgeries remotely and things like that.

Like have someone you know teleoperated one of these Da Vinci robots from across the world or something like that. If say there's one surgeon who can do the brain operation then obviously that kind of stuff makes sense. Exoskeleton is interesting like there's a lot of exoskeleletal like applications to recovery. if someone's got an injury, just putting some like, you know, robotic arm around them to protect them, but also give them the ability to like close the hand and stuff like that rather than just the old school cast system.

Like we're also seeing a lot of like mental wellness and support like companionship robots. Maybe you've seen like videos on Instagram and Tik Tok of like cute plushy toys that have a degree of like response and can like actually communicate with um elderly people. Like all this stuff is super important. Like in the UK our NHS is like the the globe the healthare system sorry is like really under strain mostly because as in a lot of the western world and world at large we have a an elderly population a growing kind of old population and it's hard to keep up with the demands required of the strain like say social system but you know robotics offers a really good means to maybe automate some of the processes reduce the cost of some of the processes and like you know provide companionship and support for old people.

So I think this stuff is really important and like there's so many different applications you can kind of like imagine with this stuff. Yeah, I think that's it's really important that you highlight you know the actual the health care is not just surgery and and medicine. It is the the kind of psychological side as well which is where again I think people are starting to claim that LLMs will be a huge have a huge role to play in that um that domain as well. Right. So being able to simulate companionship to be able to to offer, you know, we had that whole episode talking about therapy.

I don't necessarily know if it's going to replace or help, but it it offers insight. I mean, I know I've Googled, sorry, I know I've used chat GPT for for symptoms and things now already, right? So that's a kind of just the how the wet wear of robotics can help. I mean, when you overlay that with um yeah, with with the mechanical side, it's very interesting. I hate that wet wet term so much. So the next sector I think we should just skip over it like domestic and consumer robots.

I think most people understand this. It's like rumbers but imagine like when rumbers start to get more like humanoid toys like kind of dog bots and all this kind of stuff that's going to help around the home. Like there's many things you can imagine being used here. One thing that I think people don't tend to think of is construction. I think of all the sectors we've mentioned like construction is typically associated with being the most archaic. Like construction hasn't changed too much in my opinion over the last 2,000 years.

It's still brick and mortar in many ways. Surely I mean we have in some ways like automated the construction of bricks that are then used but on-site construction hasn't changed that much and I think this is like ripe for disruption. So you can imagine and this actually one of the things that a lot of my research in was like construction robots and automating construction robots but one thing I imagine a lot of people have seen is the idea of these giant 3D printing robots. So 3D printer you have plastic you extrude around like a controlled environment and then you do layer on layer to make something.

But imagine that but building sized or skyscraper sized. You have this giant huge frame around a building, concrete, liquid in there, and then a head that goes around like this 3D axis and can print layers of a house. Like this kind of stuff is super science fiction right now, but there are a lot of prototypes out there that are making buildings. Not to scale, not that like are making like a lot of money right now, but in 1015 years, I really think we're going to see this as like a potential standard for construction.

Yeah. Yeah, I had absolutely no idea that robotics was being used in construction, but again, like everything else, it makes perfect sense. I mean, I know that in some of the bigger, more futuristic projects like maybe the neon the line project, there's a lot of talk about using robotics there. I'm not sure how much that will come to fruition, but yeah, even even when you mention your work on these, what is it? It's they're like uh for for fixing potholes, right? And and like 3D printing kind of uh material into potholes.

Yeah, that that makes perfect sense. Again, it's it's jobs that no one wants to do, but there's high demand to to to solve that problem. So, yeah, it's it's another interesting one. Yeah, like the actual actuation like filling potholes or, you know, having like a brick layer, like you can imagine just, you know, I get an arm, I put it into the center of where the house is going to be. I fix it there. I put like a thousand bricks around it and it just starts to place them and put concrete.

Like, it's kind of easy to imagine how this stuff would work. But this stuff also goes beyond that. Like a big thing is 3D scanning. Like a lot of construction right now is trying to how do we make this more digital? Like how do we actually create 3D representations of these constructions that we're making to ensure that you know basically do good maintenance, ensure they've been built to spec. So you can have like inspection drones that come in and 3D scan buildings, work out where parts are so they can start to you know automate some of the processes for maintenance.

But dog bots, you know, the robo dogs are also being used have a giant 3D scanner on the back. They go around and create 3D images of the entire setup to give more information to someone who's managed it in the background. But yeah, I think another big thing is prefab. So like a lot I said that a lot of construction hasn't changed. One thing that has changed is the idea of pre-fabrication rather than taking bricks and just putting it on site, building an entire wall in like a manufacturing setting, which we know is super efficient.

We know is incredibly quick and precise. taking that on site and then just assembling four or five different large bits to make the house and basically making that on-site component much faster by pushing it to the off-site component. So I think like construction compared with every other sector like it hasn't improved in terms of productivity really like maybe like 5 to 10% over the last 20 30 years whereas manufacturing is like 100 times like multiple folds of efficiency improvements and as always there's a massive global skilled labor shortage for construction workers.

So yeah, there's a it can impact this massively. Obviously, it's very dangerous as well. Like construction is a dangerous hazardous place still. So you want to offload this to robotics or at least parts of this to robotics. So I think like I say this this whole sector is completely ripe for robot construction disruption. Robot construction disruption of that. And I think especially when we start to see humanoid robots come into the fold where they can say lift the same amount of weights that a human can lift, they can work longer hours.

I think we're really going to see them being applied to this sector because it is so dangerous, because it is so laborious and because there's so much demand and so much profitability that could come from building houses. Yeah. And I think you you mentioned the the safety being such an obvious uh benefit here which I I completely agree with. There are things that you you definitely should be replacing uh human activity with just because it's too too hazardous. Um, you also have the kind of element of precision as well.

Like you only need to have watched an episode of Grand Designs to know that when you're when you're fully reliant on humans and even even when you're you know hyper specialist and experience at doing these things just by by you know the nature of human human beings and us not being completely accurate when we're measuring things or things moving. if you can replace the kind of measurement side and the precision side with robotics, I think that is also a completely valid reason to use them, right?

Just because you're going to reduce errors. Um, and even though you can have humans, you know, again, in the loop at multiple stages, those key moments, I think, are going to be definitely replaced by by machines in the future. Yeah. And the final sector I want to mention is defense and security. Like obviously the most controversial of all of them, but it's always like with robotics there is often an element of military in there, especially when we talk about human robots, like humanoid robots. A lot of the early humanoid robots funding came from DARPA and came from the US military.

And like yeah, you could take like the ethical approach of bomb disposal, you know, autonomous medical robots, all this kind of stuff. But really, a lot of these things when we see spots or we see humanoid robots, it's how do I put a gun on there and put them into the battlefield because I don't want to lose a soldier. This stuff is controversial. I don't want to spend too much time talking about it because like it's not like there's specific elements that we haven't discussed earlier.

It's basically how do I put a weapon on these things? Drones obviously one of the big ones is like how much funding is going into drones both for surveillance but also we've seen in like the Russia UK crane war how much drones has played a part in that because they're quite cheap to deploy because they're quite impactful because they're not risking you know the kind of whoever deployed them's life in in the act but yeah I maybe there's not much more to say on that one.

Yeah I think I think this is probably again the only thing I would mention is that this is where people are most keenly aware of of the use of robots. I think whether that's from in the real world what's happening now uh in you know I guess modern warfare but also from this is where the science fiction elements that we spoke about in the first episode come back right like this is the original application of robots to most people is them taking over the world and being weaponized in in some way right so Terminator put a gun in the hand of a humanoid robot and you get Terminator exactly so I think this is this is the one where again also like this is probably where I have the least practical knowledge but the most theoretical and sci-fi knowledge and I think most people will be the same.

So yeah, I think we don't need to spend too much time on this. But defense drives innovation. So that's like most of the sectors that I want to talk about in terms of like where robots are being used today. So now I think we've kind of got a good understanding of what robots are, where they're being used. We can talk about some of the predictions and like growing trends that are quite interesting. The first one is something we have touched on quite a few times is the idea of humanoid robots.

I think it's really important to understand that so much of the world is set up for humans. Like we've adapted our environments, urban society, buildings, cars, absolutely every aspect of our life is designed for a bipedal human with opposable thumbs, right? And when we're trying to we've built robots so far to basically be simple, simple designs, simple actuation, simple ability to understand their environment. But more and more, if we want them to fit into a general purpose robot that can do almost anything that we'd find useful, they need to mimic humans.

They need to be able to open doors like humans. They need to be able to drive cars like humans. When it's unlikely we're going to adapt adapt our society for robotics, it's more likely we'll adapt robotics for our society. This is going to be one of the biggest, I think, trends and growth sectors we're going to see over the next 20 years. Yeah. And this is I think um where lots of people are anticipating you know things to get weird basically is I think um you know when when when people start talking about humanoid robots I think people start talking about or thinking about uh you know laying skin on top of these skin what that is.

You know what I mean? Like so yeah not made of to be clar clarified actually real skin hopefully. Um but yeah this is where you're you're talking about trying to again go run down a full anthropomorphization route and you're making them overcome the uncanny valley of okay this is a robot. I mean you again this has been around for quite a while right? You've had these proof concept robots where they put this rubberized face on them and they now I guess now 3D printing very realistic um kind of humanoid features on top of the mechanical side of the robot um and then starting to integrate them in in basic tasks.

I know we had we spoke about the uh you think this has been very much proof of concept stuff for a long time. Tesla had their event where they were trying to show show off their humanoid robotics capabilities, not with the the rubberized skin suits. But yeah, I I don't know what to think about this one. I think when we're talking about robots, I can see the the benefits of like a lots of the industries we just mentioned without going down the the trying to simulate humanity route.

And now with the with the rise of LLMs and everything, now that you can simulate conversational things and voices so so easily, I think we people are going to start pushing harder for this humanoid robot thing, right? But I don't know if I'm I don't know if I'm genu generally up for the world where that's happening, but I don't know. How do you feel about it? Um I think like when we start to think like it's mimicking a humans like personality intelligence like looks I think that's where it gets like controversial and I understand like the reluctance to go down that route and I think a lot of the thing people that are working on that it's mostly to get a reaction right to kind of stimulate a response.

I think like when I think of like humanform robots, I mostly think of like something that has the exoskeleton of a human or has like the outline of a human. It has fingers, thumbs, it has bipedal just in terms of interacting with the environment so it can open doors, drive cars, do surgeries, all this kind of stuff. I think when we go down the route of like how do we mimic humanity and say start to replace humans relationship wise, conversationally, all this kind of stuff, that's when I find it a bit unnerving like what does that mean for society?

Yeah, I think purely in terms of like the physical aspects of creating a human robot that can do human physical tasks. I think there is so much application there like we're seeing a population decline no cliff dive in so many countries like in terms of like say Japan Korea it's happening like a lot in a lot of the western world and a lot of people are saying okay what do we do about this like we need young labor to say fill agriculture manufacturing to keep up productivity all this kind of stuff it's not feasible to necessarily keep producing people can humanoid robots start to fill this void and I think there's a massive application there.

It's really interesting to see that, you know, some of the countries that are the most advanced in terms of robotics generally, but human humanoid robots specifically, Japan and South Korea are also the countries that are experiencing this, you know, population decline because they don't have young people because they don't have a workforce of young physical people. They're the ones that want it the most and need it the most. And I think there's a massive application for that. Even if you think like just big dumb humanoid robots that could just do those laborious tasks, like I think a massive angle for that and a real need for that for society to be honest.

Yeah. And I I I completely understand that and I I I agree that like replacing the tasks themselves makes perfect sense. I just guess where I'm coming from is um is are we just inserting the god complex here and saying, well, these tasks that need to be done, these are best done by a human, therefore we fit robotics to match our image, right? Whereas, you know, m maybe robotics can be used to replace all these human functions without looking like a human, I guess, is what I'm getting at.

But yeah, I don't know the answer to that. It's just an interesting I get I get what you mean. I think like the problem is like with that I'm just talk thinking purely like economically now. It's like if you make a robot that looks can act like a human physically, they can do almost any human task which is almost all tasks around the world. If I choose a very specific use case like okay how do I do s maybe not surgery how do I do agriculture in such a way that I don't need humans in the loop anymore I have to revamp the entire agricultural sector sector right and completely change everything so that's already two like different areas that I have to be an expert with then if I want to go to say manufacturing completely overhaul manufacturing I then have to have an expert in manufacturing be like how can I overhaul this I think it's actually a lot of these companies are basically just roboticists that don't have any kind of like

practical you people are working on humanoid robots anyway don't have any practical use case in mind because they want to keep it as generic as possible and then just apply it everywhere. Yeah, that makes sense. And I think I guess now now we're talking about it more. I think this definitely suits itself to like a transition phase between human based labor and eventually fully robotized based kind of hard labor is that you want it to be uh anthropomorphic because you want them to be working alongside humans during this transition, right?

You want it you want to not change the way the system works, just who's working in the system until you know until right at the end. Then you might decide, well, trying to do this job in the humanoid way doesn't make sense at all. We replace the whole system at the end. But yeah, that makes that makes sense. Now cars, right? Like same with cars. Like we think of autonomous vehicles, the exact same thing. It's like you kind of have to pioneer it first, simulate it first, then slowly dominate it and maybe then you can change it afterwards.

I think like in terms of the skin suits and all this kind of stuff and looking as human as possible, I think there's going to be a real resist. There'll be a real resistance to humanoid robots generally because of like workforce shortages and all this kind of stuff. But I think in terms of like humanoid robots that look and act human, I think there'll be a real resistance to that. I imagine there'll be regulations that prevent that. And I think that for the next 20 years at least, we're going to see humanoid robots that look very distinctively robotic.

Yeah, makes sense. Cool. So, that's the humanoid stuff out the way. Easy. Uh, what else should we be thinking about in the future of robotics for you? I think the another big thing is obviously we've seen this Gen AI and LLM boom in the last two to three years and the application of that to robotics is quite important like before it was very much it would be designed in such a way and programmed in such a way that it had limited responses and limited adaptability and if it needed to be updated it would be updated by a techie or and that would you know put update the programming language update the software so that it then could do more but more and more we're seeing with these the boom in LLMs and we could we're starting to see a boom in AI native robots as well.

So robots that can really speak human language or understand human language so they can interact with a normal end users not necessarily technical and basically turn that into some important update or some important reaction. I think the rise of robots that can understand human language and human kind of requests is booming right now especially in like the retail sector as Yeah. And this is really interesting, right? because uh I've I've seen a lot of people I think one of the guys from OpenAI has been quite vocal about the fact that intelligence through you know LLM is is going to reach a peak and even if we can't understand or describe um you know whe whether the int how the intelligence works right even if it's not completely explainable it will supersede humanity at some point whether it's you know want to classify as AGI or not but we're going to have synthesized a a digital or computerized AI based understanding of how to do things that human beings don't have and then robots are the

perfect vessel that you you know they are compatible with this knowledge right human beings aren't compatible with it necessarily uh you because these AI models are like somewhat in some sense a black box right it's quite hard to understand how we get the output but if you can plug it into a robot it can take that internal knowledge however that works and apply it to the real world whereas human beings we can't So in a way it's like giving a physical domain and a physical agency to the capability of LLMs right that we we can't as humans necessarily do.

So that's really interesting for me as well. Yeah. I think like one of the thing I sort of described is like this is enabling a shift from like rulebased behavior to like AI native agentbased behavior. And it's things like right now imagine you have a lab of 50 people and they're testing these robots. A lot of it's simulated and then a lot of it's practical. Imagine then you build 10 million robots and you put them into the real world. Like it's really hard to define like all the possible scenarios that can come up.

But imagine all of a sudden you have a robot that can be taught through human language. And it say doesn't know what a spoon is or something like this. And then a human says oh what go you know use that thing over there pick up that thing that's what allows you to eat soup. And because it can understand human language it can then deduce in a way that that thing is a spoon that the human is pointing at. So you can suddenly turn like all these 10 million robots with basically people around them who are human prompting them in human language into people that are training those robots and again like they're all collectively improving.

So I think purely from like an operational standpoint it makes the idea of these being deployed in the real world and updating and learning constantly with human promptable language much more realistic. Yeah. And it's that general purpose thing again. Right. If it's general purpose and it can learn, then it's even more general purpose and also personalizable, right? And and you can make it specific. So, if I have my my my domestic humanoid robot in the future, I can show it how I like to do the dishes and and you know, it's not going to use the the pre-programmed version.

I can show I like to rinse my dishes before they go in and I like to do it this way and you can teach it. That kind of thing. I think is kind of a funny example, but I think there is actually going to be a lot of use cases for that. You're definitely gonna call your humanoid robot Jack Davis and you're gonna argue about who's who's washing the dishes properly and who's not washing which one's gonna pass the cheing test. Yeah. Um so the next big area for boom we kind of talked a little bit about it already is the idea of swarm robotics.

Now swarm robotics are quite different to humanoid robotics. instead of like building big bigger more expensive robots, right? Researchers are starting to kind of take a lot of inspiration from nature and like ants and bees and things like that and are trying to evolve the idea of coordinating swarms of small affordable robots. Like I say, you can take this example from bees, ants, flocking birds. And the point here is they're quite cheap and easy to build and individually they don't have much intelligence but collectively as you know a swarm of many many individuals they can perform like quite intelligent simulating tasks like the idea of exploration on a great scale or the idea of like moving objects through things in a great scale.

I think like a real simple explanation of this is like because there's so many of them and because they're each testing things out individually like over say a million robots testing one thing simply is the idea of a brain computing one million potential simulations at any one time. So they're they're quite cheap to build. There's lots of them but they can produce great outputs because of that. Yeah, that's really actually I hadn't thought about it. I was just thinking when you said Swarm Robotics I was thinking this is the Black Mirror episode from the first series, right?

This is this is scary. Um but yeah, it it definitely makes sense from that iteration perspective. You're right. And in terms of being able to like parallelize how to solve a problem across many smaller different things. Um yeah, I think it also makes sense from like the scalability point of view as well. Like there are some problems that are just so big and so difficult to solve like you know um removing plastic from the ocean for example. I can imagine that being one of those cases where yeah, we're going to have to build a billion robots to go and do this because there's no way that a few volunteers, you know, as as as noble as that is are going to manage to do it.

So yeah, I I think that's a really interesting one. I because I don't think this is also uh something that I've person or most people will have much contact with this idea of swarm robotics be beyond the the disaster scenario in something black like Black Mirror. But yeah, I think there are going to be big applications for it too. Yeah. A good example is like the idea I know NASA's big into this for like exploration of the moon and Mars and things like that rather than sending all of their eggs in one basket and like a Mars rover sending lots of little things to explore collaboratively.

A good example as well is like detecting oil spills and things like that. Rather than having like one drone that needs to understand the entire environment and have many sensors feeding into that, you have lots of tiny little things that can say just detect a concentration of some particulate and then collectively over all of them they can work out roughly where there's more particulates and where there's more concentration and from that they can say do some action like go towards it and start to cluster all this kind of stuff.

Yeah, I think it's like quite hard to get your head around this stuff because we don't see it in day-to-day life. But the main point is rather than one big robot that's super smart, has lots of sensors, or we have lots of smaller, maybe less intelligent robots, but they work collaboratively together. And working collaboratively together allows them to create like much more impactful outcomes overall. So the idea of like exploration, detecting particulates, working towards it, that kind of thing. But yeah, I think this is a big area that we'll start to see for agriculture, for exploration, undersea exploration, and pollution is quite a good one as well.

So the next one, something that maybe a lot of people haven't really experienced much, we kind of mentioned it around exoskeletons and like medical stuff is around the idea of soft robotics. This is something my friend actually worked on previously, so I know somewhat a little bit about this stuff. But when we think of robots, we typically think like big metal actuators, like scary, aggressive, all this kind of stuff. So imagine you have a giant actuator attached to your arm and it's like helping you with like a broken arm or something like that.

That's quite scary. if something goes wrong like it can snap your arm, it can like do more damage than it can actually help. So the idea of like soft robotics is unlike traditional robots, they create they're kind of more like flexible machines that bend, stretch, and often mimic like biological mushland tissue. So it's a lot of like um actuation through expansive silicons or inflatables and things like that. You have soft grippers that are used for processing foods and they bend around the food rather than like apply lots of pressure to get the food which is obviously good for things like grabbing an egg or something like this.

Yeah, this is I hadn't really heard of this as a term soft robotics but it makes sense. And when we were talking in the previous episode about uh you know the different elements of robots and taking inspiration from human beings or biology itself, right? This is I guess where the biological inspired material science is basically the the big thing, right? It's how do you how do you mimic some of the properties and the benefits you get in the in the material science part of the biological world but but then apply the other principles of robotics on top of that, right?

So like having automation within that because I I like that idea of like it's kind of a Yeah, it's it's soft it's softer, right? It's it's less harsh. It's less risky. It should be safer in many because yeah, I guess I think I've seen this kind of thing in like replacement limbs, I guess, as well because you want them to be, you know, a little bit um more pliable and amanable to to be used with a human. You don't want to have this big bulky metal arm basically, right?

You want to have something that that feels a little bit more like it's it's real. So, yeah, I like that. Yeah. It's also like about efficiency as well. we're seeing like muscle is human muscle or I guess biological organic muscle is very efficient in terms of the strength per weight and very efficient in terms of the energy expenditure and right now actuators just don't cut it like they're not on the same level. So I think we've seen a massive um push to kind of replicate human muscle for just efficiency for human robots especially but for efficiency more generally and yeah this is a big area for development and something that's fairly novel again it's not something that most people think of when they think robotics but like you say it's ideal for working with humans um not just near them because it's safer more adaptable and much enables much more complex manipulation than an actuator that's typically that one dimensional one degree of freedom so I mean the main thing after that is going to

be how all this stuff interacts with society. And there is, we've kind of talked about it already, but there is inevitably going to be a lot of push back in terms of like regulation ethically because there's going to be so much human impact. I think of all the sectors we've seen, of all the things or topics we've talked about so far, the idea of like intelligent humanoid robots is going to be the most disruptive to society of anything that has ever been created. I really genuinely do believe that.

And with that there is going to be a push back like we've talked about the multifaceted benefits that going to come from that like the idea of replacing labor shortages bringing costs down making you know more effective and efficient systems and sectors but people will often think about this in maybe and it makes sense more reactionary and shortsighted fashion because at the end of the day this could be replacing them and that is quite scary when there's no say potential opportunity that comes about because of it.

Yeah, I think this is going to be you're right. This is going to be the biggest transformation in probably in the history of humanity so far. And we've seen how much just AI as in I guess the robotized version of human communications has upended the world and has caused a real problem with the legislators. Like how do you legislate for this, right? And replacing humans in just this kind of very soft domain. If you then take that to the real world, the physical world, replacing human labor, that as we covered at length, there's so many different things to consider here, not least from the the whole human impact, the jobs, what's the economy going to look like?

Will it still work? Will capitalism still work? Will we need universal basic income? Right? All all these things are going to come into play. And then you also have the other side of I guess that's the kind of the societal level uh legislation like how do we how does society work fundamentally you also have all the other regulations and laws you're going to need and protections around what risks are you introducing by introducing every robot and every function it does like you know self-driving cars as one example those humanoid robots that are going to start working with us out in the fields uh picking you know fruit and crops like that's going to be that's going to be interesting as well like you what we said before, if a robot bumps into a robot, that's not a problem, but what about this transition phase?

If we have a robot and a human in a field, like how do we legislate for that as well? And again, I'd point you to point you to all the sci the sci-fi literature and films for for explorations of all this stuff, but yeah, I think I'm I'm put this I'm glad I'm not a regulator or a legislator right now. Yeah, I think they're kind of like two major camps. One is that this will be lead to a utopian future where like because of the the the mass cheaper formidable um ethical labor that we can create from this.

It should hopefully create degrees of productivity and output that we haven't seen before. And that will mean absolute plummet in costs across the board. Plummet in energy, plummet in food, plummet plummet in, you know, health care costs and all this kind of stuff. And that will be good for humanity. It will mean everyone can do things cheaper. there'll be no need for us to be competitive anymore. There'll be an abundance of all resources. Great. Fantastic. Or there's the other side is that similar to what we're seeing with AI potentially right now is that it will just lead to a concentration of wealth of the creators of those robotic companies and because of the job s shortages that will be created because of the mass unemployment that will inevitably be created from this and the lack of responsiveness that will come from governments.

It will lead to a concentration of power and wealth that will lead to kind of unparalleled levels of inequality in society. And I can see both outcomes and I think what we will actually see is maybe it playing out per region or per country depending on how things are regulated and how it plays out. I understand I think there'll be two camps. There will be the hum human humanity camp that will really resist these changes, these humanoid robotics changes. And there will be like the kind of the futurist camp that needs the productivity, needs the output, cares about maybe the profitability that comes from that.

And I think that will be the thing that really defines humanity going forward is how pro- robotics and AI are you and how human focused and AI are you against? And yeah, it's quite scary. I think the other thing we're really going to see is that as humans we love to humanify things and personify things and make try and connect to things. I think it's right now it's we're already seeing people connect to AI bots, LLMs through a computer screen. Like it's bonkers that that's happening, but we have this like inbuilt desire to connect to things.

So, it's kind of inevitable that happens. As we see those things working like LLMs working on humanoid robots, there's going to be a whole like massive rise in robotic rights. Are they human? Are they conscious? And then again, I don't know where I stand on this like when they are mimicking intelligence to such an extent that we can't tell the difference anymore. Like are they conscious? Are they intelligent? But I think that is going to be one of the big angles we see is as we start to use these humanoid robots for basically 247 laborious tasks like what we typically maybe consider like slavery if a human was in those in those positions will people start to feel for them and will there be like a reaction from these super robot AI loving fat sectors of society who want to kind of consider them human as well and I think that that's kind of inevitable as well.

Well, yeah. And the only remaining question on all that is who's controlling them as well, right? Like is it and which camp do they belong in? That'll be that'll probably be the the the king making factor, right? Is who whichever company is is is owning the biggest one, whichever one's the Skynet of the future, right? Whether it's Jeff Bezos or someone else. I think that's going to be, you know, their opinion on this is going to put them arguably as the most powerful person in the world at some And you you'll see like I think there'll be a sovereign state kind of level control as well where it's like no sovereign state is going to enable 20 million robots controlled by another sovereign state to be running like all of them manufacturing agriculture and surgeries.

Right? So there also will be a control aspect as we're kind of seeing right now with AI and LLM. So state controlled swarm humanoid robots. I can't see that going wrong at all. But yeah, I think that's a good place to end this odyssey on robotics. So yeah, we've covered I think everything there is to to say robot. There's so much more to cover. Like honestly, there's about 10 more episodes on this stuff, but not in this series anyway. We'll leave it and cover some other web three things after this.

Yeah, you can reach out to Alec if you want to if you want to do I think three hours is my limit. But yeah, I really enjoyed this and yeah, thanks for thanks for another another day in the in school with you. Like it's been good fun. Um yeah, and so all it remains to say is thank you for listening and you join us next time to untangle a little more web

Topics: AI & Robotics · Blockchain & Crypto · Business & Society

Related episodes

All episodes