When robots recycle with Matanya Horowitz of AMP Robotics

0:00Wondery Plus subscribers can listen to How I Built This Early Early and Ad Free right now. Join Wondery Plus in the Wondery app or on Apple Podcast. Listening on Audible helps your imagination soar. Whether you listen to stories, motivation, expert advice, any genre you love, you can be inspired to imagine new worlds, new possibilities, new ways of thinking. Audible has an incredible selection with over 1 million audiobooks, podcasts, and Audible originals, all in one easy app. Find the genres you love and discover new ones. explore bestsellers new releases plus thousands of included audiobooks podcasts and originals that members can listen to all they want with more added all the time enjoy audible anytime while doing other things household chores exercising on the road commuting you name it audible makes it easy to be inspired and entertained as part of your everyday routine without needing to set aside extra time there's more to imagine when you listen.

1:06Sign up for a free 30-day Audible trial and your first audiobook is free. Visit audible.com slash built. I've stayed in awesome homes on Airbnb in places like Athens and Berlin and Rome, and each time these places have given our family a chance to feel like locals. It's the best way to travel. Picture this. You're spending a cozy evening out of town in a home you booked on Airbnb. The space is thoughtfully designed and you think, I bet I could host my own home on Airbnb. Your home might be worth more than you think. Find out how much at airbnb.com slash host. If you're waiting for your AI to turn into ROI, maybe you need to do more than wait.

1:51Any business can use AI. IBM helps you use AI to change how you do business. Let's create smarter business, IBM. Hello and welcome to How I Built This Lab. I'm Guy Raz. Right before the pandemic, I joined my son on a school field trip to a recycling center in the Bay Area here in California. And I was really surprised to see how it works. There was a giant conveyor belt of glass and plastic and paper all mixed up in a row of people sorting it as fast as they could. And while the workers were amazing and fast, there was a ton of stuff that was inevitably left behind and not recycled. In the U.S., the EPA estimates that only a third of all waste actually gets recycled.

2:41And one of the main reasons for this? Well, most recycling facilities just aren't equipped to sort through all the materials they receive. My guest today, Matanya Horowitz, launched Amp Robotics in 2014 to tackle this very problem. The company uses artificial intelligence and robots to sort through waste more efficiently, which means recycling facilities across the world are able to actually recycle more of the material they receive. Matanya studied robotics at Caltech, and it was there where he discovered deep learning. It's a method in artificial intelligence that enables computers to process data like the human brain.

3:22You could just start to see that machines were going to be able to see as well as a person for the first time in a very kind of fundamental and true way. And so even though that wasn't my focus area, I was like, oh my gosh, this is going to be massive. So I started focusing on that. I started studying it. And after graduating, I was looking for areas where I thought it could be useful because I was like, this might be the next mobile phone sort of wave. It could be the next internet. I don't know, but I just know it's going to be a big deal. All right. So you're seeing the beginnings of what will and is already starting to revolutionize industry in the U.S.

3:57and even consumer products. And you get your Ph.D. from Caltech in 2014. And with obviously a lot of experience working in robotics. So you start to explore, hey, what could I do? Where can I actually make a contribution? and of course we know you end up in in the sector of recycling but but how did you come to that so i i wish i had a story that was uh about me really being an expert in recycling or maybe i saw a plastic bag blow across the road and it made me sad or something but i um it was a very academic process i i said to myself okay you have a machine that could see roughly as well as a person what does that open up um there were different things in drones i was pretty interested in.

4:42I had an idea for a drone you'd kind of throw in the air and it would follow you and create kind of like a highlights reel. You could throw it in the air while you're snowboarding. I want that. Yeah. You know, I think somebody actually went ahead and did that. And I saw that you'd have to raise a lot of money for drones. It seemed like it was getting very crowded. So this is around 2013, 2014. I started researching a couple of different industries, but I visited a recycling facility in Los Angeles. And if you've ever visited a recycling facility, you'll see that the main task is people standing around conveyor belts sorting through material by hand.

5:18Yeah. You have people picking bottles, cans, cardboard, paper. It's a tough job. It doesn't smell great. And the turnover is very high. People don't kind of put their full heart into the job. And so the The material quality is lower than it needs to be. And it was really clear that this is an industry where robotics could really have an impact. I would talk to the recycling facility owners and managers, and they would say, this industry absolutely needs robots. I don't understand why we can't get robots. And with a vision system and technology that could see roughly as well as a person, you could bring that tool to this problem.

5:56Yeah. You know, I want to just step back and talk to you about recycling. Only about 60 % of American households have access to like curbside recycling programs. Right. You know, for most of us, we just dump it all into one bin, you know, all recyclables, or maybe you sort paper and then everything else, cans, glass, bottles. And then it gets dumped into a facility. I've been to one. I went to one in Berkeley, California, a couple years ago, and was amazed because it is a conveyor belt. And there are humans just sorting things out on the conveyor belt as it passes by. And a lot of things just naturally don't get sorted or, you know, like a bottle gets tossed in with aluminum.

6:37I mean, it's just going to happen. And these facilities are just super inefficient. And I was kind of surprised learning that only about 35 percent of waste is recycled in the U.S., even though much more of it is recyclable. Right. Right. Well, there's a couple kind of core challenges to the industry that you're hitting on. You know, one of the main one that we're making a big impact with is technical. All this material you put in the recycling bin, it gets smashed, it gets folded, it's dirty, it's torn in different ways. And so you have this huge stream of material that is inconsistent. And that inconsistency makes it very hard to separate.

7:19The core task of the industry is to separate out these different commodities. Yeah. You want to get all the number one plastics in one spot, the number two plastics. But if they're smashed up, if they're folded, they're kind of mushed together. It's hard to know what's what. Exactly. And that core technical problem is what's made the industry so reliant on people. The second problem is really a problem of incentives. If you can separate material out, is it going to be worth separating out? Can you sell it for a profit? And because of the need for all of this manual labor, what you find is for a lot of materials, It may not be profitable to actually recycle them, which is unfortunate because what you find is the vast majority of materials can be reused or recycled in some way.

7:59But for many of them, the two-thirds of material that isn't getting recycled today, that incentive isn't very strong. And so it's hard to build a business. In other words, it's not strong because the cost, particularly the labor cost of recycling all this stuff is sometimes higher than the price you get for selling recycled materials, right? Yeah. Well, and to give you a couple of figures that I think are helpful, most of these materials, they live in a commodity market. So, you know, oil goes up and down, plastic goes up and down. So sometimes, you know, just by virtue of what the economy is doing, recycling becomes a better or worse business, you know, setting aside any technology like what we have.

8:41But some rough figures here are that the recyclables you put in your recycling bin, depending on that commodity market, are probably worth about$100 per ton. And that's real value. Like you can pull out that plastic and sell it. The trouble is, is that the cost of running it through a recycling facility ends up being about$100 a ton. Depends on how modern the facility is and these things. But you're kind of at this point where it's a little bit marginal. But what you can see is if with technology you can bring that cost of sorting down to$50 a ton or$20 a ton, and also if you make it worthwhile to sort more materials, that$100 a ton of value can become more like$120 or$140, either by purifying the material or being more specific about what you separate out.

9:25And now recycling starts to become this phenomenal business. Wow. Now you've sort of incentivized more recycling. So a significant amount of material is recycled, but it's only about a third. And there's a lot that is thrown away. And plastic in particular, I think is problematic because not only, as you mentioned, not only does its price go up and down depending on the price of oil, sometimes it's just cheaper to make virgin plastic than it is to recycle it. But plastic also degrades after several uses, like recycled plastic over time becomes toxic, essentially. Well, yeah, there's a whole bunch of different kind of issues around plastic.

10:08You know, the one that I'm most focused on personally is sort of plastics in the ocean and how do you really keep all these plastics from getting there. Yeah. And it's an interesting story, but, you know, all of these challenges around plastic, the real hard part about it is it's kind of matched with plastics being this wonderful material. Like it's so strong, it's so lightweight, it can keep our food safe. And so, you know, as I see it, the most important thing that, you know, we can do is we make it so that the world tries to capture that plastic and get value out of it so that we can make sure it's properly handled and doesn't make it into our environment and into our, you know, food or, you know, whatever.

10:48and what I get most excited about over the long term is if you make recycling this fundamentally better business you not only create an incentive to expand recycling in the US or in Europe but if you can make it so that there's a stronger incentive to recycle in the developing world now you really make a huge impact because it ends up a lot of these plastics in the ocean are kind of coming from developing nations where they don't have strong waste infrastructure that's the the dent we ultimately want to make. We're going to take a short break, but when we come back, Matanya explains how he got AMP robotics off the ground and why that involved some dumpster diving.

11:27Stay with us. I'm Guy Raz, and you're listening to How I Built This Lab.

11:35One of the coolest things I did last summer was take my family to Athens, where we saw the Parthenon and the Agora and all those amazing neighborhoods and ancient historical sites. And one of the things that made it so awesome was the home we booked on Airbnb. We literally had a view of the Parthenon from our bedroom window. There was a hot tub on the top floor. We could sit in it overlooking the entire city. And we had a kitchen and access to markets. And we were in walking distance from all of the things you would want to see. It made the trip so, so amazing. And when you take your own vacation, that's actually a great time to host your home on Airbnb.

12:13Your place with your cool art collection and your handy kitchen gadgets, it might just be what someone else needs to feel right at home on their next trip. Plus, your earnings from hosting provide you with another income stream and could even help offset the cost of your next trip. Your home might be worth more than you think. Find out how much at Airbnb.com slash host. This show is brought to you by American Express. Running a business means making countless choices every day, some small and some that can change everything. The business owners who thrive are the ones who are ready for whatever comes next.

12:53And having the right support behind those decisions can make all the difference. That's why so many business owners I talk to choose the American Express Business Platinum Card. It's built to help them move quickly and take their business further. And with five times membership rewards points on flights and prepaid hotels booked through Amextravel.com, you can turn any business trip into a chance to earn more and invest in your business's future. Plus, you get a flexible spending limit that changes as your business does, adapting to where you are and where you're headed next. Because when it comes to growing your business, there's truly nothing like business platinum.

13:37Not all purchases will be approved. Terms apply. Learn more at go.amex.com. You know, one of the most surprising things about how I built this is just how many pitches we get every year. tens of thousands of businesses reach out hoping to tell their story. Now, while we do try to read most of those pitches, we can't possibly read all, but we choose most of our guests through our own deep research. And that means my team and I spend hundreds of hours digging into companies, what they make, who's behind them, and whether their journey can inspire you. And that's exactly where Claude has become such an incredible thinking partner in my creative process.

14:19Claude is the AI for minds that don't stop at good enough, the thinking partner that works with you to explore the things that fascinate you. Whether you're researching late-night curiosity spirals or working through complex creative challenges, Claude extends your thinking to tackle the problems together. For example, imagine I'm about to sit down with the founder of a snack company. I'll want to know what products they've released, how they've evolved, and which competitors they're up against. In the past, it could take hours to track down reliable details. We'd search articles, databases, reports.

14:54Now, Claude can give me a clear answer in seconds, and then I can decide what to dig into further. It doesn't replace my process, but it sharpens it. And that means I can spend more time doing what I love most, focusing on the human stories, the lessons, the decisions, and the struggles behind the businesses you hear on this show. Experience what collaborative thinking feels like at claude.ai.hibt and see why the world's best problem solvers choose Claude as their thinking partner. One more thing before we get back to the show, please make sure to click the follow button on your podcast app so you never miss a new episode of the show.

15:40It's usually just at the top of the app and it's totally free.

15:51Welcome back to How I Built This Lab. I'm Guy Raz, and my guest today is Matanya Horowitz, founder and CEO of Amp Robotics. All right, so you are searching around for how you can be part of this new emerging technology, AI, combined with robotics, and you land on recycling. you realize that part of the problem is just the human problem, just sorting this stuff accurately. And so you decide to tackle this problem and see if you could launch a company that could essentially automate this. So tell me about the idea that you decided to come up with. Yeah, well, I was focused on where is a spot where this artificial intelligence technology was the core linchpin to building a new product.

16:38So I started focusing on recycling and I liked it because it seemed like you could use robots that had already been around for decades if you could pair them with this vision system. And I thought that was really good. You should have one hard technical problem. That was going to be the vision system. You didn't want to have to innovate on the robotic side, at least not too much. That was at least the idea. So started kind of zeroing in on this idea of building a robot that could be bolted onto a conveyor belt and just pick stuff. Like that's really all we're trying to do is pick a bottle and pick a can.

17:08I applied for a grant from the National Science Foundation. There's a program called the SBIR, Small Business Innovations Research Program. That gave me some seed funding to get started and started building a little prototype. There was, you know, a lot of kind of good fortune on the way. Initially, I couldn't afford a conveyor belt and got a grant from Oscar Blues, the brewing company. Oh, yeah. For recycling stuff. and you know i was applying for these grants i was visiting recycling facilities i didn't know the industry that well and so i was honestly saying a lot of dumb stuff asking a lot of dumb questions and you know i'm kind of amazed looking back that people would give me the time of day but but yeah started kind of building this robot the first data sets for the vision system um so you sort of need examples of bottles and cans to train this artificial intelligence stuff yeah i did dumpster diving for them um was going to recycling centers and pulling out bottles and cans and, yeah, just started hacking away at it.

18:04I mean, this is a hardware and a software challenge, right? Because you've got to build the hardware to be physically resilient, right? Because there's a lot of just grossness that it's going to deal with. And, you know, there's probably going to be rotten milk or rotten things, right? In that conveyor belt. And then there's the software side, which is it's got to identify the different plastics. What's glass? What's paper? What's right? and then pick it up quickly. Tell me how you started to work on building a machine that could identify what it's picking up and sorting out. So the interesting thing was the computer vision technology, this artificial intelligence, actually started to work really well almost from the beginning.

18:49Wow. We had some insights really early on as to what would be important to these algorithms and had a lot of success. Initially, we weren't at 99 % accuracy, Um, but one of the nice things about recycling is honestly, the bar is a little low, like it's hard to consistently create pure material. And so as compared to automotive, where you can't run over one out of a hundred people, you see, you know, for us, if you mess up on one bottle or can, you're still doing a lot better than the status quo. So it's a very forgiving environment. And we were able to kind of meet the standards of the industry pretty quickly.

19:24The part we thought would be easy, which was using these off-the-shelf robots, that ended up being really tough. So rotten milk sounds gross, but the real problem with it is that as it dries out, it starts to get really sticky. It's sticky with a robot. It's not good. And these robots had like claws, like you'd see in an arcade with a claw machine, or how are they picking up the recycled material? So, you know, we actually did start with claws, and that was basically what my PhD was on, was having robots use claws in interesting ways. And it's really hard. It's just really hard to imitate what people do.

19:59There's so many sophisticated things you do. Like when you're picking up something that's flat, like a flattened carton, if you kind of do it slowly, you'll see you do things like you push down on a corner of it to get the opposite corner to lift up. And then that lets you wrap your fingers around. And it's very hard to get a robot to do this kind of, kind of sophisticated multi-step planning. And what I saw was it was going to be tough. and we ended up switching to a whole different technology that I didn't even know much about in graduate school, which was using vacuum technology. So basically using something that's an industrialized system, but it's not that different from your vacuum cleaner at home.

20:35And you think about, okay, I can take my vacuum cleaner and I can kind of point it at an aluminum can. I'll probably be able to pick up the aluminum can. That's kind of what we started to do. Just point a vacuum at the thing you want and pick it up. And we found we could get that to work really well. So we kind of hit the easy button on this. I mean, it's not that easy. I've spent like nine years of my life thinking about vacuums. I've seen a video of this. It's like tubes, right? There's like a few tubes over the conveyor belt and they'll just pop down and suck up each piece of trash. And it happens really fast, like in a millisecond.

21:11But how did you make it so it wouldn't just suck up like four things, right? Instead of just the actual thing it needs to suck up. you end up having this trade-off exactly like you're describing it can't be too big or it'll pick up too much it can't be too small um we rely a lot on our vision system to look at the items even if they're overlapping and only target kind of promising areas to pick things up and it's gotten pretty sophisticated there's multiple layers the thing self-cleans it reverses the airflow after every uh pick to clean itself out uh in case they picked up some plastic bags or something but it's a tough problem we're still working at it we're constantly trying to make it better.

21:48The hardest part out of all of it ended up being keeping it from getting gunked up with that rotten milk, pieces of glass debris. There's so much stuff that can just kind of accumulate in a recycling facility. And what we found was even once we made something that worked, most concepts for doing this will get gunked up with stuff within half an hour. And we needed something that would last days. And that ended up being the hardest piece of it. So do you have a sense of the accuracy? I mean, how much, so for example, in a traditional recycling center where it's sorted by humans, I have to imagine that oftentimes, you know, some glass gets into the aluminum cans, you know, and some paper gets into the plastic.

22:27How much more accurate are these machines and humans? So there's kind of two ways to answer. So one is a person can actually be very good and produce above 99 % purity, but they almost never do just because the staff tends to not be motivated and It's a tough job and you start zoning out and things. And so what you'll see is for most commodities, the recycling facilities will produce material that kind of bumps around 90 or 95 % pure. And so our facilities, the biggest thing is besides being fully automated and having this lower cost structure as a result is they're just extremely consistent, like constantly producing material of the same purity.

23:06But yeah, we'll be 95 to 99%. Interestingly, what you find is, you know, there's this broader supply chain. we're selling the plastics to somebody and they take these plastics. And because the industry can't get to a high enough purity consistently, those buyers will actually have a whole sorting process where they sort it again. They're cleaning up what they just bought. We sell those guys robots. So our robots are helping clean the stuff up. And so the result though, is, is we can calibrate the purity to what the buyer will pay. So if they pay a premium, you know, we can hit a very high standard.

23:40You know, if they're looking for kind of something that's middle of the road, more like 95%, we'll just produce at 95%. Let's talk about your customer. Who in the United States, who handles recycling? Is it because I think a lot of people assume it's like the government or municipal offices, but it's mostly not, right? It's private companies. Yeah, it'll be companies like Waste Management or Republic Services. So those are Fortune 500 companies. You'll find some small family businesses. We have one customer that's recycling and disposal services in Virginia. But the way the industry works is your city will have or your HOA will have a waste contract and they'll have a contract with one of these groups.

24:19Some municipalities decide to do it themselves, but for the most part, it's a company and they'll take your trash and they'll take your recycling. They'll bring the trash to a landfill or sometimes an incineration facility. They'll take the recycling and sort it out and run that business. And yeah, the municipalities typically aren't doing all of it themselves, but in some cases they do. So when you started to – I mean when you began this journey and you started to raise money for it, presumably you had to sort of say, hey, this is a big market opportunity, right? Because look, let's be honest.

24:55Nobody is – I mean as much as they say they are – is investing just out of the goodness of their hearts. They want to make a return. And if presumably most of the customers here are companies that are already recycling, and if recycling isn't a big driver of revenue, right, for some of these big companies like waste management, which makes its money from collecting garbage, how do you make the case that this is actually going to be, this is going to make money for them? Yeah. So fundraising was very hard for most of the life of the company for exactly this reason. We started with these sorting robots and we kind of described this one that uses a suction cup.

25:34And really that's actually not so much about the recycling business. It's really about kind of labor and automation. And so you look at the market and you say, okay, there's tens of thousands of people doing this manual sorting task. There's a potential for tens of thousands of robots. and you do the math on that and you find that that market's worth a couple billion dollars and that that can be exciting for a certain class of investor but when they look at it they'll say okay there's municipal contracts in this industry there's a high degree of concentration into the biggest waste players that makes it really hard to sell lots of units these things just slow it down and so what we found was this kind of story around automation didn't really attract a whole lot of investors.