Recycling: An Opaque Industry – Robohub

On Oct 10, 2022

The Recycling Industry in the United States is a for-profit industry. They profit from taking recyclable material, refining it, and reselling it to companies at a cheaper price than producing the material from scratch.

Supply and Demand in Recycling

If you look at the demand side of the recycling industry, an array of multi-billion dollar companies like Coca-Cola and PepsiCo are incentivized to buy recycled goods and reduce their materials costs.

On the supply side, ~300 million tons of trash are generated annually in the United States. Estimates suggest that up to 75% of that is recyclable.

On paper, it seems clear that maximizing the amount of trash the US recycles is in everyone’s interest. One issue though, less than a third of the trash ends up recycled.

Areeb, co-founder of Glacier, breaks down the multi-layered reasoning behind why the Recycling industry cannot handle this volume of trash, and what Glacier is doing to address this.

Areeb Malik

Areeb Malik is the Co-Founder of Glacier, and he is on a personal mission to fight climate change and extract value from the $123B worth of recyclables that fill the landfills and oceans. Before founding Glacier, Areeb was a Software Engineer at Facebook, where he used Machine Learning and Computer Vision to build out new product features.

Links

transcript

Areeb Malik: Thank you so much for having me, Abate.

Abate: Before we dive into what glacier is doing, I’d love to just get an overview of the recycling industry.

Areeb Malik: Yeah, sure. I think the recycling and the waste industry is one of those spots that we as consumers, you know, rely really heavily upon. Most of us don’t really know how it works. And that was certainly the case with me before I started working at Glacier. it’s pretty fascinating. It’s a, it’s a really big industry.

There are a lot of players involved and there’s a lot of money that kind of flows through behind the scenes. I think something that people don’t think about is we spend quite a bit of money every month having our trash and our recycling and our compost just magically disappear from our curbs. Right?

It’s like many multiples of what we pay for our Netflix subscriptions. it goes to, it goes to the waste industry. and the reason for that is cuz it’s a lot of work behind the scenes. generally speaking, I’ll kind of run through it at a high level.

_*How does recycling work? *_

Well, you as a consumer or you as a business, take your recycling bin out to the curb. That’s kinda the end of your involvement.

A truck comes over. These trucks are known as _*haulers*_ in the industry. they dump all of your bins into their truck and drive around, click all your neighbor’s stuff and take it to a facility. Known as a “materials recovery facility” or a MRF (Pronounced “Merf”) for shorts.

Love the name Merf.

Um, that MRF’s job is to sort. So you put a bunch of stuff into the bin a lot of it wasn’t recyclable because as a consumer, you might not know what is recyclable or isn’t. and then within the stuff that’s actually recyclable it needs to be sorted apart. Cause if you want to actually do something with those recyclable commodities, you need to sort the bottles apart from the cans, apart from the cardboard. Et cetera.

Um, this big sorting challenge is quite hard to do. we can dive more into how they do it cause that’s where our company focuses. but to just give you the next steps in what happens is, you know, this MRF creates, for instance, a big pile of aluminum cans. What it does is then condenses those cans into a cubic block, known as a bale, and sells these bales to what we call a reclaimer or a recycler who buys what they hope is just one commodity.

Right? So for instance, aluminum.

Um, they take that and they do some process with it to turn it back into either more aluminum cans or some other material that can be produced from recycled aluminum that happens for plastics that happens for cardboard. That happens for a bunch of different commodities.

Um, and the reason this is this is important is because that aluminum can and that plastic bottle, there’s a lot of energy, a lot of potential energy kinda stored in that commodity.

Um, and if you’re producing something new it’s actually a lot more* energy efficient* to use that *recycled commodity* than to go mine for raw ore, or get the petroleum for the new plastic bottle, or whatever you might be doing. And so you actually get really big energy savings by using recycled commodities instead of using virgin stock.

Um, so that’s kind, that’s kinda in the overview, obviously that, that reclaimer that recycler sells raw aluminum or raw plastic chips to some producer who makes new bottles or shirts or shoes or whatever, whatever they need to.

Abate: Yeah. And does that energy saving that you get from recycling? Does that also transfer into a lower overall cost than getting it from the Virgin material?

Areeb Malik: Yeah, absolutely.

if you’re a producer of aluminum cans, there are two ways you can produce them, right?

You can either pay for the raw materials. that’s gonna come in the form of aluminum, that aluminum that comes from raw materials is gonna have to pay for the mining and the shipping and, and everything that’s involved.

The, you know, Distillation of their raw ore into actual aluminum, everything that goes in that process versus you can buy aluminum from a recycler. that comes from all they had to do was take a bunch of cans and melt them down and clean them up a little bit. and so it’s actually a lot cheaper because that recycler didn’t spend as much time or effort kind of getting those that, that raw aluminum to you.

So, yeah. If we look at aluminum energy-wise, a recycled aluminum can takes 97% less energy to produce than a virgin can. I can’t tell you exactly what that means financially, but you can, you can kind of imagine the cost savings that come along with that.

Abate: Yeah, no, that’s a massive [00:05:00] difference. and so when, you know, you mentioned aluminum, so there are definitely certain materials that are more valuable to the businesses who are gonna be very incentivized at this point to actually invest in recycled materials than other ones. What would be the hierarchy of the very valuable recyclable components and then the ones that maybe we recycle, but it’s not as clear cut that there is a financial motive to do so?

Areeb Malik: Yeah, that’s a really good question. And it’s, it’s not a straightforward one. Unfortunately, the easy hierarchy I can give you right now is that metals are Supreme. Metals are great because, with aluminum or steel, they’re what we call infinitely recyclable. Right? You can take a can melt it down, shape it into a new can, put a drink in it, return it, melt it down make a new can and go in circles forever.

With something like plastic or, you know, we like to think of our plastics, not as just plastic, but the actual type of plastic resin. So P E T, which is your triangle one plastic as we call it. when you typically recycle plastic, you don’t actually come back exactly where you start, you tend to what we call a down cycle.

And that’s why you see a lot of companies selling shirts and shoes and bags that were once made from plastic bottles. Cause you can actually take the plastic bottle, spin it down into polyester fiber, and then sell that again. but then when you recycle that bag, right, is there a market for that?

So this is what we call down cycling. We kind of go around in a circle, but we end up producing something that just cannot be recycled at the end of the day. and so when you think about the hierarchy of commodities yeah, like metal’s supreme, Plastics are really valuable because there’s a lot of demand for them.

Paper is pretty solid. Cardboard right now, especially in the pandemic, they call brown gold because it’s super valuable as well. Newspaper, office prints, and stuff like that starts to get less valuable.* Glass is a really interesting one because it is, infinitely recyclable, but it’s also really heavy.*

And so if you have to drive a bunch of glass bottles from one facility to another. There’s a cost to that as well. and so that’s, that’s kind of mixed. a lot of facilities don’t deal with glass. A lot of them do it kind of depends on how close their buyer is.

Abate: Would the glass be melted down and reformed or it would just be literally reused as it is by the same company that made it?

Areeb Malik: Yeah. For the most part, it’s going to be melted down to reform. Cause that’s the process of glass is similar to that of metal. If you melt it down, it can kind of just reshape.

I would imagine there’s a big push by people who are they make canned soda to use recyclable metal. And so How does that motivation by these industries affect the recycling industry specifically “metal cans”?

Areeb Malik: Yeah. Great question. So if you, if you produce cans, let’s say you’re Coke, right?

And you produce aluminum cans for you, you might not care about the recyclability and, and the green thumb aspect, or you, you might and either way, you have this kind of financial incentive, right?

You say I want to produce a bunch of cans. I’m actually willing to pay 2 cents extra per can to be green and sustainable cuz that’s in our ESG guidelines company, whatever. either way, what you’re looking at is:

okay, cool. I can buy a can that’s made from virgin material or I can buy a can that’s made from recycled material.

Um, I said earlier that recycled material is cheaper, but then, the other trick there is that it’s not cheaper if there are no recycled cans to buy, right?

So if you are a aluminum can recycler and you don’t have any cans coming in, well, you can’t sell any out. And that reduction in supply is going to yield higher prices.

Um, so if you’re Coke and you’re looking at those two options, you’re gonna go with whatever is cheaper minus that fudge factor for your ESG kind of motivation.

Um, But, you know, if we can get the recycled can to be more abundant and we can make it cheaper for that recycled can turn into raw aluminum again, then Coke is all of a sudden… It doesn’t really matter if they’re a green thumb or not. They’re going to be incentivized to purchase recycled stock.

Because it’s cheaper and that’s kind of what the objective of the recycling industry is. And that’s what the objective of my company glacier is, as well as Can we manipulate the economics of recycling such that it doesn’t really matter if you’re, you know, trying to do good by the planet or not? The financially smart thing to do is to buy recycled.

Um, and that’s the, that’s the objective that the objective, I think we’re all pushing toward.

Abate: Do you have any idea what percentage of the actual goods, the cans that are being produced, are being recycled?

Areeb Malik: Yeah. it’s a great question. I think if I zoom into cans, I don’t have that number off the top of my head. And I think one of the things that you and I will kind of unearth over the course of this conversation is there’s not a lot of good data in this industry. it might not surprise you, but when you throw your bottle into the recycling, then nobody’s really tracking that bottle at that point.

Um, and so there is some information out there, but it’s not very robust. what I can tell you is that within the US, we recycle about [00:10:00] 30% of our waste stream. So the other 70% is either non-recyclable or is just never recovered. estimates right now say that we could recycle up to 75%. So if you look at the amount of trash we generate per year that’s 300 million tons within the US.

30% of that, that’s about a hundred million is recycled or composted. but that remaining there’s another 45% chunk that could be recycled or composted that is not being captured. and* that’s obviously a big problem,* right? And that’s gonna come to:

A. Consumers, not recycling properly.

B. Packaging is being made in a way such that it’s just not recyclable.

C. The markets not existing for certain types of recyclables.

For instance, a plastic bag in theory, yes. Could be recycled, but there’s no market for it. And so all of these things kind of combined together to say, yeah, we have 45% of the way stream that could be recycled. That is not currently being captured.

Um, how much of, you know, I’m sure the people who are putting stuff in the garbage can in the recycling bins, they’re not thinking that 70% of the stuff that they’re putting in there is not recyclable.

Abate: They’re thinking it’s maybe 90% and maybe threw in one thing that they’re kind of iffy about. so how much of this is a just consumer error? How much of this is the problem of the packaging? The people who make the packaging are somewhat misleading about whether or not it is recycling. And is this, a sort of government initiative that needs to happen that actually changes this?

Areeb Malik: Yeah. I would, I would say it’s it, it’s easy to blame the consumer here, but I would argue that it’s really not the consumer’s fault. I’ve been working in the recycling industry for three years now and I still have trouble. With like a yogurt container. I’m like, ah, like which bin does this go in? and it’s, it’s really hard to tell and it, the problem is like, it varies based on the individual packaging.

It varies on where in the country you are, what does your local MRF actually accept and sort? All of these questions kind of are very opaque. There’s no clear answer. And so you can’t really blame the consumer for this at this point. It’s, it’s such a hard problem to solve. No wonder they can’t figure it out.

Um, I think the biggest culprit right now is the fact that producers produce without thinking about the end of life of their product. So a lot of times you’ll go to the grocery store and you’ll get a thing a package or whatever, and it’ll be a combination for instance of multiple materials. a great example of this is Tetra pack.

Tetra pack is ubiquitous box for holding liquids or in juices, milk, whatever. and Tetra pack itself is made of several layers of different commodities. So you got paper, plastic, aluminum, all in the linings, and it does a great job preserving your drinks, but that material is super unrecoverable because how are you gonna peel the plastic and the aluminum apart, and then extract the aluminum.

All of that sort of stuff, right. There are processes in place for this, but as a consumer, like you’re not supposed to be able to do anything. And if you have to look at who’s to blame there, well, couldn’t you actually produce a container that’s designed for a good end of life, if you are some juice manufacturer and you want to make sure that your juice carton does have a good place in the circular economy?

Um, and unfortunately to my original point, right? Like they don’t care about this. What matters to them is being able to produce a lot, and do it at a cheap price. And so they go with whatever makes the most sense. again, the dream of recycling is can we make what makes the most sense of commodity that has perfect circularity?

Abate: Yeah. And so the local MRFs, these are, are these private institutions or are these government institutions?

Areeb Malik: Yeah, for the most part, they are privately owned. A couple of ’em are privately owned by public companies and about 10% are owned by municipalities.

And you know, one, one thing that you also mentioned earlier, there’s like these different categories that you can take your recycled trash and put it into metals, plastics, et cetera. But in a lot of people’s homes, all you have is a single blue recycling bin that you throw everything into. exactly.

Abate: Yeah. So, and then is this, is this a part of the issue? Is this a government thing where they’re not sending out the right bins to everybody’s houses, where they need to be able to sort it out individually by themselves?

Areeb Malik: it’s a, that’s a really fascinating question. It’s, it’s really hard to get good data on what would work here. So I’ll give you a couple examples. A lot of facil.., a lot of, a lot of communities around the country, they offer what we call dual stream recycling, where you actually have two recycling bins, right.

And the typical way they divide that is containers. That’s gonna be your glass, plastic, cans, steel, whatever. And then paper everything that’s, cardboard, mail, whatever. and there’s this concept known as dual stream recycling. It was actually a big thing back in the seventies, eighties, nineties, but it started to get phased out because they found that people, *consumers like didn’t have the mental energy to do this.*

And they found that if we get everybody to dump everything into one bin, we’ll at least capture all the stuff we care about, which is primarily metals and plastics. Right. And so this [00:15:00] change went into, into place because it just made sense for, for an industry that didn’t really have the, the strength to handle this much kind of inbound.

But what they ended up with is another problem where now everything’s mixed together. the, the other thing to think about is like what, what if every consumer was responsible for like dividing everything into every individual commodity type? obviously I, you can hear that and be like, that sounds, that sounds tiring.

Even, I would say that sounds tiring, cuz like I said before, I still have trouble splitting a thing into the rights, the right location, even with just two bins. there are communities, very small communities that do this. There’s a great one that you can read about in Japan, where I think they have like 30 or 40 different things, you have to sort your trash into you carry your bag over to a little building.

And there’s a guy in there who helps you sort all the parts and like, yeah, that like works for purity’s sake, but like how many people are actually gonna do that versus just dump it all into the, the landfill. Like it, it’s a, it’s a tricky trade off to play with and we don’t have a lot of good data to say which one really works better

Abate: An interesting thing about the United States, the scale of garbage is just significantly higher than a lot of other countries. partially because we’re producing more garbage per person and also partially is a very big country. yeah.

Areeb Malik: We are by population, I think one 20th or so of the world, but we produce 40% of its trash. So we’re a little bit skewed in our ratios there.

Abate: What is the scale of the actual trash being produced? Do you know what it is per person?

Areeb Malik: So if we look at the us again, we’re looking at about 300 million tons of trash per year. That’s that’s recyclable, compostable captured, not captured plus everything that should go to landfill of that 300 million, like I said, 30% about a hundred million is actually recycled or composted.

If you just kind of take that into, you know, the population in the US, which I think came out to 330 million in 2021, we’re looking at about for a household of four. you’re looking at about one ton of recyclables, not recycled every year. So every single house, a car’s weight of, of stuff goes not recycled right now because we don’t have the infrastructure behind the scenes actually handle it well.

Um, and if you take that number to the top, you’re looking at that three tons of stuff, consumed three tons of recyclables that you consume, as a household. and some of that gets recycled.

Abate: Yeah, which is a mind-blowing number. It’s a lot that you could be recapturing and reselling.

Areeb Malik: Yeah, definitely. It’s a lot to be recapturing. It’s a lot to think about, like, *why do we consume so much?* how are these other countries, these other people around the world getting by, without doing it? lot of, lot of questions that I raise and hopefully, you know, for the listeners now, they’re like, that’s raising some eyebrows, like, oh my gosh, my house is producing three tons of recyclable stuff, and only one of it’s getting recovered.

Abate: A certain portion of that would be businesses and other large industries that are also generating waste, not just the only the households.

Areeb Malik: Yeah, absolutely. Yeah. It’s, it’s certainly not just on households. Obviously, this number comes from you as a person. You know, you throw some stuff in your bin at home. Sometimes you go to Starbucks and you toss it into the street, trash cans, all, all that stuff kind of comes together. but it’s certainly a problem, no matter how you look at it.

Abate: So what is glacier’s approach to this problem?

Areeb Malik: Yeah. So I talked a little bit about that MRF right? Their job is to sort and the sorting, if you look at kind of the global picture of recycling from our opinion, the sorting piece has the biggest value to add with the smallest amount of effort. And so it seems like there are a lot of wins to be had in this sorting piece.

And the reason for this is twofold. One is being able to drive the cost of that sorting down will drive the cost of that material down. So we talked about that Coke example if you can get the cost of the recycle can down. Then you’re gonna be more incentivized as Coke to purchase recycled cans.

And one of the big drivers of cost is going to be the sortation piece. I will note that another big driver of cost, which we should not ignore right now is freight. The transport of stuff from one position to the other, just driving tons and tons of trash around obviously costs a lot. But if we’re looking at that sorting problem, it is a big problem right now.

And it’s a problem that could use a lot of help. so what we do at Glacier is we make technology that helps that sortation process. if you look inside one of these MRFs, you’re gonna see essentially two types of sorting processes. One, I like to consider high-quantity, low-quality sortation.

So they use machines using spinning discs or magnets or something kind of heavy metal, old school technology that does a really good job processing a lot of stuff. and it can sort, for instance, big pieces of cardboard apart from small plastic bottles. [00:20:00] but I said low quality because it’d either end the like high-density to low-density piece.

For instance, you’re gonna see a lot of stuff that shouldn’t have been there. For instance, if you’re trying to sort paper apart from everything else, using a density sorter, you’re gonna end up with a lot of plastic films, and plastic bags in the paper stream. So you have this quality that’s, that’s a problem.

Um, and similar on the downside on the flip side you have,

Abate: This is without any additional computers, or processing? This is a completely mechanical approach?

Areeb Malik: Yeah, exactly. These are completely mechanical approaches to sorting. they are getting slightly smarter, but they’re, it’s, it’s pretty kind of blind as it stands today.

And then on the flip side, there are the high-quality, low-quantity sorting techniques. What that technique is is, is people whose job is to stand above conveyor belts and sort through what these machines did not get right. and also to sort through stuff, to prevent the machines from getting hurt in the first place. You have a lot of people inside these facilities doing, and, and I think they do a great job, right?

Like a person’s eyes and a person’s hands are fantastic. in terms of doing this type of task, but there’s only so much a, a person can do. the drawback there, of course, is that people are also very expensive. and so if you’re trying to drive the costs of this end output down people’s a really tricky option to pull on.

Uh, the other tricky thing about that is that people are really hard to hire for these, these roles. as you might imagine, it’s not necessarily a very desirable job. we have spoken to several facilities across the country and all of them echo the same complaint. It’s just like yeah my people are like, it’s really hard to get people to do this job.

Uh, we talked to a facility up in Michigan and they have a station or a facility that has had 30 sorting stations available in the last two years. The most people they had in the facility on a given day was fourteen. So you’re looking at a lot of facilities that just like cannot possibly fill all the spots they need.

And because you have the high quantity, low-quality machines running and you don’t have enough high-quality, low-quantity people doing the QC you end up producing low-quality outputs. When you produce low-quality output,, you sell it for less. The buyer has to do additional sorting. The end output that they produce is lower quality.

Um, and so kind of everything is degraded by this, this sorting problem not being done efficiently and high quality enough.

So let me sit back up to “what does Glacier do?” because that was the original question. *What we do is we build robots.*

* W**e use computer vision to look down at a conveyor belt inside one of these facilities and identify every single thing that’s coming down that belt**.*

Um, we can classify. Basically, you know, recyclers think of commodities in a very specific way. And so we just know that code and say that’s P E T that’s HCP that’s an aluminum can, whatever it might be. we can use computer vision to do that. And then we pass that information to a robotic sorting system that can then separate things on a belt, however, it needs.

Um, and so it’s very dynamic. It’s very adaptable. It’s very easy to install inside these facilities. and the idea is basically here’s like a, an option that you as a facility can use. If you’re, if you can’t even hire these roles, here’s a way that you can actually drive the cost of your operations down and we hope also the quality of your operations up.

Abate: Yeah. So if you, if you were to help us visualize what this system looks like, is it a conveyor belt with a camera on top and two robot arms that are, you know, like what exactly does it look like?

Areeb Malik: Yeah, it’s a, it’s a good question. So I think the first thing you wanna visualize is these facilities these facilities, they vary place to place, but generally speaking, you have several layers, several floors, and on each of those floors, there are conveyor belts. There are these big machines I talked about earlier.

They do some sort of sorting a lot of loud noise. everything’s kind of moving everywhere at once. Very chaotic. And I think very typically you don’t have a lot of space, right? These, these places are pretty tight. Like you walk around belts and there are people saying that there are shoots over here to drop stuff in. Not a ton of space to add new equipment in which is one of the tricks of working in these facilities.

Right. You can’t just say, Hey, like here’s a 20 by 20 square… or 20 foot by 20-foot machine that can come and like, do your sorting. Cause like, they’re like, “Well, how am I gonna fit that in?”

So imagine that space, you have some conveyor belt somewhere that’s moving stuff along. what we do is we basically build a machine that straddles that conveyor belt.

Um, so without having do any retrofits or major retrofits to your facility. We just pop a machine right on top of your conveyor belt. there’s a camera that sits a little bit upstream of the robots and looks directly down at it. Uses computer vision to, like I said, classify everything coming through and the robot itself.

Uh, yeah, you actually, you got the number, right. We do use two arms. but the idea is basically a series of robotic arms that can basically respond to computer vision. pick up items off the conveyor belt. Move ’em to some location, whether that’s a bin that they’re collecting in or a shoot that takes it to another belt, or a separate belt altogether, and drive it to wherever they need to drop this item.

[00:25:00] We drop the item there.

Abate: Yeah. So, you know, and just thinking about this from a robotics point of view, there are definitely a lot of challenges with trash. they’re random shapes and sizes and maybe a little bit wet, potentially, like who knows? Right. so how, how are you able to overcome the challenge of being able to pick up individual pieces and move them to individual containers or other belts?

Areeb Malik: yeah, yeah. It’s a really good question. And that is, that is one of the hardest problems that we have to deal with, cuz you’re absolutely right. I’ll get to the wet piece later, but like, if you just imagine a plastic bottle, like just one plastic bottle, same brand, same size, like there are a thousand ways that thing comes down the belt.

Um, it can be crumpled in different ways. It can be filled with water or not. It can be obfuscated by some other material. And so you have this really hard challenge. The roboticists on here will understand, pick and place is one of the holy grail challenges of robotics today.

Um, and there are PhD programs dedicated to designing a gripper that can do as well as the human hand. our challenge of course, is that we want to make something that’s high ROI for these facilities. and is, is maintainable, right? If I build you a $500,000, you know, robotic arm that can only do you know, three items a minute, or, you know, 10 items a minute.

That’s not good enough for these facilities. so how do we overcome it? It’s, it’s a, it’s a development. I think it’s definitely like a, a point of focus for our company. And one of the things that our engineers are working really hard on thinking about, right? Like how do you design end effectors that balance high throughputs with the ability to successfully pick as much as possible?

And we have a metric internally that we call our pick rates and getting that thing close to a hundred is the dream. But if you look at just technologies out there, even the best robots, aren’t hitting a hundred percent pick rate for items. And that’s something that we have to just deal with to the wetness piece, to the dirtiness piece.

Yeah. Like that just makes it harder. Right? If you have a dirty bottle coming through. You can probably pick it up the first time, but that’s going to clog up a lot of your components and that makes the reliability of your machine more difficult. So how do you pick an item that is dirty without causing issue to your machine, or rather, how do you design a machine?

That’s easy to clean and easy to update parts when it gets too dirty. that can sense when it’s not picking well and alert the maintenance team. Hey, like, can you come clean my, you know, like gears or hose or whatever that needs to be cleaned because it’s gotten gunky. these are all challenges that I think are very accessible within robotics.

Um, but nobody’s ever thought about these problems in the context of, of trash. And so the problems that we get to solve you know, on our own over here, which is pretty exciting. but they’re, they’re I think, I think they’re, they’re, they’re doable. They’re really achievable challenges, which is, I think the exciting thing about being here.

It doesn’t feel like it’s unachievable in our current context.

Abate: Yeah. And one benefit is that, although you wanna obviously maximize the amount of items that you pick up and drop. You’re really, you’re really maximizing for getting good ROI and being able to replicate what a person can do, but do it 24/ 7 as opposed to eight hours a day and, and do it at scale. So even if you say miss a water bottle, that just goes into your percentages and you factor that out, and maybe it’s not the end of the world because a person could also just miss something or, worse off, not even be working because it’s 6:00 PM

Areeb Malik: Definitely, And that’s one of the nice things about being in our space. You know, I think about my colleagues in the self-driving car space and, and if you’re doing self-driving cars, there are lot of these other robots. Like you are not allowed to make mistakes, that robot cannot fail.

Um, and you spend a lot of efforts kind of working on the 0.99999. Right. whereas for us, like, yeah, it’s fine. If we. 0.98 like it’s, it’s like, it’s a very high ROI machine if we get to that point. So it’s totally okay if we make these mistakes. And what that allows us to do is it allows us to focus on the bigger picture.

We don’t get caught up in the details and we can actually do, is say, okay, cool. Like this machine works 98% efficiently. Like, what else, what else does does recycling need? What other robotics do we need to address your, how do we need to evolve this machine to handle this install location or this commodity type or, or whatever it might be.

Abate: Yeah. And so also to back it up to the software side, there’s also object detection and classification putting this thing into groups. is that, or are you able to leverage some open source resources out there? Is this already a solve problem or is this another large challenge of your company?

Areeb Malik: Yeah, it’s, it’s a good question. I think generally we like to build things in house here. One of the things that is unique about what we’re building is it is a robot custom design for the recycling industry. [00:30:00] a lot of times what you can do is you can go take an off the shelf robotic system and just plop it into place and say, okay, like here’s a, here’s a Delta robot.

Um, I’m going to put it into this industry and I’m gonna attach some, you know, basic processing software, but everything else kind of goes pre-written. the problem with that is that those off the shelf, robots, those kind of custom generic, or sorry, those generic robots they tend not to hit the ROI marks that these facilities are looking for.

And so we, our whole premise is if we build a custom robot, right? It’s like from the ground up, this is designed to be optimal for these recycling facilities. Will they like it more and the feedback we got really early on and we continue to get as yes, like that is, that is a huge win for them. Cuz it’s simpler.

It’s easier to maintain. It’s easier to install all of this stuff when it comes to the software. That also means we have to write it ourselves. you can do some stuff off the shelf. but typically speaking like. If you, if you want to customize it and make it very specific to a recycling facility.

Yeah. It does involve a lot of custom software to be written as well. it’s kind of fun. you get to solve challenges that are really unique. Like I said, a lot of people have solved problems for pick and place, but not a lot of people have thought about the, the specific issues that come inside the recycling space.

Like we were talking about, what if a bottle is dirty? Like what do you do about it? and so, yeah, custom software as well, custom, it’s all, it’s all custom. And the idea is that if we can customize the machine. we can build a machine that these facilities are actually gonna benefit from. That’s ultimately our goal is to help them do their recycling and their sortation better.

Abate: Mm. So, and then just to get a picture of the makeup of your team at glacier, are you guys something like 50% software engineers, 50% also some hardware and hardcore robotics people.

Areeb Malik: Yeah, that sounds about right. I think that’s that that’s a pretty good split. I would, I would say like as founder, right? I was a software engineer before this. and if I look at What I’m building today versus what I used to build. you need so many disciplines of engineering to make this work, right.

We need CV engineers, generalist software engineers, firmware engineers, electrical engineers, the whole, the whole slew of engineering. And so, yeah, we have a pretty broad spectrum. We’re a small company. So everyone gets kind of like stretch their wings and kind of explore into like different realms. but yeah, I would say on the whole we’re, we’re kind of divided along the, the software hardware line, like about 50%.

Abate: Yeah. Yeah. And so I also know your team just very recently came out of stealth mode and you’re definitely still very much a startup. Have you guys already started or planning on soon doing a pilot at a sorting facility and being able to test it out in real time.

Areeb Malik: Yeah. So I can’t name too many names, but we have done two pilots already and we have a third one that’s ongoing currently. and so we have put these machines out there. It’s great to see them run. It’s always so fun. Just turn on our, our, you know, nanny cams and be like, oh look like there’s our robots sorting through recyclables.

How cool. and so, yeah, we, we, we have a couple pilots kind of underway we’re at a point as a company where the objective right now is basically refine this machine. So it’s. It’s it’s perfect. Right. We, we know that these customers, like, they don’t understand why building this stuff is hard. They just want a machine that works.

Um, and so our objective is to get the machine to work so well, they don’t even think about it, right. As a customer, you just wanna be like, okay, cool. I’ll buy this robot. You install it. And then you never think about it again. And we wanna get to that point. So that’s where a lot of the focus is right now.

Um, in terms of, in terms of what we’re developing on the engineering side.

Abate: So you’ve now you’ve installed this system in a couple different places. Are there any key learnings that you’ve gathered from seeing this actually in place inside a, a customer’s sorting facility and they’re actually getting their hands on it and using it. And now they’re sitting there comparing this to what a person would be doing two months ago.

Areeb Malik: Yeah, it’s a, a really good question, I think from their perspective, right? Like of course they are going to compare this to a person and robots and people are they’re, you know, analogs, but they operate very differently from each other. as I kind of mentioned earlier, like a person. They’re our hands are so good.

I’ve seen people pull a can out of a plastic cup before and they take the, can they put into the can pile, take the plastic cup, they put in the… Like robots never gonna be able to do that. At least not, not in like the next 10 years. Right? That’s not something we’re trying to achieve. And so getting them to understand this difference, like this is not a defacto replacement for your people.

It is a sorting technology that helps you achieve better quality at lower costs. That’s what we’re promising. if you look at the learnings we’ve, we’ve made so far, I would say. These install locations, they are wildly different, right? You can’t just make one robot that does one thing. and put it into a, done a dozen different locations, cuz each install… the size of the belt, the, the amount of density on the belt, the commodities you’re picking, where are you taking it to?

All this stuff, changes, install to install. and so one of the, the great things about robotics is you can, [00:35:00] you can program the software to be very dynamic, right? And you can say this, this hardware. It looks very similar to that hardware, but these robots are actually very different on the inside because of how they’re configured because of how they think based on the stuff that’s coming down the line.

So being able to build technology that is dynamic enough to handle a wide variety of situations I think is, is key to our success.

Abate: Yeah. And are there other robotics players in the recyclable industry?

Areeb Malik: Yeah, there are a few that are maybe a couple of years older than us at this point. If you look at the market as a whole, it’s, it’s definitely very nascent. A lot of the facilities we talk to have heard about these other recyclable sorry, these recycling robots. and a lot of them are, are waiting for the time to be right or they’re waiting for their next retrofit to fit them in.

Um, but if you look at kind of the market as a whole. Most facilities have heard about them, but have not installed our robots. if you look kind of beyond the robots of like, what else is gonna happen to these facilities? You step inside one of these facilities and you’re gonna be quick to see like, oh, there are like 15 optimizations we could do.

Um, and so, you know, robots first you know, there’s comp competitors, but there’s a lot of space. And then beyond the robots, there’s a ton of opportunity to just build out technology that helps these facilities and helps the recycling industry operate more efficiently.

Abate: mm-hmm yeah. So what’s next at glacier?

Areeb Malik: What’s next at Glacier, will we take over the recycling industry and fix the world?

Is, is the end?? That’s next.

what are we doing? So I kind of mentioned we have these two prototypes the objective technically is to get the machines, to be as good as possible. and then start getting them into more you know, facilities across the country at that point, right? Like once we get this machine down and we can actually just like, press the rocket ship button on it get these things everywhere.

My vision is that one arm of glacier is these robots. And the question becomes, what else? What else can glacier do with our status in the industry, with the technology we have with the people we have on our team? How else can we positively impact how we as a society handle our waste?

Um, and there’s a lot of opportunity there. There are a lot of ideas. There are a lot of things we need to vet out in the markets.

Um, but as I mentioned before, this industry is really big and it could definitely use a little bit of love from the tech industry. So we’re hoping to kind of get the robot up and running. Let that kind of be an arm of the business and then focus on what comes next.

How do we next improve our you know, ability to, to take stuff away from the landfills and get it back into the economy? is it compost? Is it data? There’s a whole, whole slew of things we could be going into. we, at this company, never call ourselves a robotics company. We consider ourselves a recycling technology company.

Um, and so what comes next is, is whatever we can do best to help this industry align.

Abate: mm-hmm. awesome. Awesome. What’s a good way for people to follow up on glacier and your progress?

Areeb Malik: Yeah. So we’re, we’re working on our SEO, but if you if you go Google search us, you’ll find our LinkedIn, our Twitter, all that sort of stuff, which we’re obviously you know, posting on regularly with updates. we also have a website which is endwaste.io.

Um, and so you can go over there to check out a little demo of the machine and kind of see what it looks like, cuz I can’t explain it well enough with my words. but you know, take a look at what it looks like, see our open roles, all that sort of stuff.

Abate: Thank you. Thank you for speaking with us.

Areeb Malik: Of course it was, it was great and it’s like I said, it’s, it’s very exciting to be sharing this information with people because I think everyone needs to know a little bit more about how recycling works.

transcript

tags: Algorithm Controls, c-Environment-Agriculture, cx-Environment-Agriculture, Industrial Automation, podcast, Robotics technology, startup