I think sustainability and security are totally wrapped up together. It was the drought in Syria that led to a crop failure that led to a humanitarian crisis and refugees spreading all over that region and Europe. When Ukraine goes into war, we lose access to food, and food security goes out the window. I think it happens that those things are interrelated and can be solved together with these sorts of smart technical systems.
I am pleased to be joined today by Will Marshall, who is the co-founder and CEO of Planet Labs, which provides governments, businesses, and research institutions with the highest frequency satellite data, images, and analytics available today, including the photographing of the entire planet at least once per day. He and I discussed the current projects at Planet Labs that are working to improve the health of our earth's ecological systems from tracking illegal deforestation in the Amazon to identifying methane leaks that might otherwise go undetected. Coming from fracking operations. Prior to co-founding planet, will was a scientist at NASA and the University Space Research Association, where he was a systems engineer on lunar orbiter missions and was the technical lead on research projects in space debris remediation. Alongside past and future projects will Marshall shares the mission behind Planet Labs work, which is to image all of Earth's landmass every day and make global change visible, accessible, and actionable. In my opinion, the data coming from Planet Labs is one of the most positive use cases of artificial intelligence, with the potential to enable an unimaginable number of other projects and initiatives that are in service of life. Additionally, there are a few images from Planet Labs referenced throughout this episode. If you're listening on an audio only platform, you can find these images under the show notes on the Great Simplification website. Before we begin, if you enjoy this podcast, one of the biggest ways you can support us is by subscribing to it on your favorite platform and sharing this episode with someone who might also enjoy it. We believe in making this content free and accessible to as many people on Earth as possible, so we appreciate your support. With that, please welcome Will Marshall.
Will Marshall. Great to see you.
We are here at Planet Labs, headquarters where you just gave me an incredible tour. This is, a lot of impressive technology for a Wisconsin boy. So you are, a physicist and, you make satellites, send them to space to look at earth. How did you get started in this? What was the led you from physics to creating Planet Labs?
Ever since I was a kid, I was a really interested in astronomy and I even built my own telescope when I was a teenager. And what you learn when you look up, and I continued to do this through working at nasa, which we'll get into in our founding story of your planet, is that we've still not found any life there is. It's mostly rocks and dead matter out there, and that only makes you more and more amazed at what we find here. So if you'd have told my, 7-year-old self that I'd be building telescopes looking down, from space, I'd have gone, well, that's the wrong way round. But apart from that, it would've made sense. The in incredible, vastness of the universe contrasted with the fact that we still haven't found life anywhere. So it's either very rare or we're the only. Life, place of life in the universe. It gives an almost cosmological duty to take care of life on the earth. And so that's what was the initial motivation, through, my career.
You ever seen a, little short movie called The Overview Effect? Has that happened to you where when you started to work and you started to see the images coming in from your technology, did it actually intensify the emotion that you just described?
Yeah, and, definitely the challenges that we see, which is, you know, devastating destruction of the rainforest and the coral reef systems dying and things like that. For sure. That's, that, that's hard. And, it only make MO motivates you even more. It also motivates you even more when you see how they can help people to fix the system. And we see that in abundance in some of our partnerships around the world where people, if you give them real time information, can help fix some of those things.
So give us an overview of the projects. how does planet access this real-time data? what, are the main arcs of, your work and, your hopes, and success?
Yeah. I mean, at a top level, what Planet has done is we've launched a few hundred satellites that image the entire earth every day. Now, why do we do that? We did that because you need real time information to take care of the planet. it's actually an engineering principle. You learn quite early on, in, if you go into technology that you can't fix something if you can't, if you don't have data on it, on the relevant time scale. So if you spend a spacecraft into space and you're in a tumble. If you measure it once an hour, but it's spinning once a minute, then you can't stop it. Or if you're in a spacecraft and the oxygen rates are going down, or the CO2 is going up, if you're not measuring it on the relevant timescale, it's gonna be too late before you fix it. But there's no such system for the earth. We are a collective ecosystem of life on the earth, but it wasn't, we didn't have the relevant, data on the relevant timescale. So the mission we set ourselves out was to image the entire earth, land mass once per day to get into the human time timescale of evolution on the planet. You know, if there's deforestation in the Amazon and takes place over days and weeks, there's no point measuring it once a year. And that's roughly what we were doing before planet came along. And we basically took that to a daily scan so that we can get inside that decision making loop and help make PE enable people to make smarter decisions.
How many satellites are there now?
We have about 200. we have the, two fleets of satellites, mainly, one that does this daily scan and then a higher resolution fleet that can zoom in on any particular location up to 10 times per day.
And like when you say zoom in, like how, what's the resolution?
The present best resolution we have is about 50 centimeters. but we're just launching a system that's gonna get even better than that down to 30 centimeters.
Can you give an example of what that would be like?
Yeah, I mean, that's like seeing, you know, you can easily see vehicles and the, and, even identify the kind of vehicle from, and we're 500 kilometers away, so that's like taking a picture from here in San Francisco of Los Angeles and seeing, vehicles and things like this.
So the satellites are 500 kilometers up in the sky. And you're doing real time, minute or every hour or continuous once per day, the whole earth, and then up to 10 times per day for any particular location where we are getting, where we have particular interest. And some of those have been going for months, years now.
We've been doing this total daily scan now for about seven years.
So you have compiled an enormous amount of like time lapse images of everywhere on earth.
Yeah, on average we have about 3000 images for every point on the Earth's land mass.
Do you ever, like on a Friday night, this is a Friday, do you ever like grab a beer and just look at some place in the Central Africa or in Indonesia, just out of curiosity, or are you overloaded with so much all the time?
Yeah, I sometimes look at the imagery like that and we have a slack channel called Interesting picks that always flashes up crazy things that either people or the machine learning is detecting. And then we're like, what is that? So that's the channel I tend to tune into when I, have a free moment. And it's incredible. And what you learn is the earth is constantly changing. So when you think about maps, you think about whether it's a classic old map or Google maps online and you go to the satellite layer, you think of it as a static thing. But actually when you look at our images, you realize it's an evolving thing. and that this dynamism of the earth, it really speaks very quickly if you start looking at a time-lapse.
So has it, accentuated your love for earth or also the tragedy that's unfolding, or a little bit of both?
Yes, exactly. Both. I mean, as I said, we've almost, I feel, have a cosmological duty to the protection of life, given the rareness, of, life in the universe. And I see planet's role as helping to hold up a mirror so that we can understand those changes, in the relevant timescale to make smarter decisions.
So if your company and your work, are wildly successful, what could you envision 2, 3, 4, 5 years from now? that's outcropping of, all this?
Well, look, I hope that we can, make a serious positive impact in two main areas. I think about. Sustainability and security and the sustainability realm. Our hope is that people can take this data and integrate it into smarter decision making, sustainable agriculture, stopping deforestation, helping us to manage fisheries sustainably and so on and in, security, how can we shed light that helps drive accountability, and reduce warfare. you know, basically in bo underpinning both of those areas is data can help drive accountability. It's basically we're bringing a big transparency maneuver over the planet and with that information, it comes the power to make smarter decisions. So, you know, the deforestation example is kind of obvious. You know, you get the data of the person that's doing the deforestation in the relevant timescale, then you can go and stop it. Hap happening.
Oh, it's almost like the candid camera, environmental police, of sorts.
Yeah. So lemme give you this specific example in Brazil. We monitor all of the Brazilian Amazon. Every week we do a search for new roads across all of the 8 million square kilometers of the Brazilian Amazon. That's almost the size of the United States. And when we find a new road, that is normally a sign of either illegal deforestation, illegal narcotics, a, new illegal mine, the federal police. 3000 interventions using that data last year, so almost 10 per day, where they send people to those locations in helicopters and what have you to stop that action, it confiscated $3 billion worth of illegal equipment, and most importantly, significantly reduced deforestation rates.
So you send a dossier of the recent images to your colleagues in Brazil, for example, and the AI on top that gives them the alerts of where the problems are happening because they can't very well look.
I mean, so we download about 4 million images, 47 megabyte images every day. So there's no way a human can scroll through those. A single human is not gonna look through all this imagery, but we can have, use AI as interns if you like, have thousands and thousands of ais that look across, AI interns looking across all of this data and finding the things that the humans need to look at.
So that's. A check mark on one positive thing that AI might be able to do for us, and that's in the sort of sustainability realm and in the security realm. Take Ukraine. we can bring accountability. what's going on there? When, people found mass graves, our data could show who did what when. because we can go backwards in time and do the forensic digital forensics. when there was claims that they, Russia is not bombing civilian targets. We could show on the front page of the New York Times the next day, here are the five hospitals and schools that were bombed, et cetera, et cetera. I think that becomes in time a deterrent from doing those sort of things, because. And, actually slightly bigger picture, back through history. we've seen most wars have begun when there was mis lack of information or people took a lack of information and, had to take the worst case scenario, and then they do something risky. And then you fall into a war. It's when you have informa. It's when you have a lack of information. You're more likely to fall into war. When you have information about where each other's military capabilities are, you are less likely to fall into a war. War often happens through mistakes, not deliberate. Action.
Wouldn't governments have their own satellites doing this stuff?
They do. but they don't have as many,
you have more satellites than governments?
Yes. in earth imaging we've got the largest imaging fleet of any actor. quite a bit bigger than anyone else's. and, but they specialize in typically is super high resolution. So I said 30 centimeters, well maybe they're three centimeters. You know, they're much higher resolution. But they are looking through a soda straw at a tiny area of the earth, and they spend a huge amount per satellite. Like the us government sometimes spends three or $4 billion on a single satellite. So these are huge school bus size satellites with massive telescopes looking down so they can get exquisite resolution. But what that trades is coverage. So what they can do is zoom in on a particular location and know exactly what's going on with better precision than we'll ever do, probably. But what they can't do is scan the whole earth and find new changes with that sort of system. So we're like the peripheral vision, for humanity. And sometimes these governments,
is Moore's Law apply to the, 30 centimeters and the one centimeter? Is it headed in that direction? the resolution like three years from now, they'll be able to detect from 500 kilometers, a human face. Is that gonna be possible?
There is a. Improvement curve, but it's not anything like Moore's Law on the optics. Okay. So where there is improvement is in the electronics. So typically when we're at nasa, a typical mission costs a billion dollars, sometimes a few hundred million, sometimes a few billion, but all of magnitude a billion dollars per satellite. And the main trick that we did was focus on leveraging consumer grade electronics. Like the phone in your pocket and the stuff that's developed there. And, that tends to be miniaturized and it's pretty robust. And we leverage that technology to build our satellites. And that is following Moore's Law. And what that leads to is more processing power in orbit, more faster radios, better sensors. so the data side of it and the compute side of it is improving very fast along the lines of Moore's Law. In fact, we have a mantra where we're trying to strap space to Moore's law so that every time we launch a satellite, it's. It's, gonna produce three times more data or 10 times more data than the previous version. But the one place that doesn't apply is the optics because no one's figured out a way to improve that in a, in any sort of digital way. That makes sense. So an optic is an optic, and so you just have to get to a bigger telescope or fly lower to get to a higher resolution. And doing that a 10 x on that is not easy.
So we met a couple months ago, and one of the things you told me is that, you believe that biodiversity and eco side are more urgent, lower cost, high impact environmental areas than climate change, and that you are playing a role in that with your, earth imaging. can you explain what you meant by that?
Absolutely. I mean, look, the planet obviously faces a number of sustainability challenges. Climate and global warming, biodiversity loss and eco side. Also pollution in rivers, various other pieces. I like the Planetary Boundaries Framework is a, as a quick snapshot of how we're doing across these different measures. I would say the. Climate one has grabbed the headlines, and I'm not sure that's smart. I think we're much further along in the eco side, and I think it's much more profound. I think also from a marketing standpoint, if I can think about it like that, it was an unwise choice of focus, for the sustainability and environmental community because climate is distant and it's hard to, understand. and whereas losing life, losing nature around, everyone cares about that. Ask a child, they care about that, they care about, other birds. Are there less birds in the forest nearby? Are there less animals? And, there are, we've, dramatically wiped out life on the earth. And it's, the other reason I think of it is that it's primarily a. It's a problem that we've already created, whereas a lot of the climate stuff is in our future. This is already, we've already wiped out 70% of life on the earth, roughly speaking according to IPBS reports and, well, wildlife Foundation, se 70% of average population size on average population size, which is a massive yeah. Ecosystem. And the question is only when you start whittling it down to that, at what point do you lose the ecosystem viability. And we dunno.
So I imagine you have hundreds, thousands, tens of thousands of time lapse loops even in the last seven years that are colorful that almost probably look like the proliferation of a cancer cell, time-lapse. Is there a place that you are these are available to the public to see things or how do you highlight this?
It's a great question. I mean, we, do share time lapses with media when they ask for them and NGOs. We have a huge program with universities to provide our data, to universities to help, to, there's, I dunno, 30,000 university researchers that are using our data, for example, to try to get that data out there, to understand this more as well as bring great awareness to the challenges.
So gimme an example. Like what are we looking at here?
Well, here's an example of, deforestation occurring over time. people are probably familiar with this sort of time-lapse. You see, the, inroads first of, the loggers before they even take down trees. And this is how we can detect the deforestation before they even take down trees. But the logging roads go in first, and then you can see the deforestation starts happening in, around those roads. here's another example where we see, some, ocean, some coral reef bleaching. that's probably because of increased, temperature locally in that, area of the ocean. And this is a final example of a new satellite we just launched, which is a hyperspectral satellite. So what you're seeing here is a set of. Methane plumes of gas, of methane rising up from these different, mainly gas, oil and gas facilities. And the great thing here is we can actually pinpoint the location of these gas emissions so that operators of those can go and fix them. And nobody wants the gas leak. the environment certainly doesn't want it, but nor do the operators, because it's just a waste. And so by helping to point them out, we can help fix that.
So what are the filters that are being used? Because methane is invisible, correct? It's a, it's a colorless, gas, but.
It absorbs in a particular spectrum. So this is a hyperspectral instrument. So instead of having three colors like the eye, we have 420 colors, across a slightly wider spectrum. I think it's from about 400 nanometers to about 2.5, microns. and they, each different element has slightly different features, in different, spectral ranges. And in, in methane, in this particular case, there's a deep absorption band, a particular wavelength. And so when we detect that, depending on the depth of that absorption band, we can estimate the amount of methane in that pixel.
So. For non climate reasons, there would be a business reason for oil and gas companies to say, Hey, we didn't know we had this leak here.
This particular one is, was, is one of my favorite examples. the, this is, we, released the first 300, the last COP meeting and we released them to the public just to showcase this technology. And within a week or two, this operator, as you can see, has turned off this leak and that one leak is 7,000 kilograms per hour of methane. And I did a rough calculation that's about the same as 10,000 cars going full speed on the motorway. Or if you take into account the duty cycle, that's an equivalent of about 300,000 cars being taken off the road. That one leak. That one leak. So quick maths, if you did about a thousand of those per quarter, which we think there's one satellite could find a thousand leaks like that per quarter around the world. You could, it's almost the equivalent of about taking all the US cars off the road.
And so that's, you're planning on doing that?
Absolutely. Yeah. It's, the satellite hasn't quite finished. It's, commissioning, but as soon as the operation.
But this isn't, this isn't like putting a tax on carbon or anything. This is part of business, doing business. You're getting paid by oil and gas companies for this information.
Yeah. so we've got a number of partnerships. it's useful for the NGOs that are trying to track this and the universities that are trying to research it, but most practically, it's useful for the oil and gas companies and waste facilities and ag companies that are the dominant sources of methane emissions to get, have that, to know where their leaks are, to go and fix them. And then for the regulators. it is a, cheap and efficient way of regulating this, without having to send people everywhere with these little sensors, they can just look in our imagery.
So do you have either an AI intern or a real human that gives you a weekly Monday morning briefing? Like, this is like stuff we'd never seen before and you need to look at this boss?
No, but I should. Good idea. I'm gonna work on that. I mean, I would be a kid in a candy store because I love the earth so much, and the fractals that you can see from space and all the colors. We do sometimes print out our beautiful images and take a look at it and, but yeah, we, should do more of that.
it's so beautiful. We definitely take pride in that here at planet. The beauty of the imagery that comes down and also taking pride in the satellites. We have art on the side of the satellites, art on the side of the ground stations and other things because. We want. Yeah. We just want our, work to be beautiful as well as impactful.
I mean, this is a naive question, but there should be like beautiful glossy coffee table books and an art exhibit with only planet images. Like do Are you doing that?
Yeah. I've got a book for you before we leave.
Really? Okay. Because this gets to the heart of it. Like Carrie Stover, works with, planetary Boundaries and Johan Strom, yep. And, they're doing this, annual health check.
How does your work overlap with something like that? They're trying to do an annual health check of the planet on the different boundaries?
Well, certainly that's a really important initiative and, we are helping, as best we can with that initiative. some of those parameters we can measure, some we can't, but where we can, we're certainly, willing to help provide our data.
So one of the questions I asked her, when we did a podcast was. Is it a lack of information? Is it a lack of seeing these time-lapse photographs of what's happening to the earth? Oh my gosh. I love the forest. I love the animals. I love the biodiversity. This is happening, but we kind of know enough and we haven't really changed our behaviors. So this, what you're doing is definitive, undeniable, like real-time data. So maybe that will change at a higher scales, but what are your opinions on, that whole trajectory?
Well, I think it's a great question because it's, but it's a sociological one, and I'm not a sociological expert. But what I would say it's a mixture of, I think this data can be because it's more real time can be much more practical when you're told, well, roughly speaking, we have this ecological problem, not here's the incursion onto the park, near your house yesterday. It's a very different contemplation. Right. One, you're like throwing your hands up. You're not, dunno what to do. And the other you're like, okay, well there's something I can do here right now. So it's actionable and specific as opposed to general. The second thing is that just whilst it is true, so I think it can ha help, but the bigger thing is sociological changes. Right. We've got to, behavior change tends to happen when you look back where only when humanity decides something is important. Yes. When we had covid we decided it was important and we changed our behavior. We didn't shake hands, we didn't go to the office, we flew with a mask on we, these other changes. We as a species basically haven't yet decided that the environment is an important issue. and it's partly, I think because of timescale. The Covid, pandemic onset was four months here. The onset of that biodiversity loss we were talking about earlier has been 40 years. Those couple of orders of magnitude make it harder for humans to go with. Was it that grand said there were more birds here, or, you know, it's like, what does that mean? You know, it's, hard to when you are, when your timescale is slower, but I, think people are starting younger generations see it more. They feel it more, they understand the urgency. They know we're in a predicament. I think that a bit like, the other, the good thing I see is that, again, when humans decide something is important, they will do something about it. We did with Covid. I actually think our changes of behavior to help the planet are less than the changes of behavior we did for Covid.
