Transcript from 501: Inside the Armpit of a Giraffe with Meredith Palmer, Akiba, Christopher White, and Elecia White.

Welcome to Embedded. I am Elecia White, alongside Christopher White. Our guests this week are animal- Wait, no. Our topic this week is animals. Our guests are Meredith Palmer and Akiba.

CW (00:00:19):

Hi. Thank you both for being on the show.

MP (00:00:22):

Not that we are not animals. <laugh>

A (00:00:27):

<laugh> I guess, technically, yeah, we are all- Everyone is an animal.

MP (00:00:30):

Everyone is an animal. It is fine. <laugh>

EW (00:00:33):

Meredith, Dr. Meredith Palmer, could you tell us about yourself, as if we, I do not know, met at the Monterey Bay Aquarium Research Institute, in a random talk you gave.

CW (00:00:45):

<laugh>

MP (00:00:46):

Oh, my goodness. If only I did things that cool. Yeah. My name is Dr. Meredith Palmer. Meredith Palmer, just like the character from "The Office." I am a wildlife ecologist, with over a decade of experience in the field, and advancing technological solutions to overcome conservation challenges.

(00:01:05):

I started out, again, over a decade ago at this point, doing ecology and conservation of lions in East Africa. I have since supported carnivore conservation and rewilding programs, across Africa and around the world.

(00:01:21):

I am currently a conservation scientist at Yale University, where I lead and support projects, integrating in situ sensor networks. So, camera traps, acoustic sensors, eDNA. All of those great tools with AI driven data processing. Again, to drive impact in the conservation space and in global policy.

(00:01:44):

Essentially I am interested in using emerging technologies, both to gather "big data" that enables us to better understand biodiversity at scale, while also working with new types of sensors, to develop data-driven strategies for protecting and restoring ecological systems.

EW (00:02:03):

Akiba, what is it that you do?

A (00:02:07):

Yeah, so I guess my background is- Actually, I was originally a professional hip-hop and break dancer. That did not work out for me. I went back to school, and ended up as an FPGA and ASIC designer, way back when. Since then got into firmware and system level hardware design.

(00:02:27):

I also discovered that I had a passion for wireless sensor networks, and designing technology for harsh, resource constrained, outdoor environments. That took me into developing technology for international development. And now, for the past ten years, developing technology for conservation.

MP (00:02:55):

Honestly, the best origin story of anyone working in the field. <laugh>

EW (00:02:59):

<laugh>

CW (00:02:59):

<laugh>

MP (00:02:59):

I do not think I know anyone coming from the hip-hop scene. It gives me a lot of street cred to say I know someone who did.

CW (00:03:06):

Yeah, I really just want to say, "How did that happen?" But we probably should not spend the whole show about that.

A (00:03:13):

I think you should ask Meredith about roller derby. <laugh>

MP (00:03:19):

Oh. Guys. That would definitely be, we would be on the phone for forever. <laugh>

EW (00:03:24):

Okay. We are going to do lightning round. We are going to try to make this fast, because we ended up with listener questions, as well as our own, as well as all of the historical ones the listeners know we are going to ask. So let me get the first one out of the way. Have you ever pet a penguin?

MP (00:03:41):

Oh! I have seen plenty of penguins in the wild. I do a lot of work in Southern Africa, and we do get African penguins. They are called "jackass penguins."

CW (00:03:51):

<laugh>

MP (00:03:51):

I do not know if you have to bleep that out or not. But it is because the sound they make sounds like a donkey, not because they are horrible penguins. But I do not think I have pet one.

A (00:04:04):

Yeah, we are part of the US Arctic Research, I guess, organization.

MP (00:04:10):

Shut up. Are you?

A (00:04:11):

And then- Yeah. ARCUS. We missed out on the community call last night. It is like 2:00 AM Australia time.

MP (00:04:17):

Ah. Amazing.

A (00:04:19):

Yeah, so we have a project for polar bear tracking, but we have not actually interfaced with penguins, unfortunately. Even the animals that we work with, we never get a chance to pet them. <laugh>

CW (00:04:30):

That is a shame.

EW (00:04:32):

Not even the polar bears?

A (00:04:34):

You are not-

CW (00:04:35):

I do not really think you should do that.

A (00:04:38):

Well, there was a- One lady from a zoo offered to- Was interested in being involved in the project. She offered to train the polar bear to present its ear to Jacinta, to mount the tracking tag. And we are like, "Oh. That would be awesome." But yeah, we never had a chance to. Ethically, I think we are not supposed to touch animals in the wild.

EW (00:05:05):

Okay, so theoretically, have you ever pet a lion?

MP (00:05:11):

You are just asking to lose limbs at that point. <laugh> No.

EW (00:05:16):

Seriously? You have not tranqed one and then checked to see how soft it is?

CW (00:05:18):

That is not very ethical.

EW (00:05:18):

<laugh>

MP (00:05:21):

My grad school advisor did once milk lions, at one point. That is one of my fun cocktail party facts.

EW (00:05:27):

Milk?

CW (00:05:27):

That is what I heard.

MP (00:05:28):

Yeah.

CW (00:05:28):

Okay.

MP (00:05:30):

Yeah, like a cow. Except it is a lion.

EW (00:05:32):

<laugh>

MP (00:05:32):

I have never done that myself.

CW (00:05:35):

Well, that is how they get those Tiger's Milk bars. Remember those?

MP (00:05:37):

<laugh>

EW (00:05:37):

<laugh>

CW (00:05:37):

<laugh>

MP (00:05:41):

Yeah, no. I think the problem with working at the technology and all of this in situ sensors, is like the entire point is to do everything non-invasively. Unfortunately I built my whole career around being as non-invasive as possible, so no. I have been chased by lions. I have been-

EW (00:06:01):

<laugh>

MP (00:06:01):

I do not know. There has always been incidents. But I do not think I have actually ever pet a lion, personally.

CW (00:06:08):

All right. Favorite fictional robot?

MP (00:06:11):

Akiba, you can start on this one. I have got to have a think.

A (00:06:14):

Oh. I would say KITT from "Knight Rider."

CW (00:06:16):

Oh, right.

A (00:06:16):

It is not a conventional robot, but I think it was the prototype for all the autonomous vehicles that are coming out now.

MP (00:06:28):

My first thought went to RoboCop, just because Peter Weller has such beautiful lips.

CW (00:06:31):

<laugh>

EW (00:06:31):

<laugh>

MP (00:06:32):

But I know he is not a <laugh> real robot. So maybe those two legged robots in "RoboCop" that take Peter Weller down.

CW (00:06:42):

He was mostly robot by the end of that.

MP (00:06:44):

Was he? Oh. But he had those beautiful lips. Anyhow. Yeah. Sorry. Off topic. <laugh>

A (00:06:49):

Technically, I guess he would be a cyborg?

CW (00:06:51):

Yeah, yeah. Well, we will allow that though.

MP (00:06:53):

<laugh>

EW (00:06:55):

Do giraffes make sense?

MP (00:06:57):

Can I take this one?

EW (00:06:59):

Yeah.

MP (00:06:59):

Okay.

A (00:07:01):

I have nothing to say about giraffes.

EW (00:07:02):

<laugh>

MP (00:07:03):

I have so much to say. Okay. My brief moment of fame, like the only time I have ever been on the international stage in terms of popsci, has been my work with giraffes, where we used camera traps. The core device, that is part of the BoomBox that we are going to talk about, we have a gazillion of them set up across Eastern and Southern Africa.

(00:07:23):

We discovered in this camera trap footage- So researchers do not go out into the middle of the Serengeti at night, because hippos and things that are- They are going to eat you, essentially. It is so dangerous. No one does that.

EW (00:07:36):

Hippos are herbivores.

MP (00:07:39):

No. They eat you. They are the most dangerous thing in Africa.

EW (00:07:40):

They will masticate you. Yeah.

MP (00:07:42):

After mosquitoes, the last thing you want to mess with is a hippo. Anyhow. At night, full of hippos, very dangerous, no idea what happens in the savannah at night. But we set up these camera traps, which are these cameras, you set them up on trees and things. They run 24/7 taking photos of wildlife. So we get that full 24/7 picture of what is happening in the African savannah.

(00:08:04):

We discovered that there is a species of bird, that instead of roosting on trees like normal birds, like regular birds do, this species of bird has decided that the best place to sleep at night, is inside the armpit of a giraffe.

EW (00:08:20):

<laugh>

CW (00:08:20):

Oh my God.

MP (00:08:21):

We have all of these photos of birds roosting in the armpits, and in the groin pits. I do not know how to say this politely, but right up in there. That made "National Geographic." That made everything.

CW (00:08:36):

<laugh>

EW (00:08:36):

<laugh>

MP (00:08:36):

That is like- I am a very serious researcher. I work at Yale, I worked at Princeton. And my moment of fame was an article I published on birds that spend the night nestled up next to giraffe bits. So yeah. No, giraffes, they do not make sense. And the things that sleep in their armpits, also do not make sense.

A (00:09:01):

Did the camera trap record pictures of birds flying out of the butthole? Or just-

EW (00:09:06):

<laugh>

MP (00:09:07):

<laugh> They were not that up close and personal.

A (00:09:09):

<laugh>

MP (00:09:09):

I do know a surprising amount about things that do live in butthole. Like, if you want to- So hippos. There is a kind of leach that has evolved to live inside the hippo butthole. I could talk forever on this. If you ever want to do a side chat about things that live in butthole, this is the ecologist for you. These birds did not live inside the butthole, but they were pretty close.

CW (00:09:35):

So I am enjoying this conversation very much, but this will be the last episode of this show, of all time.

EW (00:09:40):

<laugh>

MP (00:09:42):

<laugh> There is a lot you are going to have to bleep out here. I am so sorry. What is your audience? PG?

CW (00:09:48):

Well, I will make up a determination while I edit. <laugh>

EW (00:09:53):

Okay. Christopher has declared the end of lightning round, even though I still have questions here.

CW (00:09:58):

I am just worried where they are going to go. Go ahead. One more.

EW (00:10:03):

Do you have a tip everyone should know?

CW (00:10:05):

That sounds innocuous enough.

EW (00:10:07):

I know, but I am really- Leeches in buttholes have already come up, so I am pretty sure we are going to go off the rails. Akiba, why do you not start?

A (00:10:17):

I would say a USB rechargeable electric screwdriver, is probably one of the best pieces of field kit you could have.

MP (00:10:29):

Is this a "Doctor Who" reference? Are you telling us to have a sonic screwdriver with us at all times?

A (00:10:35):

<laugh> No. It is something that we just use it all the time. And there are these cheap Chinese ones that you can recharge via USB. That is great, because you can just use your phone charger to recharge your screwdriver when you are in the field, especially if it goes down. You can fit drill bits onto it and other things too. We use it all the time.

CW (00:10:57):

Oh.

MP (00:10:58):

You never gave me this hot tip when I was working in the field. I feel a little bit cheated here.

A (00:11:05):

It is a trade secret.

MP (00:11:08):

<laugh> I feel honored.

EW (00:11:10):

Meredith, a tip everyone should know?

MP (00:11:13):

Oh goodness. I am going to go real classic and say back your data up in three places.

CW (00:11:16):

<laugh> Yes.

MP (00:11:16):

There have been so many times where we have had data that has been like eaten by hyenas, or stolen out of our luggage on international flights. Anything that could happen, does happen to your data. And so my religion is backing data up in three places.

A (00:11:36):

On that note, I would also say that one of the BoomBox devices that Meredith had deployed, got actually eaten by a hyena. It was recorded on camera, too. So it was really fascinating.

MP (00:11:47):

Oh! We have the final moments. It is so good.

EW (00:11:51):

The final output moments? Or just the final input moments?

CW (00:11:55):

What?

MP (00:11:56):

You have the hyenas jaws closing in over the camera. You have got teeth, you have got tongue, you have got like-

EW (00:12:05):

But not an endoscopy.

MP (00:12:08):

We are not going down the digestive tract.

EW (00:12:09):

<laugh>

MP (00:12:09):

No. There is no butthole involved in this story. <laugh>

CW (00:12:13):

It is not a diagnosing camera. They are not checking it for colon cancer.

A (00:12:19):

This interview is going to be like, "Beep hole. Beep hole."

MP (00:12:22):

<laugh>

A (00:12:22):

I can send over a link. I think I have it. I have the video on YouTube somewhere.

CW (00:12:33):

All right. We will put that in the show notes.

EW (00:12:37):

Yes. Please.

(00:12:37):

<music> And now a word from our sponsor, Mouser Electronics. If you are interested in the intersection of neuroscience and engineering, you might want to check out what Mouser Electronics is doing with Brain-Computer Interfaces. It is all about how you can control machines with your mind. It is one of the coolest areas of innovation right now.

(00:12:57):

Mouser's Empowering Innovation Together site has great content on Brain-Computer Interfaces. From videos to in-depth articles, and podcasts that break down the tech. If this piques your curiosity, head over to mouser.com/empowering-innovation and explore what is happening with BCI, and other exciting developments in the world of design and engineering. <music>

(00:13:24):

You mentioned BoomBox, which- I know what that is.

CW (00:13:27):

We are off lightning round, right? We are in the show.

EW (00:13:29):

Right. Because BoomBox is one of the things we were going to talk about.

CW (00:13:33):

All right.

EW (00:13:33):

I assume this is not the thing you hold over your head, when you are trying to make up with somebody, after-

CW (00:13:39):

In the eighties.

EW (00:13:40):

In the eighties. In a movie.

CW (00:13:41):

<laugh>

MP (00:13:43):

I feel like a Akiba's hip-hop background really comes into play, in the naming of this device.

A (00:13:50):

<laugh> Yeah. So BoomBox is a device that connects to a camera trap. So a camera trap records- When it detects an animal using a PIR motion sensor, then it will start recording.

(00:14:12):

BoomBox will actually pick up on the trigger signal to start recording, and then wait a few seconds and play a sound. Then you can actually monitor the animal's behavioral response to the sound. Actually, Meredith is famous for this, so I do not even know why I am explaining this part of things.

MP (00:14:33):

Because you are the genius behind BoomBox. I can talk a little bit about the origin of why we would want a device that does this, in the first place.

EW (00:14:41):

That totally makes sense. Every time I see a bird, I open the Merlin Bird app and I totally want to replay the song, even though I know you are not supposed to do that.

MP (00:14:54):

Do not do that. Do not do that. Ecologists-

EW (00:14:55):

It confuses the birds. I understand. But I want to. And you get to.

MP (00:15:01):

Yeah. I think the origin story here- Again, as we were talking about before, as conservationists and ecologists, which again, I am a conservationist, my introduction to electronics came solely through this project. I am not a technically minded person.

(00:15:22):

Conservationists and ecologist, we were a field that does not have a lot of- There is not a lot of industry. There is not a lot of scaling. There is not a lot of finances for ecology. We are not out there developing our own things. We are out there stealing tools from other fields, and adapting them to our own needs.

(00:15:44):

So we have things like drones, which we took from the military. And environmental eDNA, which we took from forensics. Even GPS tracking devices, that all came out of navigation tools for aviation and navigation for ships. So we are very, very good at taking other tools and trying to make them work for us.

(00:16:06):

But we do not really have a lot of capacity as ecologists, to create data collecting devices that do exactly what we want them to do. So that was a problem I was running into with my research that I did during my PhD and my first couple of postdocs.

(00:16:24):

As I mentioned at the top of the show, my background, I have studied African lions. I am into a large African carnivores. I am quite interested in understanding carnivores from a prey's point of view. So I am interested in putting myself in the mind of a zebra or a wildebeest, and trying to understand what those prey animals do to avoid becoming something's lunch.

(00:16:52):

Understanding all of those tactics, those behavioral tactics that wildebeest and zebra do to avoid getting eaten, is really important in context, like conservation context, where we are interested-

(00:17:02):

For example, if we are restoring an ecosystem, if we have this park or reserve that we are trying to rebuild from the ground up, and we want to reintroduce predators. We want to make sure that the prey animals in there, that these herbivores, that they know what to do. That when you throw a lion in there, that they are not just going to stand there and let themselves get chowed down on.

EW (00:17:23):

<laugh>

MP (00:17:25):

All of my research is on understanding, "How are prey supposed to respond to predators?" So that we can try and reinstill those responses into these naive prey populations, when we are chucking a whole bunch of lions into a reserve, that has not had lions in decades.

EW (00:17:41):

Why do the prey populations get naive? Is it because their matriarchs and stuff are killed off?

MP (00:17:49):

There are a lot of reserves in Africa- Predators are problems for people. They eat our wildlife. They sometimes kill humans. People and large carnivores typically do not get along very well. That has historically been quite a big problem in Africa, where we do have half a dozen of these really large carnivores. It is not just lions. It is leopards, cheetah, hyena, wild dogs. I do not know. There are gazillions of them. And so-

A (00:18:21):

I just wanted to interject really quickly here. Because in the predator-prey behavioral ecology community, there is this wonderful term called "the landscape of fear." I just thought everyone should know that.

MP (00:18:35):

Yeah. This is what my whole research has been on, is landscape of fear,. Which again, is the best buzzword of all time. So you have this landscape of fear, which is this idea that in a savannah for example, there are different places in that savannah based on how your predators hunt, that are very, very dangerous for you to be as a wildebeest.

(00:19:00):

So lions for example, they are not very fast, they are not very sprightly. They are not going to do a long distance marathon. They are more of an ambush and a sprint. They are short bursts of energy. So they are very good at sneaking up at you, from behind rocks, behind trees, at waterholes. So those parts of the landscape, rocks, trees and waterholes, very dangerous for you as a wildebeest. Those are the hotspots in our landscape of fear.

(00:19:27):

Whereas other parts of the landscape, the big open savannas, no trees, you can see from miles, a lot safer. So that heterogeneity, spatial heterogeneity in predation risk, is what is called "the landscape of fear."

EW (00:19:42):

Is that also Akiba's latest album name?

CW (00:19:47):

<laugh>

EW (00:19:47):

Because it would be a great one. I am sorry, Meredith. I am taking this seriously. Landscape of fear. Go ahead.

MP (00:19:54):

Oh. No. I literally called my PhD thesis- Going back to the sonic screwdriver thing. I called my thesis, I think it was- Oh, what was it? It was something about the spatiotemporal distribution of fear in time and space. I tried so hard to make the acronym "TARDIS"-

EW (00:20:13):

<laugh> Yes.

MP (00:20:13):

That it was almost painful at the end. <laugh> So in intact ecosystems, you have this landscape of fear. But people do not like predators. They are the first things that get hunted out or extirpated in these landscapes, in Africa and around the world.

(00:20:31):

So you quite often have these landscapes where 50, a hundred years ago, all of the lions were shot. All of the leopards were killed. All the hyenas were runoff. Because they were killing cattle, harming people, harming livelihoods.

(00:20:46):

So then for 50, a hundred years, for multiple generations of wildebeest or impala, these prey animals grow up without any kind of lived experience of being eaten by anything. Of being afraid of anything. Of having to make decisions between having a drink at the waterhole, and potentially being eaten by a lion.

(00:21:09):

And so in these intact ecosystems, where prey do live with predators and coexist with predators, they have all of these behavioral tactics that they use to trade off that risk versus resources. How they navigate that landscape to avoid predators, but also what they do when they encounter a predator.

(00:21:28):

Because again, in Africa it is not just a lion. It is lions, leopards, hyenas, wild dogs, the whole litany of predators. These predators hunt in different ways. These predators have different superpowers for catching you. So different responses to those predators.

(00:21:45):

Like running away is going to be effective for a lion, because again, they are not really fast, they are kind of slow and sluggish. You are not going to run away from a wild dog. You are not going to outrun a wild dog. Those guys are sprinters. When you see a wild dog, what you need to do, is a completely different behavior hypothetically, than what you would do if you faced a lion or a cheetah or a leopard. And I am saying this-

EW (00:22:09):

What is the strategy, just in case I need to know?

A (00:22:11):

<laugh>

MP (00:22:12):

Well, that is exactly it. So this is what I am saying right in here is the theory. The theory is predators hunt in different ways. They are successful in different ways. Prey should respond in different ways. They must do because we still have predators and prey coexisting. The prey have not been hunted out, so they must do something different. But we had never been able to study that before.

(00:22:36):

This goes back to the idea that we do not as ecologists have tools that collect necessarily the data we need to answer these questions. So the question we are trying to answer is we have this community that has half a dozen carnivores, and we have a gazillion herbivores. We have 40 different species of large herbivores in the savannah. I was interested in maybe half a dozen of those.

(00:23:01):

But if you think about doing robust research, we need a big sample size. We need a lot of replicates of interactions between each predator and each prey, to be able to say that we figured out what survival strategy prey use. Versus, Impala versus lion. That factorial is enormous. That is huge.

(00:23:25):

I have been driving myself around in the Serengeti in a Land Rover for a decade. I have seen maybe five hunts, ever.

EW (00:23:32):

Really?

MP (00:23:34):

Ever. How are you supposed to create or simulate hundreds of interactions, between all of these predators and all of prey? You could spend lifetimes in the savannah waiting to watch these happen for real. That is simply- It is impossible.

(00:23:54):

So we have never been able to answer this question before, because the amount of data we would need to observe actual interactions between predator and prey, is ridiculous. So we need an experimental approach. We needed a remote approach. We needed an in situ approach, to do this that we just never had the capacity to do before.

(00:24:17):

This is where Akiba and the BoomBox came in. This is where Akiba created for me, the perfect device, tailored for all of my ecology needs, to do exactly this experiment, but at scale. And that is what the BoomBox is. This is what the BoomBox unlocks, is our capacity to ask these questions in ecology, that are just- You cannot ask them doing ecology the traditional way.

EW (00:24:45):

It is funny. We see "National Geographic" and David Attenborough going out, and every time we see them, they are in the middle of a hunt. Or we are seeing camera captures of amazing lions doing...

CW (00:25:00):

Lion things.

EW (00:25:01):

Lion things.

MP (00:25:02):

<laugh>

EW (00:25:02):

And you do not realize that we are only seeing the special stuff. We are not seeing day to day.

MP (00:25:09):

I have worked with plenty of film crews. The amount of time they spend in the field, to get that one moment, oh, it must be so mind-numbingly boring for them. Also happy 99th birthday David Attenborough, yesterday.

CW (00:25:23):

Yes.

EW (00:25:23):

Yes.

MP (00:25:25):

Still going strong. Speaking of.

EW (00:25:30):

The different responses between the different types of prey? Like a wildebeest, which is a big strong animal, versus like a dik-dik, which is like a tiny deer. One of those would try to kick an animal, and one of them would run away in all cases? Is the landscape of fear very different for different prey animals?

MP (00:25:54):

It is. And it depends on a lot of different traits, both of the predator and the prey. Again, this is the theory that we are testing with the BoomBox. But hypothetically, again, you have these predators that hunt in different ways.

(00:26:08):

We typically separate them broadly into categories. Of coursing predators. So those are the ones like your wild dogs or hyenas. They are the long distance runners. What they do is they just chase you, until you fall over exhausted. Versus your ambush predators, like lions. That again, just find a rock and they jump out at you and eat you. So there are predator traits that play a role in this.

(00:26:30):

And then there are prey traits. We have everything from body size, sociality. So if you are a herd of buffalo and there are 2,000 of you, that is a great defense against lions. There is the morphology. So some of these animals have horns. They have hooves. Zebras can kick like no one's business.

(00:26:53):

And then there is like, can you run away quickly? Can you vocalize? Can you make an alarm call? There are questions about sensory modalities. We talked about giraffes. Giraffes have great vision. They are very high up, they can see very far away. Other animals rely a lot more on smell or on sound, to detect predators.

(00:27:12):

Again, it is this factorial combination of there are all of these different traits, and ways of attacking prey, and escaping from predators, and sensing predators, that all make this question really complicated. Which adds to the amount of data we need to collect, in order to start addressing, teasing this apart, in a really scientifically rigorous way.

EW (00:27:36):

Okay. So you have the BoomBoxes. How is it working?

MP (00:27:40):

Yeah. So the idea behind the BoomBox, like Akiba just said, is we have- Ecologists rely a lot on a device called "the camera trap," which is a camera. It is rugged. It is waterproof. You set it up in the wild. You attach it to a tree.

(00:27:53):

It has a PIR sensor, so when something warm moves in front of the camera, it triggers- The heat and motion triggers the sensor to snap a photograph or take a video.

(00:28:04):

Akiba, I realize this is a gross generalization of how this works, so please jump in with-

A (00:28:11):

No problem.

MP (00:28:12):

Technical details. But in layman's terms, that is how camera traps work. I have used camera traps for decades. They collect really great data on where animals are, and what they are doing, and who they are associating with.

(00:28:26):

But the problem is, is again, these predator-prey encounters are just so rare to encounter in the wild, and so rare to capture. That even with a hundred camera traps set up in the Serengeti, you are still not capturing enough of these predator-prey interactions, to understand what prey are doing.

(00:28:47):

So where the BoomBox comes in, and what is really clever with the BoomBox, is the BoomBox is not observational. It is an experiment. Experiments are really powerful in ecology and science.

(00:28:56):

So the BoomBox device is this- Akiba is going to be rolling in his grave, <laugh> all the way over there in Australia-

A (00:29:05):

No. No, no. No. No.

EW (00:29:05):

He is not dead yet. <laugh>

MP (00:29:06):

He will be, when I finish describing the BoomBox, the way I understand it.

CW (00:29:11):

<laugh>

A (00:29:11):

No, no.

MP (00:29:11):

The BoomBox is this modular attachment. We essentially hack into the circuit board of the camera trap. And like Akiba said, the BoomBox device attaches to the PIR sensor. So when the PIR sensor is triggered by an animal, and it triggers the camera trap to, in this case take a video, it also triggers the BoomBox device to play an audio recording.

(00:29:32):

These audio recordings could be anything. It could be disco. It could be hip-hop. In our case, we programmed it with predator calls. I had all of these predator calls. Calls from lions. Calls from hyena. Calls from cheetahs. Calls from wild dogs.

EW (00:29:47):

Hippos? Hippo hop?

MP (00:29:49):

<laugh> "Hippo hop" would be the great band's name. I would buy tickets to see that.

A (00:29:53):

I did not expect that one. That was great.

MP (00:29:57):

The "Hiphopopotamus," right? Classic "Flight of the Conchords."

CW (00:29:59):

<laugh>

MP (00:29:59):

But yeah. So the BoomBox device, what it does is- The experiment is, when an animal walks in front of the camera, the BoomBox device roars like a lion, or it twitters like a wild dog, or it growls like a leopard, and then it records- The camera trap records the prey animal's response to that "predator."

(00:30:20):

So that is simulating for us that encounter between a predator and a prey animal. If you deploy two or three dozen BoomBoxes, and leave them going for multiple months, you are capturing- Oh God, I should know this number because it melted my brain.

EW (00:30:38):

I believe it is a crap load.

MP (00:30:39):

<laugh>

A (00:30:39):

<laugh>

MP (00:30:39):

Exactly. A butt ton of recordings.

A (00:30:45):

I stopped hearing from Meredith for a year. I think that was when she was processing all the videos. <laugh>

MP (00:30:52):

Ohh. My goodness. It melts your- Processing videos is the absolute worst.

EW (00:30:55):

Yes.

MP (00:30:55):

You do get some gems. I would love Akiba to share some of these with you. We do have some really funny- As a scientist, I should not find it hysterical how scared animals respond to predators. So if my advisor is listening, please tune off now. But it is really funny.

A (00:31:17):

It is like you are pranking the animals.

EW (00:31:19):

<laugh>

MP (00:31:20):

I know. It is. It is like "Candid Camera" for animals. It is robust scientific work that is producing great papers, great inference. But it is also really funny watching some of these animals respond to predators. Because we just never see this. We just do not know how this works.

(00:31:39):

There are decades of anecdotal evidence. We kind of understand how this works. But we finally got the thousands of videos, the thousands of recordings. Robust, verifiable video recordings that we can sit down and we can watch, and we can rewatch, and we can extract all of that data we need, to be able to start answering these questions in a really rigorous way.

A (00:32:05):

It was really interesting, because the device was specced by a guy named Justin Suraci. I think he wrote a paper on it. And then Meredith had posted somewhere, that she was looking for someone to help her develop a device like this.

(00:32:22):

So when Jacinta and I got involved- Because we were working on an electronic board game that had audio sound effects that you could via MP3. We were like, "Ah. I wonder if that would work." And then we were discussing with Meredith, and she specced out a lot of what she needed. We put together a prototype.

(00:32:44):

The weird thing was that this had to happen in- Was it like three or four weeks? Because you were about to go deploy in Serengeti. So we are like, "Ohh. I guess we have one shot the design. So no PCB respins or anything. Let us just cross our fingers."

MP (00:32:58):

I was so desperate. <laugh>

A (00:33:02):

It worked the first time. But later on we had to make a lot of modifications to- Over the years, we have made a lot of modifications. But luckily it worked that time to collect the data.

CW (00:33:13):

How did you approach development of something like that? Because if it is me and I am given a deadline like that, it is like, "Oh. I am going to go shop on SparkFun. I am going to plug several boards together, and hope for the best."

A (00:33:22):

<laugh>

CW (00:33:22):

But maybe that does not work for something that needs to be in the field, and sitting in the sun, and-

EW (00:33:28):

Hot glue and epoxy.

CW (00:33:29):

Well. Okay. Yeah. So how did you approach this? <laugh>

A (00:33:38):

Because we had already designed the audio subsystem- That was probably the hardest part, is the audio and the amplifier. Interestingly enough, the hard part was the power supply, because the audio- An amplifier requires quite a bit of power. But a lot of that was already designed, so it was just modifying our electronic board game PCB into- Adding inputs for the camera trap trigger, and things like that. So it was kind of serendipitous.

(00:34:12):

Also it was like, "Ah. You know, let us give this a shot." I did not expect it to actually- I expected Meredith to be the only one in the world that would be interested in something like this. But later on we found out that there are actually quite a few researchers, that are interested in this kind of thing.

MP (00:34:32):

Oh. It has blown up. Again, it is because ecologists, we do not have anyone like you Akiba. We spend all of our time trying to make stuff that is not built for us, work for us.

(00:34:45):

I just think that the potential people saw in a device like the BoomBox, to actually do the things we need technology to do as ecologists, it opens a box for a lot of people. And what? You have got projects in New Zealand and Italy and North America and Africa. Where do you not have BoomBox projects running now?

A (00:35:05):

Yeah. Since then we have had a lot of projects, and they have been adapted for a lot of things. Now we are getting into more human-wildlife conflict and non-lethal deterrence. Now we are putting together-

MP (00:35:22):

Talk about BoomBox Disco.

CW (00:35:23):

<laugh>

A (00:35:25):

Well, there is BoomBox Disco. Actually, we are doing a high powered, large amplified version of BoomBox, which is called "Blaster."

MP (00:35:34):

Blaster?

A (00:35:36):

Yeah. So it was kind of the ghetto blaster. The-

CW (00:35:37):

Aah. Okay.

MP (00:35:39):

Love it.

A (00:35:46):

Meredith had mentioned there are different modalities too. So we added extensions to do lights along with the sounds. There is smell. So we used a nebulator.

CW (00:35:59):

Oh wow.

A (00:36:00):

They are just like ultrasonic piezoelectric actuators that will-

EW (00:36:07):

The little fog machine things.

A (00:36:09):

Yeah. Ultrasonic humidifiers use them. Those cheap USB ultrasonic humidifiers. But you can actually nebulize different water soluble things. The first thing that people always ask me is like, "Have you tried pee? Have you tried nebulizing the pee?" And I am like, "Oh. No. I am not going to do that in my lab."

MP (00:36:30):

I have done a lot of work with pee. I did, before the BoomBox. I spent a lot of time wandering around fields with buckets full of wolf urine. So do not diss the pee. It does the job. <laugh>

A (00:36:44):

<laugh> Ecologists are fine with that. They are fine with shoving their hand into a giant mound of poop. And then-

MP (00:36:52):

I have done that. I have done that.

A (00:36:54):

Yeah.

EW (00:36:55):

Well, yeah. You need to see what it ate.

A (00:36:57):

Yeah. Totally. There are different- Yeah. It is quite interesting. It is fun hanging out with ecologists, because they are really into- They are a quirky bunch of people.

MP (00:37:11):

We are gross. We are gross, is what you are saying.

A (00:37:11):

<laugh>

MP (00:37:13):

Can I actually jump in really quickly, and just pick up on something that you said there? Is you mentioned that the devices that you were integrating into these BoomBoxes were cheap. That is something I really want to underscore, the importance of cheap for ecologists as well.

(00:37:29):

Because not only do these devices have to be rugged. They have to endure, like you were saying Chris, wind and rain and sun and dust and ants and-

EW (00:37:38):

Hyenas.

MP (00:37:38):

Hyenas. But also we are ecologists. Our budget every year is like a hundred bucks. Maybe not a hundred bucks. But I have definitely had years where I have $5,000 in my pocket, and I have to save the tigers. That is a problem with a lot of technology. And then also a problem with developing and scaling technology, is that it is not a cheap process necessarily.

(00:38:03):

So the work that Akiba has done, and other people in the conservation technology space, not only to make these devices collect the data we want, work for us in the field, but also be accessible at a price point. For both academics and researchers, like in developing economies and the Global South, places where we have biodiversity, that is incredibly important as well.

(00:38:27):

Something that I spend a lot of time when I interface with technologists, underscoring is again, we have such important work to do as conservationists, but our budget is Monopoly money <laugh> essentially. We do not have a lot to work with. So these low cost open-source, like Arduino, Raspberry Pi kinds of devices, are game changing for us.

A (00:38:52):

I would also say, because there is a huge problem in financing conservation- We are actually giving a talk next week in Thailand with the UNDP on biodiversity finance. There are a lot of really interesting things happening, in looking for novel ways to finance biodiversity initiatives. There is a lot of discussion at the moment about exactly what Meredith was talking about.

EW (00:39:22):

Going back to the human-predator interaction. Human-bio interaction? What I understood-

A (00:39:31):

Human-wildlife conflict?

EW (00:39:32):

Yes. We do not want people to kill off lions and other cool animals, because they are really cool.

CW (00:39:41):

<laugh>

EW (00:39:42):

But for us, it is not a matter of them eating our house and home. It is a matter of going to visit them and seeing, "Oh! They are cool." But there are a lot of people who do not want lions nearby, because they are dangerous.

(00:40:03):

I saw on the freaklabs.org website that the BoomBox works with that as well. Do you wait for a lion to come up, and then you play a lion roar from somebody else, so the lion wanders off thinking this is covered territory? Or how does that work? How do you warn off predators, so that they stay away from humans? Is that a goal?

A (00:40:34):

It is kind of a Meredith question, but I can answer part of it, which is that BoomBox- So we do not really advertise that it tackles deterrence directly. Because the thing is that deterrence is really, I feel like is, highly dependent on the specific animal. You can look at what things could potentially scare off an animal. I think lions, they do not get really scared by much though. Meredith, I think you have a much better idea on that though.

MP (00:41:09):

Yeah, so that is not what I use the BoomBox for. Again, I am interested in scaring prey, more than I am in scaring predators. Human-wildlife conflict is a little outside the direct work I do. It is a consequence of the work I do, but not something I work on specifically.

(00:41:28):

We do get predators in our BoomBox videos, responding to other predators. It is mostly with curiosity. We do have some great videos of things like lions coming up and investigating the BoomBoxes, because they are curious to see what is making these noises.

(00:41:44):

Traditionally when you do- So one way of counting predators is to do a playback, where you play the call of an unfamiliar lion or hyena. That actually brings out the lions or hyenas to come investigate that call. So I do not think playing calls of predators, is going to deter predators.

EW (00:42:06):

Okay.

MP (00:42:06):

Maybe the lights and sounds of some of these new BoomBoxes, could be used as deterrence. Akiba, is not Justin Suraci doing BoomBoxes in New Zealand, to scare cats away from endemic birds? I do think there is some utility there. But it is not something I have worked with myself specifically.

A (00:42:28):

Yeah. That was one of the projects we helped out on also. It was basically to deter feral cats, to prevent them from hunting endangered birds during the nesting, the breeding, season. It was really interesting, because one of the things that are scariest to feral cats, are people talking in normal voices.

EW (00:42:56):

<laugh> We are terrifying. You should totally play this podcast for them.

A (00:43:01):

<laugh> It would probably work. One of the sounds were just like two women having a conversation, and it would be terrifying to the feral cats. It was just like, "Oh, my God." I think humans are probably one of the main animals that most other animals are scared of.

EW (00:43:22):

Changing subjects. Meredith, you mentioned that your tip was to back up your data in three places, and a hyena's gut. But you on your website have a link to DataDryad, which I had never seen. Could you tell us about it?

MP (00:43:41):

Oh! Yeah, no. Of course. Going back to what we were talking about earlier about camera traps. Before the BoomBox and during the BoomBox and since the BoomBox, a lot of the work that I have done, has been in deploying these large scale wildlife monitoring networks.

(00:43:59):

So I have helped set up and run sensor networks of hundreds- Thousands? Thousands. Thousands of cameras across Eastern and Southern Africa and North America, so all around the world. We are setting up these networks.

(00:44:18):

I am interested again in predator-prey interactions and in conservation. But the thing with these sensors, with camera traps, is that they are not just snapping pictures of wildebeest and lions. They are getting aardvarks and zorillas and ground hornbills and giraffes and armpit birds.

EW (00:44:36):

<laugh>

MP (00:44:36):

They are capturing these entire wildlife communities. It is more data than I could ever process or analyze or be interested in doing. I need lifetimes to answer all the questions that we could ask, using all of this data we are collecting.

(00:44:55):

My lab and myself personally, we are big advocates of de-siloing data, of sharing our data with other researchers. There has been a bit of a paradigm shift in the field of ecology in the last decade. Where before, when we did not have these sensors and these technologies and these tools helping us collect again "big data."

(00:45:16):

To get the data we need to do our thesis, to answer our questions, to write our papers, we are spending decades in the field in a tent away from our families eating rice and beans, three meals a day, you put your blood, sweat and tears. Baboons will destroy your tent and <beep> on your pillow. That stuff happens. That data is valuable. You do not give that data up. You do not share that data with other researchers, because you put your life into that.

(00:45:45):

But now we are getting all of this big data. Researchers are collecting a lot of data. We are collecting it a lot more easily. And we are realizing the potential to really use these data to address biodiversity questions at scale. So now again there is this movement, where researchers are sharing their data with other researchers.

(00:46:07):

So there are sites like Data Dryad, like Darwin Core, like Wildlife Insights. Like lila.science is the big one we use, where we have put up hundreds of thousands of camera trap photos for anyone to use. So we get other ecologists using those. We also get a lot of computer scientists using these data sets, to develop computer vision models. I would say those were one of our biggest audiences.

(00:46:34):

So all of our data, our research information, goes online. We love to collaborate, we love to share that data. We want to see research coming out of this information. And yeah, we are really moving into this kind of warm and fuzzy phase, where collaboration is the name of the game in ecology. And we are here to work together to answer these big questions, before it is too late.

EW (00:46:59):

If I wanted to just look at stuff, which one should I start with?

MP (00:47:05):

Oh, if you want to just look at stuff- If you do not want to just look at stuff, how about this? Let me sell you on this idea. You do not want to just look at photos. You want to help researchers analyze data. You want to go on your armchair safari, but also contribute to conservation science at the same time.

EW (00:47:21):

Okay.

MP (00:47:21):

Okay. Okay?

EW (00:47:22):

I will buy that. Yeah. I do. I mean I do.

MP (00:47:23):

Okay, so what we do. Again, thousands of camera traps, collecting millions of photos a year. How do I as one researcher, look at a hundred million photos and write down what is in each photo? Right?

EW (00:47:44):

Interns.

MP (00:47:44):

<laugh>

EW (00:47:44):

Sorry.

MP (00:47:44):

There are not enough interns in the world.

EW (00:47:46):

<laugh>

A (00:47:46):

Undergrads.

CW (00:47:46):

<laugh>

MP (00:47:46):

Believe you me. We rely a lot on interns. But it takes- I once calculated it that for one single project, for example, that we run in the Serengeti, it is 200 camera traps. If I or an intern, they actually measure this in grad student hours. This is a unit of measurement that I find really sad and really funny.

(00:48:05):

But if I was to sit down and extract the data I need- Not doing analysis, not doing insights. But just like, "This image contains a giraffe," from all of the images we collect at that one site. It would take me seven years to process a single year's worth of data. It is bananas.

(00:48:26):

Now we do rely a lot on AI, to help us analyze what is in those photos. But there are some things that AI just cannot do, at the moment. Camera trap photos are not great. They are not like that snapshot you take at a wedding. It is not an animal perfectly framed with wonderful lighting. It is an animal two kilometers away, behind a tree in the rain in the evening, and you just see a tail.

(00:48:49):

So there are some things where we really still do have to do- We use AI, but we still rely on the power of the human brain. What we do with those images is we crowdsource them. We put them online, and so anyone can go to our websites, and look at all of our photos.

(00:49:07):

While you are looking at them, there is a little sidebar with some filters and some tools, to help you understand what animals we have in these different sites. But you just tell us- You look at a photo. It is a beautiful zebra. You tell us, "It is one zebra. It is having a snack. It has a baby." There is some ecological information we want from that photo. You enter that information. We show you another photo.

(00:49:30):

That is such a good way. We call that "citizen-" Or "participatory science" is the term we use now. But you can go and look at any of our photos. Snapshot Safari is the website. That links to, again, dozens of camera trap projects we run around Africa. You can go look at all of our photos. But while you are looking at them, help us out, tell us what is in them.

(00:49:52):

It is real data we use in our research. It is the help of members of the public, of volunteers who look at those photos, is something that keeps the wheels of our research running in a really important way. It has direct on the ground conservation impact. We are so grateful for everyone who tunes in. Yeah, it helps us look at all of those photos, because there are a butt ton of them.

CW (00:50:20):

Are the camera traps devices that are taking advantage of continuing to technology development, and things getting less expensive? Or, I feel like for some of these things it is like, "Okay. Here is the Camera Trap 2000, from literally 2000. One company makes it, and they are the only ones who make it."

(00:50:41):

Is this something that is getting cheaper and easier? Or, is it a legacy technology that you have to buy from one place? Or something like that?

A (00:50:49):

I would say that actually they are really cheap.

CW (00:50:51):

Okay. Good.

A (00:50:53):

It is like a commodity product from China. I think there are only- Truthfully, from the research I have done, I think there are only maybe three or four factories, that actually make all of the different variations and brands of camera traps. They are actually called "trail cams," and they are made for hunters, but-

MP (00:51:12):

We stole them.

A (00:51:13):

Yeah.

MP (00:51:13):

Ecologists stole them.

A (00:51:14):

<laugh>

CW (00:51:14):

<laugh>

EW (00:51:17):

Ecologists' thievery here is very high.

MP (00:51:18):

<laugh> We are resourceful.

A (00:51:23):

Yeah. There totally is a huge amount of innovation within ecology. But, yeah. The trail cams, I take them apart all the time. They basically use image processors from a single company.

(00:51:40):

You can actually make your own camera trap as well. But the problem is that you run into this weird intersection, where you need to have very low power and then a very fast response time waking up from a deep sleep mode. A lot of people try to do Raspberry Pi cameras and stuff, but you just cannot get that, because you have to boot up a Raspberry Pi.

MP (00:52:08):

For a little context on that, is the camera traps, we are not checking them every day. We are checking them every three months or every six months. Which is where the power resourcefulness has to come in, and the going in and out of sleep mode. Because this is not-

(00:52:25):

The point of these sensors is to be remote and deployed in the field and be undisturbed. So we leave them in the field for months on end. If the batteries die in the first five days, then we are screwed. So that low resource environment is really key.

EW (00:52:45):

So somebody goes out in the middle of nowhere and drops a camera. Then three months later comes back, swaps out the SD card, swaps out the battery, and leaves it there?

MP (00:52:58):

Yeah. It is not just- I get asked this question- I think there are a lot of things, again that technologists think there are easy solutions to this, and there are really not. I get asked a lot, "Why do you not just ping the photos over the GSM network? Or send them to a satellite? Why do you need to go check the cameras at all?"

CW (00:53:17):

<laugh>

A (00:53:19):

Like, "Send the data over the 2G network that is available there." <laugh>

MP (00:53:23):

But that is- Okay, A, that is ridiculously expensive, because each camera is picking up thousands of photos a day. B, the cameras that are enabled to do that, are so expensive. We lose 10% of our cameras every year to things like hyenas, weather, poachers-

EW (00:53:41):

Hippos.

MP (00:53:41):

Yeah, hippos. We cannot afford those cameras. But also vegetation will grow up in front of the camera. We do have to physically visit the camera every couple of months, to trim the grass so that the camera continues to have that really great field of view for taking photos.

(00:53:59):

So there are these trade-offs between, we could use really fancy expensive high tech stuff, where we would never have to visit the field. But actually we would have to visit the field, and it would all get stolen anyway.

CW (00:54:11):

Yeah. Yeah.

A (00:54:11):

I would also clarify that, because a lot of researchers do not just take still images with these trail cams. And so one of the big- That is actually not a very difficult problem to solve, to do still images. But what is really hard is to do video.

(00:54:29):

Most of the camera traps do H.264. We are looking at trying to implement H.265. But that is one of the big issues, where you just use up a ton of bandwidth. So video over communication networks like LoRa or Cellular, just does not exactly make sense.

EW (00:54:53):

Or satellite.

MP (00:54:54):

Remember, our budget is like $200. So an Iridium subscription, we blow through that in two days.

EW (00:55:03):

Okay. I do have a question that is related to the cell modems. How do I get a live stream with a lion point of view camera?

CW (00:55:12):

<laugh>

A (00:55:12):

<laugh>

MP (00:55:14):

You give me a grant for $2 million, and then I will set that up for you. <laugh>

A (00:55:21):

I remember, the "National Geographic" has an explorer technology team. Which basically implements a lot of the conservation technology for Nat Geo. They actually have set up action cams that go on animals. You just strap it to an animal, and then it just goes and takes video of everything the animal sees.

(00:55:48):

Unfortunately, the larger problem is the battery life. If you are just constantly recording, then it is not going to last very long. Whereas you can have a streaming camera. They have the great ones with the grizzlies, and especially for Fat Bear Week.

CW (00:56:04):

Mm-hmm. Yeah.

EW (00:56:04):

Yeah.

MP (00:56:06):

Yeah. There are a lot of webcams set up at waterholes around Africa, that you can tune into and take a look. They are not for research, they are for fun, again, because streaming that much data is just so exorbitantly expensive. And the questions that I am trying to ask, you do not necessarily need real-time data.

(00:56:28):

I think there are some conservation questions- There are acoustic devices, for example, that detect gunshots and send real-time alerts to anti-poaching units. Or detect chainsaws. I think there is a role.

(00:56:43):

I think Panthera is a conservational organization, that makes a device called the "poacher cam," which has embedded AI. When it detects a poacher, like a person, it will transmit that image real time to HQ. So they can deploy the men, I do not know, to go find that poacher.

(00:57:02):

So there are some conservation tools that are real-time. But they are deployed very strategically in the field, again, because it is so resource intensive to do that.

(00:57:14):

I trade off in my research, like the scale of data I collect. I collect big data, a lot of data across big spatial and temporal scales. I do not need that data in real time, so I do not do live streaming. But that is not to say that there are not devices for conservation specifically that do not do that. There is a very big role for that. It is a different use case, and a different audience. But those devices do exist.

A (00:57:42):

I would also add that- Because there is a lot of discussion about say, AI on the edge, or AI on camera traps, and things like that. I think from a practical practitioner standpoint though, the problem is that it is so energy intensive to use these AI models. There is not a massive benefit, because you take such a big hit to the battery life.

(00:58:03):

So the practice still is basically just use larger SD cards and collect more data, and then go pick it up. Rather than using AI to select the specific animal, and then stream that up to the internet or something.

CW (00:58:24):

Yeah. And as good as that is now, it seems like you would miss stuff, right? You want all the data.

MP (00:58:30):

A hundred percent. I was just going to say there are concerns about false negatives and false positives. So missing critical incidents, because the AI does not recognize it. But also, again, speaking with my ecologist hat on, AI only recognizes things it has been trained to recognize, at this point.

(00:58:48):

So my armpit birds would not have been recognized by AI. Right? Because we would have never have known to train AI, to look for birds hanging out in something's armpit. So that is also a big concern for us, is missing data and missing critical events with using AI.

(00:59:08):

If it is something like a person, like a poacher, AI is very good with people in vehicles. I think those are tasks that are quite simple to use edge or embedded AI on, for alerts. But for some of the natural history stuff, we just want all the data we can get.

CW (00:59:24):

That problem extends to your post-processing problem of seven years for one year of data, because the same issues exist. It might miss stuff, like you might say, "Okay. I want to filter all of this for lions," but it might not be good, as you said, seeing a lion behind a tree or whatever.

MP (00:59:42):

That is why we use the humans in the loop.

CW (00:59:46):

Things are developing there. There are new ML techniques that incorporate new things, but-

MP (00:59:52):

We just put out some really cool papers on large language models and large multimodal models and generative AI, and exploring their applications in the field of ecology. So I think that there is a lot of exciting stuff with things like zero-shot learning, that are going to revolutionize how we use AI to help process data like this.

(01:00:10):

But at the moment, it really is this beautiful partnership between AI taking a pass, and people taking a pass as well.

CW (01:00:17):

Yeah. Yeah. Which seems like the appropriate way to do it.

A (01:00:20):

I would also say, I think what is really interesting is, that within conservation and also ecology, there are real and impactful applications. A lot of the bleeding edge technology that is coming out, I think there is a gap between people that understand and can implement the technology, and the people that actually really need and use it, and can use it.

(01:00:50):

That is where- I think it is starting to close, but there are a lot of issues. There are huge funding gaps. But what I really like, is you get to work with these really big, messy problems. There are just so many of them. I could just throw a stone and find eight different applications, that need to be implemented or done.

MP (01:01:22):

Oh. Amen. That is why we really need people like you, Akiba, in this space is-

A (01:01:27):

I think there are a lot of people like me in the technology space. <laugh>

MP (01:01:32):

Not people who are willing to listen to ecologists, and again work with us in our tiny little budgets, and with our difficult problems. I find it so frustrating as an ecologist. I do a lot of work, probably more so with AI computer engineers, than I do with hardware people.

(01:01:51):

But people will come in, and they will take our messy data, and they will develop something amazing and new. Like a new algorithm that solves all our problems. They will write the code up in TensorFlow and put it on GitHub, and then they will publish it and disappear.

EW (01:02:04):

Mm-hmm. Mm-hmm.

MP (01:02:04):

And I am there in the Serengeti, in the field, with my Windows 95 PC, and my Tanzanian field technician who has never gone to high school. We do not have internet and we do not have power. I am looking this code and it is like, "Well, I cannot actually put this into practice."

(01:02:21):

This amazing breakthrough in computer vision, is great for computer vision as a field. But the implementation gap, like Akiba was saying, is so frustrating. Because we need people who are willing to translate those achievements in these more academic spaces, and actually wrap them up into tools that we can deploy on the ground.

(01:02:44):

That does not happen enough. It is so frustrating. It is so hard as a conservationist to look at this amazing world of technology. Like, I can look at my phone and it recognizes my faces. The technology that we need is there. But dragging it down into the field of conservation and making it accessible to practitioners, not enough is happening in that space.

(01:03:06):

We really do shout out to any engineers out there, "Please come help us, because we need the support of people in your spaces, to come and hold our hands and make those tools work for us."

CW (01:03:22):

Okay. So I am going to ask a question that has been on the tip of my tongue, and you have set it up beautifully. How- If people, say a burned out engineer who is tired of working on-

EW (01:03:32):

Boring things.

CW (01:03:32):

Boring things, or something.

EW (01:03:34):

Boring BLE nonsense.

CW (01:03:36):

<laugh> Light bulbs that connect to the internet.

A (01:03:37):

<laugh>

CW (01:03:37):

How does somebody get into working with conservationists, or science in general, in a collaborative way, that might not be super well compensated, but might be very fulfilling?

EW (01:03:53):

<laugh> I wonder who you are talking about.

A (01:03:54):

<laugh>

MP (01:03:56):

How did I find you, Akiba? <laugh>

A (01:04:03):

There are a lot of different avenues into it. I would recommend- No, yeah. It is a little bit difficult to say, because I kind of fell into it. There are sites like WILDLABS, so if you check that out, wildlabs.net or- Anyways, there are sites like WILDLABS that focus on conservation and conservation technology.

(01:04:32):

But if you really wanted to get into it, I think it is like within different schools. Or even finding something you are passionate about. Like if you are interested in volcanoes, hanging out in the volcanologist community and then seeing what is needed. Because there are so many- Volcanoes are actually quite fascinating.

EW (01:04:53):

Oh. Yeah.

A (01:04:54):

But there is so much technology that is needed, in so many different areas. It is not that hard to get involved in the development.

(01:05:05):

I think the larger problem is really the longevity, being able to stay involved. Especially because a lot of other industries do not have, say the technology pay scales. The funding is much lower, so that becomes tricky.

(01:05:25):

The one thing I would say though, is that I think it is really interesting, because I never really expected engineering to be as exciting as going into the field. And truthfully, risking my life in a lot of situations. Like, in international development, we would be going into conflict zones, in order to check out water infrastructure, in order to automate water distribution.

(01:06:02):

But there are so many different problems like that. I think it is really about- Part of the journey is searching out what has meaning to you, in order to apply your skills. And I guess it is general for a lot of things. The way that I would put it is there are different pathways in. I do not think there is a set pathway, and it is not a common route to follow.

CW (01:06:30):

Sure. Yeah. Yeah.

EW (01:06:32):

But there are a lot of scientists out there, who need the technology transfer. Some of them have at least a little bit of grant money to pay for it. A lot of them do have power needs, that are not being met by people who are grad students, because the grad students need to focus on their own things.

(01:06:52):

You said wildlabs.net, is that a place that people post for wild things?

A (01:07:02):

Yeah. They are actually quite a large community of people, that are somewhere in between conservation and technology. So there are a lot of technology people, a lot of conservation ecology people.

(01:07:16):

I think, like Meredith was saying though, it is the longevity. When I was working with World Bank, there was a concept called, I think "the parachute-" I forgot what it was. Basically it is people that parachute in, do something, take a lot of pictures about what they do, and then leave <laugh>. I think that happens a lot.

(01:07:52):

But to have meaningful impact, I think it would probably take three to five years of consistent effort at a project, where then you would start seeing real impact and change.

MP (01:08:05):

I once went to a conservation technology conference. I went to a talk for technologists, that was literally, "How to work with ecologists." The opening statement, the opening premise, was like, "Ecologists are really annoying, and hard to work with. So this is what you need to know. They have no money. They do these field seasons. The prototypes never work! It takes years!"

(01:08:30):

Honestly, that was so helpful, I think, to see that laid out. Because, like Akiba was saying, it is not like a quick fix. It is an iterative process like we discovered with the BoomBox. Yes, we whipped up- We <laugh>. Akiba and Jacinta whipped up an amazing prototype in three months, but there have been subsequent years of refinement on that program. And so again-

A (01:08:55):

Was it three months? It felt like three weeks. Everything felt so rushed.

MP (01:08:59):

<laugh>

A (01:08:59):

Wow, okay. Three months is a much more workable timeline.

MP (01:09:04):

I felt every second of that three months. My PhD was riding on that. But, yeah. I would plug WILDLABS quite a bit. I think it is, as an ecologist, where do I go to find engineering support?

(01:09:20):

When I was looking for support to develop the concept of the BoomBox, I went to my university. I went to our engineering department. I looked around our town at commercial businesses. There was not anyone really who could help me do what I needed to do, to do this innovation, this prototyping.

(01:09:38):

So wildlabs.net was a place I went. I would also like to plug Sara Beery, who is a researcher at MIT, runs an AI for Conservation Slack. Which is another community that brings together engineers and technologists and ecologist and conservationists into one place.

(01:09:55):

But yeah, I think that those melting pots are far and few between. So I think that conversations like the ones we are having now, highlighting the needs that these different communities have, and what each community can bring together, are so important. For making people aware that this is a problem we would like help solving, and then directing people to those places.

(01:10:26):

So even going to universities, going to ecology departments and asking, if you are a technologist, "What are researchers working on? What do you need help doing?" I think a big part of the problem we have is as people who are not necessarily in technology, is we only think inside the box. We are not recognizing the potential of emerging technologies. We do not really know what is possible.

(01:10:54):

I would love to see a world where technologists spend more time just going to ecology talks, and just being like- There are so many conversations I have had with technologists, where a problem that has taken me- Data I have collected over ten years, someone has been like, "Oh yeah. If you just daisy-chained that with Bluetooth, you could have done that in five minutes." And it is like, "Oh. I did not even know that was a technology that existed." Right. So I think-

A (01:11:17):

That would be Elecia and Chris, I think.

EW (01:11:19):

<laugh>

CW (01:11:19):

<laugh>

MP (01:11:21):

We need that because we do not know what is possible. We need some help thinking outside the box. I would love to see more of that, those conversations happening.

EW (01:11:30):

That would be fun. Okay. I have one more listener question, from Sahil who requested the funniest field debug story. There is a pun in there, but I was not sure whether it was "field" or "bug." Akiba, do you have one?

MP (01:11:48):

Is the question? <laugh> Sorry.

EW (01:11:50):

No, no. Go ahead.

MP (01:11:52):

Oh. If the question is have I ever had a camera trap full of ants? The answer is, "Yes."

CW (01:11:56):

<laugh>

EW (01:11:56):

<laugh>

A (01:12:00):

I remember Meredith told us this weird story about a friend of hers, that was cultivating a parasite inside of him, or something like that also.

EW (01:12:09):

Oh, is this not the fly that all ecologists have to have?

MP (01:12:12):

Oh! Yes. I have had a botfly too.

CW (01:12:15):

Okay. Well, it has been a good show.

EW (01:12:16):

<laugh>

A (01:12:16):

<laugh>

EW (01:12:17):

I already knew.

MP (01:12:18):

<laugh> You do not want to talk about butt holes for five more minutes?

EW (01:12:23):

Well, botflies are not- Those are just things that you host, because you are a weird ecologist who decides that, "Sure. Lay your eggs in me."

CW (01:12:36):

What? On purpose?

EW (01:12:36):

Nothing can go wrong with that plan.

MP (01:12:38):

It is like a child. But no, I think the friend Akiba is talking about, he films the entire process. It is on YouTube. His name is Piotr Naskrecki. It is a beautiful, heartwarming video about his botfly child, and I cannot recommend it enough. Ten stars. Please go see. <laugh>

EW (01:12:57):

<laugh>

CW (01:12:59):

That is a literal debugging, I suppose.

A (01:13:03):

One of the field stories- There are so many, but one of the field stories is- Because on our deployments here- I am in Melbourne right now, and we have deployments in the Australian bush. Our wireless sensor nodes that we put up- What happens is there is this gap in the back of them, that is a really great habitat for the Australian redback spider. Which is one of the most poisonous spiders in Australia.

EW (01:13:35):

It is Australia. Everything there is venomous.

A (01:13:39):

<laugh> Yeah. Even the cockatoos try and throw plant pots at you from the roof or something.

EW (01:13:49):

And the drop bears.

A (01:13:54):

Huh. I guess that would be my- It is not funny though. It is just more like when we- Every time- Almost every device that we decommission and pull off, we have to check the back, because there are always these redback spiders on them. Before we would do it barehanded. Now we have to wear gloves.

EW (01:14:13):

Have you considered covering that hole? Just a thought.

A (01:14:18):

No, it is like a gap. There is the mounting pole, and there is this small gap that- Yeah.

CW (01:14:26):

A concave thing, right? It is not that it is a hole.

A (01:14:29):

Yeah, yeah, yeah. So it is open at both ends, and I guess they nest there, and other animals come in and they just eat them, I guess.

CW (01:14:38):

Because the cameras are warm.

A (01:14:39):

<laugh>

EW (01:14:43):

All right. I should have really emphasized the "field," instead of the "bug" part of that.

CW (01:14:48):

<laugh>

A (01:14:48):

<laugh>

EW (01:14:48):

But I think now we can- Meredith, do you have any thoughts you would like to leave us with?

MP (01:14:58):

If you do not want to talk about bugs, I do not have anything.

EW (01:15:01):

No. Go ahead. Go ahead. Bugs, butt holes, whatever you want.

MP (01:15:04):

No. Going back to the field debugging, I do think some of my- The wonderful world of electronics that Akiba introduced me to, had me lugging suitcases full of soldering irons and wires and bits and pieces, into the depths of Mozambique.

(01:15:22):

And you are sitting there in your tiny tents, forgetting of course that the power voltages are different in different countries. Blowing out multiple soldering irons. Trying to hook them up. Desperately trying to rewire a camera trap, that fell apart in the bush plane in transit to the field, because you did not solder it right the first time.

(01:15:46):

There have definitely been a lot of those fun moments, which is such a beautiful marriage for the traditional field ecology that I do. Which is tents and dirts and bugs and what have you, with this more lab, technical, sterile, cracking open a camera trap. And there is a shiny circuit board, and I am trying to solder things together. It is really interesting juxtaposing one into the other. I have had a lot of fun with that. Not a lot of success per se, doing field repairs, but a lot of fun.

A (01:16:21):

<laugh> Field repairs suck!

MP (01:16:24):

<laugh> So, yeah. My closing thought then is, just a massive thanks to Akiba for the work he has done both on my project, and it has clearly, as you have heard, been a game changer for a lot of different ecologists and conservationists around the world.

(01:16:36):

And yeah. Just very excited to see over the lifetime the course of my research, the tides turn in terms of creating these sensors and devices and experimentation units, that allow us to explain the unexplainable. We have never been able to answer these questions before, and that is so cool! I do not have a more profound thing to say, other than it is really cool that we can do this now.

EW (01:17:07):

Yes.

MP (01:17:08):

Yeah. It is so exciting to me as an ecologist, to see this happening.

EW (01:17:13):

Akiba, do you have any thoughts you would like to leave us with?

A (01:17:18):

It was really interesting hearing Chris talk about a burned out engineer. Because I was one of those burned out engineers. Back in my life as more in commercial technology. I was always wondering, "What is the purpose of all of this stuff that we are developing? And all of the hours and days that go into learning these new technologies?"

(01:17:47):

I feel that I am really lucky to have an opportunity to collaborate with a lot of really interesting people in a different domain, and learning from them. And getting to work on these really big, messy problems that we are facing. It gives me- It is a chance to really hone my skills in technology, and feel like I am having some kind of an impact.

(01:18:28):

One of the reasons I think I am doing it, is because it is also helping me to find and probably deepen my relationship to technology, rather than seeing it as just an income source. I am not really sure. It is still an ongoing thing for me. But I guess that is what attracts me to this. And you get to see a lot of cute animals too.

EW (01:18:55):

But apparently not pet them?

A (01:18:56):

<laugh>

MP (01:18:56):

You can get to scare them. The scaring them is the best bit, guys.

EW (01:19:01):

<laugh>

A (01:19:01):

<laugh>

EW (01:19:05):

Our guests have been Dr. Meredith Palmer, a scientist currently at Yale University, who specializes in predator-prey research and behavioral ecology. And Akiba is the CTO of Freaklabs, an organizations specializing in technology for wildlife conservation, ecological restoration, and international development. We will have lots of links in the show notes, so please check those out.

CW (01:19:33):

Thanks to you both. This was really fun.

A (01:19:36):

Thanks everyone. It is great to finally get to meet you and Elecia.

MP (01:19:40):

Thanks for the opportunity. It was great to talk a little science with you guys.

EW (01:19:43):

Thank you to Christopher for producing and co-hosting. Thank you to our Patreon listener Slack group for their questions. Thank you to Mouser for their sponsorship. And thank you for listening. You can contact us at show@embedded.fm or hit the contact link on embedded.fm.

(01:20:00):

Now a quote to leave you with. This is from David Quammen. He writes a lot about science and ecology and biodiversity. His latest book, "Wild Thoughts from Wild Places" is a lot about conservation. "Humanity badly needs things that are big and fearsome and homicidally wild. Counterintuitive as it may seem, we need to preserve those few remaining beasts, places, and forces of nature capable of murdering us with sublime indifference."