Unexplainable - A brainless yellow goo that does math

Episode Date: November 10, 2021

Slime molds can navigate mazes, control robots, and make complicated decisions, all without a central nervous system. If this weird gooey blob can think, does that mean scientists are thinking about intelligence all wrong? For more, go to http://vox.com/unexplainable It’s a great place to view show transcripts and read more about the topics on our show. Also, email us! [e-mail] We read every email. Support Unexplainable by making a financial contribution to Vox! bit.ly/givepodcasts Learn more about your ad choices. Visit podcastchoices.com/adchoices

AI was supposed to take over the parts of the job you hate. Turns out, it made your job even harder. Instead of doing the work, it gave you homework. ServiceNow's AI specialists get work done from start to finish. Cases get resolved. Loops get closed. With ServiceNow, you can do the parts of your job you're best at and delegate the rest. To put AI to work for people, visit servicenow.com.

Some gifts say, I thought of you. The best ones help you discover more. This Mother's Day, give her something personal with Ancestry DNA. Now, up to $75 off. Explore her origins and discover the journeys that made her who she is. Save today. Give her something unforgettable, thoughtful, meaningful, uniquely hers. Give more than a gift for less.

Give AncestryDNA. Visit Ancestry.ca. Today, offer ends May 10th. Terms apply. It's unexplainable. I'm Noah. Hassenfeld, here with senior producer Meredith Hadna. Yo.

What is going on, Meredith? I really wanted to tell you about this, like, ooey-go-y creature thing that has been blowing my mind recently. Okay. What kind of creature? Slime mold. Slime mold. Slime mold. What is that exactly?

What does it look like? Yeah, so slime molds, they can come in a whole lots of different types and sizes and shapes and shapes and stuff, but I fell down this rabbit hole for one slime in particular. I'm going to have a little trouble with the Latin, but it's fizerum polysexual. Oh boy, it's hard to describe. So it feels kind of like slime, like mucus. This is Tanya Laddy. She's a professor over at the University of Sydney.

But if you look at it, it tends, especially in the lab, you start to really see the structure, which kind of looks like a series of veins. It almost spread out like a tree. It's bright yellow and it like squidges out in this blob. And then the whole thing is just gradually moving forward, tiny, tiny bit by tiny, tiny bit. It can move? Yeah, very, very slowly, like four or five centimeters an hour. Yeah, it's the kind of speed where if you go to the bathroom and come back, it's like, oh, did it move?

I'm sure it moved. You can almost just kind of see it. When conditions are right, it can, like, ooze out to the size of a bath mat or even bigger, like, meters. Okay. And the crazy thing is it's all just one cell. The whole... One cell. Bath mat of slime mold is one cell?

What does that even mean? The easiest way to think about a slime mold is to imagine an amoeba and then make it big. So it just has one big cell membrane. And what is it exactly? Is it like an animal? Is it a fungus or something? Neither. So if we think of the tree of life, you know, we all have a common ancestor and the different

lineages kind of branched off at different times. The group that contained slime is branched off from that tree before fungi split from animals. So that's a really, really long time ago. So these are like further from us than any other animal or plant. We're definitely more related to mushrooms than we are to a slime mold. So it looks really weird. It is really weird. What makes this so interesting?

Why did you fall down the rabbit hole on it? So even though these things like, they couldn't be further from us on the tree of life, they're weirdly super smart. Okay. We don't really know how they're pulling this off. But the more scientists learn about them, the more they're thinking that slime molds could potentially redefine how we think about intelligence. Hmm. Okay. Okay. What makes them so smart?

Well, okay. They don't have a brain. Sure. They don't have a central nervous system. It's just all goo all the way through. Okay. But it can do all these tricky things. For me, that's really mind bending because we just sort of assume that brains are like the best, you know? And that in order to be able to solve problems, you must have a brain.

And I think that's caused us to potentially overlook all these really interesting alternative ways of solving problems, you know, like the slime molds must be using. What kind of problems can slime mold solve? All right. You ready for like some mind-blowing stuff? Hit me, let's go. Turns out, slime molds can navigate mazes. If you put a slime mold in a maze and you have a food source at the entrance and one at the end, the slime mold will connect up. those points through the shortest possible path.

And it's particularly motivated by a ferocious love of oatmeal. Get out of here. Loves oatmeal. So like if you put oatmeal on the couch or somewhere and you have a slime mold, like the slime mold will slowly inch towards the oatmeal? Oh, yeah. Huh. What other kinds of things can it do?

Yeah. So instead of putting food in a maze, researchers put these dollops of oatmeal on a map. Okay. on all the major population centers of Tokyo. And the slime mold, it connected them in the same way, using the most efficient means possible, and just independently happen to map out the entire Tokyo metro system. Just so I can be clear, like, it's a problem that I assume

would require a bunch of engineers to do a bunch of complicated math to be like, what's the most efficient way to map these things to connect them, and the slime mold can just figure it out. Exactly. It just connects them with the fewest amount of resources. There's so much more it can do. People have built robots controlled by slime mold. It can escape from traps, and it can make some pretty complicated decisions. What kind of decisions?

All right. So let's say you're a slime mold. You're in the slime mold like mindset, oozing. I can ooze. So you come across two sources of food. One is super delicious oatmeal, your fave. Yum. Obviously.

Love it. The other is just like a few oat flakes. Ah. So what do you do? I go for the oatmeal. I mean, obviously, easy decision. Mm-hmm. So it's really easy to make a decision when it's, you know, really good or really bad.

Like, you take the really good. Of course you do. But let's say that the good food choice was dangerous and the bad food choice was safe. So the delicious food is under a bright light. It's super exposed. And like the crummy food is tucked away safe in this nice, deep dark shadow. It's much harder to make that decision. if really good is also guarded by a monster.

And now you've got to make that decision, like what's more important? The food quality or the danger factor. And it turns out that the slime mold only goes for the good food if it's five times better in quality than the bad safe food. So the slime mold can do multiplication now. Yeah. So it's like if you had to cross a highway for a donut,

that had better be one freaking amazing donut to be worth it. I mean, this now brings me to my, the question that I've been wanting to ask for a while, which is just like how, like how is it doing that? Yeah. So scientists don't really know. But I think we've got some ideas. Tanya suspects it has to do with collective behavior. She studies the collective behavior in social insects like ants or bees or soldier flies. And the patterns that I was seeing in the slime molds after.

I fed them, looked really similar to what I was seeing in the ants. When ants explore an area, they'll lay down a trail of these messenger chemicals for other ants in the colony to find, like a little post-it saying like, hey, over here, found food this way. And the other ants, they'll stumble across this trail and start following it. And as they go, they'll lay down their own chemicals. And soon the whole colony is, like, following this trail, right? So, like, all of this, this is an example of a positive feedback loop.

And we think that that collection of behaviors is really at the heart of slime mold decision-making. Wait, I thought slime mold was like this one big amoeba. How would it be doing this through collective behavior? Right. It's totally one giant cell. But inside it has tons and tons of nuclei, the like center organelle, each regulating their own little DNAs, right? So if you cut up a slime mold, within minutes, each of the pieces are fully independent. So you can take one giant slime mold cut into a thousand pieces. They're all sort of individuals, even though they were not individuals the day before.

You put them back together again, they'll join up. So the whole idea and line between like what's an individual and what's like a community or collective is really blurry. So a slime mold is sort of like a bunch of nuclei coming together surrounding. by goo or like a big sack of ants. But with the ants, they all have brains. Like, how do the parts of a slime mold do this? It's very weird. So if you look closely at a slime mold, like under a microscope,

within that bright yellow goo, there's these pulsing veins. And that's driving the flow of goo. And it's all flowing, and that flow is being driven by these pumping motion of the veins. But not all those veins are pulsing at the same. rain. So if a slime mold finds something it likes, the veins in that region start pumping faster. It's like, oh yeah, go and get that oatmeal. Okay. And it influences the regions around it to also pump faster.

And that pumping faster, it's moving more of its goo self in that direction, and it's moving that goo self faster. And then that goo starts pumping faster, and there's your positive feedback loop. They're probably not doing very sophisticated behavior other than speech. behavior other than speeding up and slowing down. But on a whole, collectively, they're able to do really complicated things. So it's not exactly like the slime mold is making a decision. It's more that the parts of a slime mold react in certain ways, then that causes a chain reaction where the other parts of the slime mold react in different ways.

And that ends up sort of looking like a decision. Yeah. I mean, that's, I'm, it's, I'm struggling just because, like, how do you define a a decision versus not a decision. I mean, like, there's this thing called swarm intelligence, where the collective actually does have access to more resources than any one individual. And just through these relatively simple mechanisms, you see this with ants, you see this with bees, but also like schools of fish and flocks of birds and even like crowds of humans, individuals all making individual decisions, but you bring them together and the collective is much

smarter than any of the individuals. So yeah, I'd argue that sure, one little region pulsing isn't effectively making a decision, but the collective of this positive feedback loop actually does amount to more than the sum of its parts. Yeah. I mean, it also makes me think about our own brain, right? Like, what does a decision mean in my own brain? Who am I to make the decision? Or is it a bunch of little neurons that's each reacting to little stimuli and all acting together. Obviously, we're different from slime molds, but... I mean, not as much as you'd think. If you took a neuron out of your brain and studied it in isolation, you'd think, well,

it's very simple. You know, a neuron can't do calculus or, you know, produce podcasts or write books or do any of those things. But if you stick billions of them together and let them communicate, all of a sudden you have this really sophisticated decision-making system. that you would never predict from looking at the neurons in isolation. It's only when it's networked and within this collective and building on all these feedback loops

that our own consciousness can arise. On some levels, I think we're starting to see these really similar similarities between very different types of decision-making systems, almost like we're running similar software on really, really different hardware. What does that mean? So our brains use feedback loops too, just like slime mold uses feedback loops.

That's the software. We're just running it on very different hardware. We have these big, squishy brains that we carry around in our skulls. Well, slime molds, they have goo that they go around with. So on the surface, slime mold couldn't possibly be more different from us. I mean, there are these little goo piles that are one cell that are obsessed with oatmeal. Okay, that's not totally different. but they seem very different from us,

but I guess in the sense of how they work or how they even think if we can use that word, they're not so alien from us. We're not that different. Yeah, I mean, true. In one sense, they do kind of work, like our brains kind of work, but they still don't have a brain.

Right. They still don't have a central nervous system. It's just this weird sack of goo with a collection of nuclei, and it's doing all this crazy stuff. stuff and it's freaking nuts. At the end of the day, we thought brains were necessary for intelligence, and they just might not be.

There could be all these possible ways of getting to the place where something is intelligent, and we just need to open our eyes to the possibility of what that could look like. Could look like a sack of goo. Coming up after the break, how intelligent slime molds are leading researchers to reconsider long-held beliefs about some of the smallest, simplest forms of intelligence. That's next. It's all about you. And when you fly with Virgin Atlantic in their upper class cabin,

they take the VIP treatment to the next level. With a private wing to check in and your own security channel at London Heathrow, you can glide from your car to their clubhouse, a destination in its own right in 10 minutes or less. On board, you can treat yourself to your own private suite to stretch out in with lots of storage space, a lie flat bed and delicious dining from beginning to end.

Just be sure to be sure to be. leave room for dessert. Their mile high tea with all the little cakes and sandwiches is a showstopper. Go to virginatlantic.com to learn more. Visit BetMGM Casino and check out the newest exclusive. The Price is Right Fortune Pick. BetMGM and Game Sense remind you to play responsibly. 19 plus to wager. Ontario only. Please play responsibly. If you have questions or concerns about your gambling or someone close to you, please contact Connects Ontario at 1-866-531-2-6-600 to speak to advisor free of charge. BetMGM operates pursuant to an operating agreement with Eye Gaming, Ontario.

All my circuits will be right back after this word from slurm. It's highly addictive. Okay, unexplainable. We're back here with Meredith. Hey, yeah. So we've been talking about slime molds, which are these weirdly huge cells with lots of nuclei inside that are somehow super smart. Scientists don't really know how, but they think it's because of some sort of swarm intelligence. where the whole is greater than the sum of its parts. Yeah. Slymold's super impressive. I've been, like, losing my mind about it for weeks.

But scientists are also starting to wonder, are slime molds unique as single-cell organisms? There are lots of single-celled organisms out there. Lots of them are much simpler than this, like, crazy collective that slime molds are. Are there other ways that a single-cell can be smart? Okay. This gets into even muddier territory than we were in before, but it really pushes that question forward,

the question of like what do we even think is capable of intelligence? Could it be some of the smallest, simplest, simplest life on Earth? How would that work exactly? We don't know, but people have been asking this question for a long time. Learning in single cells is an old idea. It goes back to the days when people were first starting to think about learning from a scientific point of view. So I talked to Jeremy Gunnarwardna. He's a biology professor at Harvard, and he's been doing some historical research on this scientist named Beatrice Gelber. She was a PhD at the University of Indiana in the 1950s.

So Gilbert had this idea. She wanted to test complex learning in single cells, the way that most people would test complex learning in animals. You know Pavlov, Pavlov's dog? Sure. Heard of Pavlov. Right, right. So like Pavlov, he fed his dogs and they would drool and salivate. And then he would start ringing a bell every time he fed them, conditioning them to link the food with the bell. Uh-huh. And after a few rounds, he could ring the bell and the dogs would start slobbering even if there wasn't any food around. And that's like the dog learning to associate the bell with food, right?

Exactly. But with Gelber, instead of dogs, she was looking at ciliates. Ciliates? Okay, so ciliates are a group of single-celled organisms. And before the advent of kind of serious multicellular organisms, before the advent of animals and plants and fungi, ciliates were probably the apex predators on the planet. So like a little single-cell shark.

Okay. And then instead of dog food, she gave the ciliates. their food, bacteria. By coating a wire in bacteria and then putting that wire in the solution with the ciliates and counting how many of the ciliates came over to munch on that lunch buffet of bacteria. Mmm, salty. This is like, you know, feeding the dog some food and also ringing the bell.

Right, exactly. So after a few rounds of this, Gelber puts a clean wire into the solution. No food, no bacteria. and then she counted how many of the ciliates swam over. Okay, so she's conditioning the ciliates with this wire. If they see a clean wire and they swim over, that sort of means they've learned to expect that the wire means food? Totally.

So her experiments in her hands showed that the ciliots could learn. They swam right over to the clean wire. Okay, but at the top you were saying that no one really knows if tiny single cells can learn. Isn't this proof? that they can? Well, Gelber's research wasn't very popular. There was this sort of very strong bias towards believing that you needed nervous systems to do sophisticated tasks. People thought she was projecting too much onto these little single cell organisms. Like, she was reading into this as

intelligence when it was really just random movement. And when they tried to replicate her results, they couldn't. And so her work was dismissed and lost a history. If scientists tried to replicate her work and they couldn't do it, why is Jeremy now interested in revisiting her work? So Jeremy thinks that Gelber didn't get a fair shake. One of Gelber's most serious opponents tried to reproduce her experiments. I suspect he didn't believe it worked and therefore he just didn't pay really serious attention to the kind of parameters that Gelber was using in her experiments. And the moment he couldn't reproduce it, he was very quick to say, well, you know, it doesn't

work. The scientists that were replicating her work in the 50s, they didn't pay close attention to her protocols. They used different concentrations of bacteria. It's a really complicated, really fiddily experiment. So Jeremy's looking to replicate Gelber's work with modern tools and techniques. And he's already replicated some simpler experiments. for simpler forms of learning on single cells,

and his research looks really promising. But ultimately, we're just still not sure if cells can learn at this level. Okay, so we have research that looks promising that your average, tiny single cell might be able to learn. And then we also have this knowledge that one kind of weird, huge single cell, the slime mold, can learn.

So it feels like we have, you know, results, in two different directions, at least pointing this way, what would it mean if scientists actually proved that all single cells could have some form of intelligence, like they could learn? If single cells could learn, it would totally expand our idea of what's even possible in biology. And just even how we conceptualize or frame problems, like cancer research. Often cells are seen as these little robots executing instructions in the genes or DNA. if there's a problem that leads to a tumor or something, there was either a problem with the DNA or a faulty instruction

or a problem with how the cell carried it out. But there's not a lot of agency there. Right. If cells can learn, that totally shifts the paradigm. If we can show what we believe is the case, I think the ramifications of that will actually go right through biology because it will alter the way we think about all organisms, because all organisms are ultimately composed of cells.

It's the same molecular machinery that basically runs all of life. Right now, we're still at the stage where we're trying to figure out if cells can learn. But scientists don't really have a great explanation of how that would be possible. Right. Like with slime mold, at least we have some theory of swarm intelligence, but when it gets down to regular teeny cells without the help of multiple nuclei, If those guys can learn, I think it forces us to reconsider what's possible. Like, maybe we need to be a little bit more humble about our place in the world as these big fancy thinking machines.

Because at the end of the day, like, we really don't know that much about what intelligence means and what it can look like. This episode was produced and reported by Meredith Hodnott. Noam Hassanfeld wrote the music and edited the episode, along with Brian Resnick and Catherine Wells, with help from Bird Pinkerton. Manning Wen, check the facts, and I, Christian Ayala, did our mixing and sound design. And Liz Kelly Nelson is the VP of Vox Audio. If you have thoughts, email them to us, Unexplainable at Vox.com. And if you feel like leaving a nice review on Apple Podcasts, I know we'd all really appreciate it. Unexplainable is part of the Vox Media Podcast Network.

We'll be back next week.