Bacterial Computers

recording on the same day as the other episode. There's a peak behind the curt most But anyway, we still want to talk a little more about bacteria and uh. And in the last podcast, we kind of talked about bacteria, about how they how we how we've used bacteria to create stuff, whether it's drugs or chemicals of some sort, or even possibly harnessing bacteria to form nano wires or even to do physical work on a very tiny scale. But we wanted to talk about a different application for

bacteria in this podcast. Yeah, we wondered, have you ever considered the fact that the bacteria that gave you food poisoning when you were on vacation um and kept you inside for a week might be smarter than you are, or at least better at math. No, I have not. I thought that I was dumb forgetting it, but I didn't think that the bacteria itself was smarter. Well that's what you get when you eat the seafood special at a gas station. Yeah. Yeah, never again, well maybe one

more time. But yeah, we wanted to talk about using bacteria to to do computations and the fact that a bacterial computer is in fact a possibility for very specific types of problems. People are probably going like, wait, wait, wait what Yeah, no, you heard us right, like, using bacteria to solve computational problems. It's real. We're not talking about you need to clean your keyboard, although that's also true. Gross. Yeah,

this is also not just crazy future is speculation. We're talking about stuff that's already been done in the lab. So so, Jonathan, how what what are we talking about here? What's involved? We're mostly talking about NP class problems. So these are problems that are classically very difficult for computers

to solve. NP problems tend to have lots of variables, and they tend to have a lot of potential uh pathways that you can take and Classically, the way a computer eliminates possibilities is it goes through every single permutation and checks to find out whether or not it's a valid response, which means that it has to go through them sequentially and one at a time, one at a time. So, for example, the one of the problems we can talk

about is the Hamiltonian path problem. Uh, there's also the burnt pancake problem, which do you want to talk about that one? First, let's start with an example. What is the first explore and what the problems are? Okay, so the burnt pancake problem, Well, it starts with the pancake problem. And this is a sorting problem in computation. Um. And so a lot of what computer programmers do is come up with ways of arranging data in a certain way,

like from smallest to largest. Um. So the pancake problem tells you to imagine this. Okay, So you're you're like a waiter running around in a restaurant and you've got a stack of X number of pancakes in your hand. Um, that's not sanitary. It should be on a plate. Well, okay, imagine it's on a plate, but you're you're holding the bottom of the plate with one hand. Um, and while

you're running around the restaurant. Um, you want to arrange the pancakes so that the largest pancake is on the bottom and the smallest pancake is on the top, thus ensuring pancakes stability. Right, very important. Right, you can't put the tallest pancake on the top it say it was just the topple over. It's madness. Uh so how do you arrange it? Well, you've only got one free hand, so you can't go sticking inserting all over. You can

only flip stacks of pancakes. So imagine you've got a stack of pancakes and you can stick a spatula in anywhere along the stack and flip everything that's above the stat the spatula over. So yeah, so you would want to try and find the way of doing this with the fewest number of flaps. What's the fewest number of times you can stick the spatula in and turn over everything on top of it to get the stacks sorted from largest to smallest. See my solution to this problem

is to order waffles. But you could have some waffles bigger than another. Oh well, probably not actually, because the waffle press has a standard size, whereas the pancakes that they spread out according to the poor exactly anyway. So that so that's smart. That's a more symmetrical order. Okay, well, um, so so you're sorting like that. So that's one type

of hard problem. And what you want to do when you've got a pancake sorting algorithm is to figure out the least number of flips, like we said to get the stack in order. The burnt pancake problem adds another complication.

It says that each pancake is burned on one side, got you, and so you want to get the stack in order, but you also want all of the burned sides facing down, got youa so you, So it may require extra flips because some of the when you when you arrange stuff, it may turn out that the one that would have been in the right position is it's from based on size, it's perfectly fine, but because the burnt side is up, you have to flip that. But that means that everything on top of it also gets flipped.

So then you have to figure out how many more flips do you do to get all ordered largest on the bottom and burnt sides down. If you just stop and think about this for a second, this is a

really hard problem to solve. Yeah, it's not something that is you know, you just look at Especially it gets harder the more pancakes you add, right, And that's that's another thing that we can talk about that a lot of these NP problems fall into this category where they get uh much more complex when you start to add extra variables in they get what would is it exponentially more complex? I would need to look into it to make absolutely sure that exponentially is in fact the correct term.

I would say at least geometrically more complex. In other words, that it's it's increasing at a very um predictable rate, because exponentially would mean that it's increasing ten times over each time or what. Okay, yeah, we'll settle that later. Yes,

but anyway, it does get far more complex. And once you get to a certain level of complexity and you're using a classic computer to try and solve this, remember that classic computer is going through every single variation of the solution of that problem, so it has to wait until it's finished with one before it can begin the

next one. If you have a multi core processor, then it can divide that up among the various cores and try and solve the problem more quickly by each core we're working on a different variation of that, but even then you're still only reducing the overall time by a tiny amount compared to how you know, this could take decades or centuries to solve, depending on how complex the problem is. If you're using a classical computer UM and the other NP problem, we're going to talk about the

Hamiltonian path problem. This is one you guys may have heard about. The idea is that you've got these different nodes within a system, or you can give it as cities UM. Sometimes this is a variation of this is called the traveling salesman problem. The traveling salesman problem is a little bit harder because it involves calculating a minimum distance. Right.

This is a little different in that the ideas that you've got Let's let's say you've got five cities and you want to visit all five cities, and the cities have roadways, but not every city is connected to every other city, right, So some cities are connected to maybe just one other city. Other cities are connected to two or three, and you have to figure out which way you can take where you visit each city once and

only once, but you do visit all the cities. So again, a classic computer would have to sit there and go through each variation and validate or throw out the results based upon what happened. So depending on how how many cities are there, that can take a really long time because there's so many different options. So instead of trying each thing, instead of having one smart computer try each

option one at a time. Is there something? Is there something that can do lots of different variations all at the same time. There are a couple of things, and one of them, of course, is what we're talking about. Bacteria. Uh. It turns out that if you are using bacteria and you encode d n A so that it essentially simulates whatever the problem is. Uh, And then you kind of sounds crazy, but you're kind of shaking the bacteria altogether.

What what results is the answer to the problem. So both of the problems we've just described, the burnt pancake problem and the Hamiltonian path problem, believe it or not, have been solved by bacteria. Shut your mouth, yes, okay, bacterial computers what some people call uh back toputing. I wish I had not heard that, But alright, we're bacterial puting, or you're not making it any better pupingpingputing, coputing, okay, okay um. Bacterial computers, literally colonies of bacteria have been

made to solve these problems. Well, how on earth could they do that? They're they're able to do this in parallel as opposed to sequentially. So the classical computer is trying to solve the problem sequentially, like we said, it's going through all but bacteria can actually examine all of these kind of at once. And it's interesting. They researchers had for the Hamiltonian problem. What they did was they

only used three cities, which dramatically reduces the complexity. But it's a proof of concept and it's and they showed that it's a sound proof of concept. What they did was they modified the DNA of E. Coli. Bacteria are good old buddies who make you feel not so good after you've had that lukewarm hot dog. Uh. But they modified the DNA and they ended up shuffling this DNA together and the bacteria ended up producing the correct answer. The back to the DNA was to give the bacteria

command to either glow red or green. And then uh, they put the bacteria in different combinations that represented the various um options for traveling through this this these nodes, and the one that was the actual answer ended up turning yellow. The bacteria turned yellow. And so in a way you're thinking, all right, well, you're just throwing all the options into variations and then you shake it up

and see what comes out. And that's actually true. You know, you're using all these but it is processing in parallel. But but it's useful. I mean, it can solve the problem. It's like even if the bacteria don't know what they're doing. Of course, one might point out that the silicon and

your computer chip doesn't know what it's doing. Yeah, and often I don't know what I'm doing, so I don't see where the But so the bacteria don't know what they're doing, they don't understand that they're solving the problem. But it just happens to be that we can set it so that the ones that do solve the problem make it clear. Yeah, exactly. So what we can do is you create the bacteria with all the different variations

that you need to address whatever the problem is. You do all the different combinations that those bacteria can be in, in other words, like the variations, and then you you take a look and see the results and whatever. Whatever the right answer is is going to be evident based upon whatever you've told the bacteria to do. When it

when it's the result you need. So the what I was reading, if I understand correctly, the way they ascertained the success of computing the burnt pancake problem in bacteria was they I think they introduced an antibiotic into the bacterial colonies and basically they had worked it so that bacteria that solved the problem correctly would end up with a resistance to the antibiotics, so those bacteria would survive

in all the rest. Ye off, Yeah, yeah, that's and that that's a very quick way of taking a look and saying, all right, here's your answer, because it has to be. Uh, that's it's kind of an interesting approach. The other the other big method I've heard of for tackling these sort of problems, besides using something like a graphics processing unit which has lots and lots of different parallel processors in it. Uh, they that would still take a long time to solve. A lot of these NP

problems um depending upon how complex they were. But quantum computers are very take a very similar approach. But that's a whole other episode. It's a different episode. But just really quickly, a quantum computer, you know, a classic computer uses bits which are either a zero or a one, so that that binary digit um, that's what bit stands for. Well, quantum computers use cubits qu b I t s, and these are bits that exist in superposition, meaning that they not just a zero or one, they are both a

zero and the one and technically everything in between. And theoretically, if you had enough cubits together, you could process these problems in parallel, where each cubit is either a zero or a one or both or whatever, and you process all the all the possibilities all at once, assuming that you have enough cubits to tackle whatever the big problem is. So you would have to first create enough cubits to be able to do this. If you did, then you

could solve the problem in parallel. We're really good at manipulating cubits, aren't we not. Unfortunately, no, no, because if you it's very it's a very um delicate system. Anything time you're you're talking about quantum effects. You are talking about a very delicate system that can very simply break down. And if it broke down, your quantum computer would essentially become a classical computer. Do we do we interfere with the state of the cubit whenever we look at the cubit?

I mean, that's the problem, right. And then also I should point out that the answers you get from a quantum computer are probabilistic. They are not definitive. So you would get essentially a level of like a percentage of how accurate is this answer? Like is it it's there's an eight percent chance this is the answer kind of thing like you're never gonna get response where you're going to know for sure that you've got the right answer.

That's again grossly under stating what a quantum computer does. But you know, for the purposes of this discussion, I thought it was interesting just because again it was one of those things where as I was reading about these bacterial computers, I thought, well, that's kind of similar to what a quantum computer does. It's just doing it with d n A and doing it in a you know, not in a quantum like a weird quantum effect kind

of way. Ps right to us and tell us to do an episode on quantum computers and we can convince our bosses to do it. Yeah, we need enough of you guys to say I really want to know about this quantum computer. Also we can talk about quantum encryption, which is wicked awesome. But yeah, you need to write inn until So Yeah, so we have solved problems with bacteria. Yes, what what else do we need to do to make

a bacterial computer. Well, you've got to be able to go beyond just the prototype stage and make sure you have. You have to. It's all a question of scale, right, So the scale is an issue. And also you know you got to feed your computer as it turns out. Well, maybe here's how I should rephrase it. Are the ways that we know we can use bacteria in uh, or at least we think we will be able to use

bacteria in in computing in the nearer future. In a in an indirect way, there are bacteria that can produce essentially magnetic material magnetite, right, magnetite, it's the most magnetic natural mineral in them yep. And there are there's bacteria that that consumes iron and creates magnetite, and so you know, if you've ever wondered about how how a hard drive works, a classic hard drive, those are magnetic drives. They are

storing information magnetically. That's also why if you've ever worked with any kind of computer or magnetic storage, you know you're not supposed to bring powerful magnets anywhere close to that because you could accidentally wipe whatever it is you're working with. Um. Well, there there are bacteria that produce magnetite, and there's a possibility that we could use these to

create new types of hard drives. Because one of the challenges we're encountering is designing uh computer components that are on an increasingly smaller scale. It's hard to get to a level of precision once we get down below a certain level, and this might give us a chance to create even more precise means of constructing something like hard drives. So it wouldn't be a full computer, it would be a computer component. But that's an actual possibility that we

may see within the next few years. That's pretty cool, now, that's pretty awesome. The idea of harnessing bacteria to create things that that will increase our own ability to to store information and access information. That's interesting if you've listened to our last episode that that isn't the only thing, uh that bacteria can manufacture. Now we can create the

like wires I mean, and the wires thing. To me, the wires thing is really interesting because the idea of being able to create wires that can exist within an organism means that you have a lot of options when it comes to cyborg implants. And as we've already covered in a previous episode of Forward Thinking, I'm an aspiring cyborg. So that'll just pave my pathway to robotic uh superiority that much more quickly. So you know, stay on. That

gets a crazy look in your eye. I pretty much always have a crazy look at my eye, but I have a have a specifically crazy look at my eye here. I'm I'm mostly imagining that scene in Terminator where you see the robot on the mound of human skulls. That's you know, that's my happy They end up there. They don't necessarily start there, but I find a way to find a way. Yeah, how does the the naked in high school dream turned to well, standing on the skulls?

Clothes don't don't transfer over when you go through the time point. That's correct. So okay, speaking of robots, let's do that. I had a kind of crazy idea. Um and and here we're going to utter speculation. But in the last podcast, we talked about how bacteria can be used to generate mechanical motion. Right, so you can, um, you can regulate the amount of oxygen and fluid um with bacteria suspended in it to get the bacteria to push micro gears. They're actually they're turning gears. They can

operate tiny machinery. Okay, so you've got mechanical motion. We just talked about computing. Those two things put together are the building blocks of something we would call a robot. Yeah. Now, if you're talking about a bacterial robot, uh, this is not a nanobot. It's going to be larger than the nanoscale, but it would still would be made of bacteria. Yeah, you can have a in theory if we were able to harness this sort of power, then yeah, I guess

I imagine. So, I mean you would still need uh, some means of processing that information because right now, what we're talking about with the computers is mostly you know, put all the dice in a in a in a bucket and shake um. And when you remove the bucket and you reveal what the dice have rolled. Because of the way that this system works, only the dice that are necessary to answer the question show up and you

get the answer. Going from that to some sort of computer that can actually process information in a meaningful way and give commands, that's a big jump. It's not to say that we won't make it so. And also, like we've said before, if we're if we're talking about building something at a very tiny scale, it it behooves us to look at nature because they're already these natural organisms that are on this tiny, tiny scale that do very well.

So if we can emulate that or manipulate them in some way, then we can create things on that scale much more quickly than we would if we just build from the ground up, so to speak. So could we have a bacterial robot? I don't see why not. If we did, what would it be good for an autonomous bacterial machine? I would imagine that that would have medical uses.

I mean something that could especially if you had any way of guiding it in within a larger organism than you could end up using that to either administer medicine or do even you know, think about cellular level surgery, you know, I mean talking about surgery on the cellular level where you're actually repairing injured or damaged or or diseased tissue with something that has a level of precision far greater than even the sharpest scalpel. So that's a

potential use for it. Whether we actually ever see at I don't know. I'm a I'm a little skeptical. It sounds like it's a little science fiction e to me. But then I'm sure if we had talked to people, you know, fifty years ago, they would say that the stuff that we take for granted now is the stuff of science fiction. It would never come to pass the Internet, the Internet, having the Internet in your in your hand,

having things like tell Us surgery. I mean, we already have tell usurgery where people can can perform surgeries from across the world. Well, it seems clear that if there ever will be a bacterial robot, it's a long way off. But another thing that seemed possible to me is a use for a bacterial robot is um So how do you get your other bacterial machines to do their job. I mean, if we have like microfluidic machines that are controlled by bacteria, it could be useful to have autonomous

bacterial machines that know what they're doing via the computing power. Yeah. I would imagine you would mostly have to have some bacteria owen sentives in there, like some bacterial coffee, uh, like a good bacterial life work balance, you know, some maybe you could work from outside the organism three days all the week, a bacterial vacation plan. Yeah, yeah, every

now and then your holidays in Jonathan Strickland's lower intestine. Yeah, well, boy do they But to seriously answer your question, that's that's a good that's an excellent question. And honestly, it's one of those things where again, like I can kind of see where the individual pieces would need to come

in from. I have no idea what the execution would actually be, but it's one of those things where I'm sure there are people who are working on that, and they're also you know, keep in mind, we've got people working on on manipulating and altering bacteria from multiple disciplines, so we're going to see advances from many different disciplines all kind of combined together in the future to h to create stuff that we can only dream about right now,

or some of us can't even dream about it. It It may be beyond even the crazy stuff we're talking about here. So that's the coolest thing I think about this kind of research is that it will lead us to discoveries that you know, we can only speculate on, and I doubt we'll even get close to the cool stuff that

comes down the line. This world is so weird to us, the very very small, Yeah, and also outside of the way we think that I always say, like, there's certain worlds that are largely a mystery to us, the very tiny and the extremely large, like the cosmological scale. But there. But the thing, the thing that makes them really different in my mind is on the galactic scale, there are no machines there. Well, yeah, not that we know. I mean,

God would be terrifying. We could be living in a computer simulation, in which case there could be galaxy sized cells of bacteria. Yeah, um, but as far as we know at the at the galactic level, there are no machines. At the micro level, there are tons of machines. There's there's a machine for almost everything, and if there and if there isn't one, we can manipulate it to do it. Yeah. Yeah, it's pretty awesome. So uh yeah, well, we'll have to see how will harness bacteria in the future and what

kind of interesting applications we can have. We've really just scratched the surface here, and there are a lot of incredibly smart people who are really pushing the boundaries of what we thought was possible even as little as five years ago. I think if you had taught to some of the researchers five years ago about some of the stuff that's going on now, they would have thought you were crazy. So it'll be interesting to see how this

goes in the future. Guys, if you have any suggestions for future topics of forward thinking, please get in touch with us. Let's know, like, let's say you want to know about quantum computers and you think I hope they get around to that. It will help if you're write in or if there's any other topic you want, let's know our email addresses FW thinking at discovery dot com

or go to FW thinking dot com. That's the website where we have all of our videos, the podcast, the blogs and all of the links to our social media. So if you want to get in touch with us on Twitter or Facebook or Google Plus, that's where you can go to find us and drop us a line. Let's know how you're doing, you know, tell us what you're excited about in the future, and we will be sure to take the end of consideration. We're excited to hear from you, and we will talk to you again

really soon. For more on this topic and the future of technology, visit forward Thinking dot Com, brought to you by Toyota. Let's Go Places