 I am going to talk to you today about three hard problems. Let me try this. Let me see if this works. Oh, very good. One of them is old. One is a construct, and you're going to get angry about it. And the last one is the one that we're going to solve together. So this first one, I'm going to start off. This is going to get weird, guys. So just get this thing. This is my daughter, Natalie. And she's in sixth grade. This is one of her homework assignments. And if you don't mind, let me just read this out to you. And this will make a lot of sense while we're getting into this. So Jenna tells her parents the secret number she's thinking of is the largest common divisor of the number she tells each of them. If she tells her mother 510 and her father 306 and she is telling the truth, what is the secret number that Jenna could be thinking of? OK. So the reason Natalie's doing that is because she's got to do like 50 of these problems. And this is a talk about problems, specifically hard ones. And so this is how you solve these problems. At least this is how you're told to solve them. You're supposed to take this number and then break it up into its prime factors. And so here you get 510. It's equal to 2 times 3 times 5 times 17. Great. This one here, 306, it's 2 times 3 times 3 times 17. Great. And then what you do is you kind of collect all the prime factors together. You see the 2s matches, the 3s match, the 17s match, and the 3 and 5 don't. See, knock those off. Then you re-multiply all back together and it's equal to 102. And they sell it like it's a game, like it's a guessing game. It is not a fun game. But we're going to talk about this problem is much harder than you might imagine. So it's called prime factorization. And it's a hard problem. And I'm going to define hard here as being given the size of the problem, the amount of time that it takes to solve it is way beyond what you'd ever thought it would be. So, but it's easy to write a program to solve. It's easy to write the program. The question is when you're going to get to the answer. So for example, this one, if you're just trying to figure out, like, hey, what is 5? What are the factors of 510? I just write a big old loop. And I could just keep dividing by all the prime numbers till I get them all out. And here's one to try. And I made it for you just this morning. I promise this is a product of two prime numbers. And you can take a picture of that, if you want, or remember it, because I'll give you a nice present if you can figure out what those two numbers are. And actually, this problem is so hard, will actually be remarkable if you're able to do this. But if you look at, if you're to just do that program that I've talked about up there, just write a loop and try all the prime numbers. If you just look at, irrespective of how much computation it takes, if you just took it like how much electricity, like the electric bill of the computer to run that entire program to try all those things, it would exceed the world domestic product by a factor of 10 to the 44. This is what we're calling a hard problem, OK? So you get back to here, and what's kind of interesting about this, this isn't actually the problem that they asked Natalie to solve. They didn't ask her to factor these things. They actually asked, what's the greatest common divisor? And there is a fun solution to this. And to go find that out, we'll travel back in time to Alexandria at 300 BC. And there's this mathematician, Euclid, and you might have heard of him in terms of Euclidean geometry. And he is a mathematician, but also a lot of the mathematics that we knew, even up to, say, like the 12th or 13th century, were basically distilled by this guy in a giant book. He's basically going around and getting all the mathematics of his contemporaries. And so it's ancient, ancient math. In fact, this book that is here, Euclid's Elements, it was mass, I mean, he wrote it back in 300 BC. But it was basically the second book to be mass produced after the Gutenberg Bible. And all the way up until about 1910, if you wanted to be considered an educated person, at least in the Western world, this was a book you would have read at one time. We don't read it anymore, because it's so fundamental to mathematics that are in here. It gets into our public education system. These are things about what is a prime number, Euclidean geometry, Pythagorean theorem, and so forth. But here's a fun one. The greatest common divisor of two numbers is the same as the difference between those two numbers. So if we go back here to what Natalie was doing here, and this is what we had to do before. We had to guess this thing. And that was the hard part, was guessing what the factors were. But if instead, we can just take the difference of them, and we're sure that the greatest common divisor is always the same, we can do this. So we first subtract these two numbers, we get 204. What Euclid is saying here, these all have the same greatest common divisor. Amazing. So then we can just subtract them again. These all have the same greatest common divisor. We subtract them one more time, and we just get the same number out. Nice. And that took no work. So you're talking about the gross domestic product, 10 to the 44 factor. This just flies right through. And what's remarkable is this algorithm is still in use today in a place. Let me take a drink and you can guess where. On the internet, it's hard to find a modern map of the internet, so excuse this, 2006 map. But the internet's everywhere. And where this is used, it's actually used so many places. I just want to give you an example of like, this is a thing that you encounter every day and are probably encountering right now if you're surfing and even your phones are probably doing updates and getting at this thing. This S that sits on front of a URL, this HTTBS, it stands for Secure. There is a cryptographic algorithm there that I'm not going to get into, but involves this kind of stuff, like giant primes and all kinds of great numbers here. And the difference between not having that S on there, it's the difference of being at the coffee shop and somebody listening in on your Facebooking or whatever you're doing, or even worse, injecting something into the internet stream to be like, oh, what was your password again and getting it back? Two, getting a kind of communication that if somebody's trying to monitor on the internet, it's so remarkable. It'll simply look like random noise coming out of your computer. And random in the sense of statistically testing for random is looking as random. This will be relevant. We're gonna get back to all this stuff later in a way that will be surprising. So in a real sense, this is the mother of all algorithms. It's the oldest non-trivial algorithm that we still use today. And we're definitely gonna see it again. Okay, oh, that's the first hard problem I was talking about. Okay, great. So the next hard problem, at least topically in terms of the internet, there we're gonna talk about here, Bitcoin. And the reason why we're gonna talk about this, actually how Mooney asked me as a favor, he was like, do you mind taking some of your talk to talk about this? If you don't know how Mooney is most famously one of Gretchen Daley's postdocs, probably the most famous one. And specifically the reason why he asked me to talk about this is he formed me in this article from the BBC, about 50 cent, finding $8 million worth of Bitcoin. He's like, can you just explain this thing and sorry, let me just take, this will be weird that I would otherwise be talking, let me just take 30 seconds to talk about how this happened, okay? So back in July 2014, how he was trying to buy this book online, Animal Cognition, some kind of evolutionary psychology text. And look, we have all had an autocorrect mistake in our lives. And instead what he ended up buying was the album, Animal Ambitions. And it just coincidentally was the time when 50 cent was like the first really big musical artist to accept Bitcoin payments on his album. So that's just why we're talking about that. And just so you know, 50 cent he's in a bankruptcy hearing now and so when this article came out the bankruptcy court's like, wow, you have $8 million. And he's like, no, actually, I don't have $8 million. Okay, okay, so let's get into what Bitcoin here is. So Bitcoin, you have probably heard about it, but if not, I'm just gonna give context for what it is. It's, they call it a cryptocurrency. And the premise of it is a kind of currency that is outside any control of any particular government. Like it's a really interesting problem. Like how can we have two parties that don't trust each other, maybe don't even know each other and guarantee that they've exchanged some kind of currency in a way that they cannot lie to each other. Everybody else can see and you can maintain anonymity. That's kind of the premise of what Bitcoin is. The other reason why I talk about it is because it just recently got very expensive. And so this is a plot over time of where it used to be worth like two bucks. And now it goes up and down. I actually don't know what, I snapped this the other day so it was at $8,000. I don't know, it'll be up there. Thousands of dollars for a thing that seems where it used to be kind of worthless. There's a, there's another use of this. I was talking to my friend Rob as an economist here. It's also that like, as soon as you get into something that's some illegal activity, this is Bitcoin is the thing you want to be able to do. And the other thing I want to talk about it, this recently has come up a lot in, my friend Lisa's forwards me these articles all the time of like Bitcoin is going to save the planet. And so I think it's topical about why we talk about it here. And it's also a really interesting problem. And they'll lead on to the thing we'll finally get to. Okay, so I think a good way to talk about how Bitcoin works. Let's just talk, let's just go through what a transaction would look like. So the heart of Bitcoin is this, it's called the blockchain, but I think it's better to think about as a ledger. A giant book where every transaction of every Bitcoin ever has been recorded, that's public that anybody can look at at any time. And at this point it's, oh maybe somebody knows, it might be like 10 gigabytes big or 100 gigabytes big or something, but you actually just page through it and you can just see like here, this person transferred this much Bitcoin to somebody else. And I had a thing on here. Yeah, every transaction's recorded here. And it's a very hard problem to make a fraudulent entry. Hard problem like the running through the prime factors earlier. Then every computer that's on the Bitcoin network, it also is like just the nature of the way that this system is set up. It incentivizes everybody who wants to join it, that like you should participate in it or the whole thing just doesn't work at all. And so you're just like, and we're gonna get to this concept a little bit later too, but I think this is a really interesting one. So let's just walk through what happens here in a Bitcoin transaction. So how it comes up here, and the transaction is like very simple. The first thing Hal has to do is prove that he owns Bitcoin. And the way that he proves it is he points at the ledger saying like here, look here's where I was transferred some Bitcoin. So obviously I have it. And I wanna send it to Caroline Records. But if you remember what Rob was saying here before, here's the other part of Bitcoin. Is you don't have to be Hal and you don't have to be Caroline Records. You can be like Secret Man one, wants to transfer to Secret Man two in a way that isn't just anonymous and random. It's anonymous in a way that only Hal could ever prove that's him. Nobody else could come up and be like, actually that was me at that time. It's really interesting concept. So once you get that, that's actually the transaction. Secret person has this Bitcoin, wants to transfer it to this other secret person. Then some computer over here, the payment processing thing for Caroline Records will come along here. They'll check the ledger to make sure that works. And they'll be like, that's your transaction. There you go, you did it. And then Hal and Caroline Records can go do their business. But there's this one other part. So when they say that the transaction was good to go, they didn't write it into the ledger yet. And this is kind of the heart of Bitcoin. And this is the whole reason why there's a network of computers around the world. So as soon as a transaction goes out, there's this race to solve a really hard mathematical problem. You'll see it's not an interesting mathematical problem the same way that Natalie's guessing game was not an interesting or fun game. But it is very hard problem. So here's the first thing you do. So I'm gonna like wave hand stuff. And you know, as you're watching this, like you'll get the sense of like, oh, I kind of get what the numbers are doing and back and forth. And we're gonna tie this very at the very end into a software platform for the natural capital project. So I don't know, just let it flow over here if you like it, let's just see how it goes, okay? So the first thing we do, here's this transaction. Remember we saw this before? You can just take my word for it, we can turn this into a number. And there's lots of ways to do it. Just take every letter, sign in a number and stick them all together. So it's like literally a literal number is the thing you're getting at. Then there'd be a computer on the Bitcoin network who does this. Pick a random number and just add it to it. Just whatever random number you wanna get. And then you're gonna calculate a hash function. We're gonna talk a lot about this in the next couple slides. And you get this very weird, that's actually a number two, it's using hex notation, but another weird string out of it. And then here's the rule, here's the Bitcoin thing. Hey, the thing you just calculated, does it have 18 zeros on the front of it? Very specific, 18 zeros. If no, just try it again. Try a different random number. Try it again, like a loop. And if yes, great, you won. We're gonna write it in the ledger. Why would you bother to do that at all? Because there is a reward. You get 12 and a half Bitcoin. And that's worth about $100,000. And so there's a financial, well, I mean it used to be worth $100,000, but it is now, and that's why there's so many computers on the network now. And the other thing is like, well, how often does that happen? It actually is tuned, the whole network is tuned, so it happens every 10 minutes, no matter what. If more computers go on, it still takes 10 minutes. If more computers drop off, it still takes 10 minutes. So it's this chunking problem that will never end and will always be driving towards this kind of thing. Okay, I don't see what popped up there. Oh, probably that green thing. Okay, we're gonna talk about this part here because we're gonna use this in our natural capital software platform a little bit later. Okay, so to get to the hash function, we traveled to another place. So rather than now examine it, we traveled to the National Security Agency of the United States. I don't know if you ever thought about what their building looks like, but like black glass and steel and like miles of concrete around it just seems just right to me. Just really beautiful stuff. Okay, so Bitcoin uses a hash, it's called the SHA-256, stands for Secure Hashing Algorithm as published in 2001 by the National Security Agency. In general, what a hash function does is it takes some kind of input and it mixes it up in a way that's always consistent. So if you give me the same input, I will always give you the same output, but it's gonna be, that output's gonna appear to be random. Again, in a test for randomness, if you're looking at the values that are coming out of this function, it will appear to be random numbers, except it's not, because so long as you get the same input, you get the same back output. But the premise here is it's so, so, I mix it up so, so much, even though you know the steps I took to take it, it's gonna be very, very hard for you to figure out what my input was. And so in some sense it's like a really complicated shell game, you know, like where you put a thing under a cup and then you mix it up a whole bunch, except you can imagine like they're always gonna mix it up in the same way and then the ball is somewhere else, except it's more like this, except this is only like, I mean, whatever you wanna take from this, this is only like 10% of the algorithm on top of it. And you can see there's like 61 different rows that don't happen to be in there. This is kind of a big, like this is the thing the national security agency like, look, here's what we made and you should use this as your hashing algorithm. And we do. And I just wanted to, oh yeah, and Euclid, this guy is all over in this stuff. Like just that ancient mathematics is just deep embedded in this stuff all the time. And I just wanna mention this, Becky asked me to say this. This is something called a one-way function, meaning it's very easy to calculate one way, sometimes kind of trapdoor function. We don't actually know if these exist or not. There's been no proof of it. So whatever you wanna take from that, but it's a really, I mean, you're already factoring this giant prime number so you can figure this thing out for me later too. Okay, so this is what it looks like in practice. So it's coming along here and there's a computer here on the network. Here's how's transferring you add 123 to it and you call a hash on it, you get some random thing out here. No 18 leading zeros. So we tried again. No 18 leading zeros. So we tried again. Oh, look at that. And then you get your $100,000 here. Okay, so let's just see. I wanna get into like how hard this problem is. I don't know if I'm gonna have time to say why I have the grand canyon here, but a lot of computers are declining. Oh my gosh. Dang it. Let's cruise to the end. Oh gosh, I'm so sorry. What do you wanna do? By four minutes, do you mean four real? Here, you talked about it. Okay, so the total power of the Bitcoin network, 256 times the top 500 computers systems in the world, except that was in 2013. And if you look at it now, the total hash rate, like how fast those things are going that I just flew on the slides, you're getting 27 quadrillion of those a second across the entire network. And that's like some really, really small fraction of the top supercomputers in the world. In fact, it's an insignificant fraction of their total computing time. And it's a thing where like if you toss these things right into the computational network, it's not even gonna significantly reduce the time of when anybody was gonna get that $100,000 reward. And if you wanna get angry about this, you can look at this of like what are those computers doing? They're doing things like modeling Earth's weather systems to predict storms, to figure out if we need to evacuate a coast or not. And the holy grail of weather prediction, they're thinking to go all the way out to maybe 2030. We might get a supercomputer that has like this much computational power, and then we can model everything and have like a really good prediction of weather patterns up to like two weeks out. That's like less than 1% of the Bitcoin mining network that's just sitting there and solving this, just irrelevant problem. And if you're gonna get one more thing upset about, how much energy does it use? Here's a recent estimate of it, and you can visit this website, and it's interesting to see how they predict that. But it's basically 1.3 of the total US energy usage. 5 million US homes could be powered by it. And even if you just engage in it and just use one transaction, the electrical energy that's used to verify that transaction, it can power 26 homes for a day. This is the largest coordinated computer system ever built. And the solution to it is just an artificial constructed problem. And you think, what's left? Someday Bitcoin will lose its value. This always happens to human constructs. And all we'll have in the end, we'll just burn a lot of electricity. No interesting problem will have ever been solved, even though we've solved, I mean, I could really not even know how much you're gonna get up on that, okay? Okay, so let's go here. Okay. So after a decade of work, there's a question I think we feel comfortable answering. What services does nature provide? And here's some ways that we do that. One way is just through expert opinion. You get a bunch of scientists who know what they're doing. They've been trained up for years. They all have PhDs. They work on a particular project. They work there a few months, if not years. And you get a very high quality, but I would argue expensive result because it takes so long to just get humans trained up to that stuff, right? Another way you can approach this is you can just develop new science. How does water fall down? And how do trees make less water or more water, depending on how you wanna look at it, or protect the coast with mangroves and so forth? And then this is core science. So then you don't necessarily, what you could potentially do would be like, well, I care about a coastline, so I'm gonna go read a journal article about how natural capital helps that coastline. And that's still a good quality result, except it's still a lot of effort because you have to digest a science article and get that stuff in. And then here's what we try to do with software. We try to take that stuff. We try to abstract it into a language where practitioners are actually speaking it, usually through GIS maps and so forth. So long as you have a map of your area and biophysical data, then we can give you probably a pretty good result with pretty low cost. And our platform looks something like this. In the middle there's a bunch of models. We have a big computational backend that's super hot and goes fast. We have API that sits on the front that anybody could develop their own software on this. In fact, we go to develop, we in fact develop our own software. We develop, invest off our own abstraction of what this software is. There's data that fit outside and then sometimes we expose the things on the computational backend. Like you wanna delineate a watershed really fast, we can expose that to you. And is there anything wrong with this? Not really. But let's look at this. So this is a project that Becky's been working on. Becky and many other people have been working on for the past nine months of like, well, let's do this at a global scale. Not like let's get really coarse data, but let's get down into it and get the best data we can get and the most computational power we can get. And let's see how it's going. So I'm definitely not gonna talk about the science in here. But what I wanna talk about, I wanna give you a sense that the work that we did here was very computationally complex. And so much so, I mean the first thing, it's just a lot of data. I didn't even know that your home internet had an upper limit of how much data you could download and I found out this way. So the first thing, just take rasters, just stack them, do multiplications, all that stuff. That's what our PyGeo processing package does for you really well and it's in a super fast. You will not get a memory error. Another one is it's a little bit different problem. Here's, you know, Florida. There's a, so we drop a whole bunch of points all across the world. We walk up all the coastlines and then for each point, we have to figure out like where the wind is hitting and if it's near habitat and if like what's the slope gonna be like and who knows what else. And there's, you know, hundreds of thousands of these points but they all stand alone and we have another computational framework here to help with that. If you have a problem that stands alone, like you can calculate one point over here and one point over here and they don't talk to each other. What this thing does, if you program in it, it'll just like toss those out to a bunch of CPUs and later on computers and it'll just rip through those problems. And it'll also remember if you've done it before. So if you make one little change, you can rerun the thing really fast or if you just hit control C or your power goes out, it'll pick up where it left off. And the last one here with water, we basically put both of these things together and it was great. So now that we're done, now that we're done, here's like questions. I'm only using this as kind of an example of like you have probably encountered this before where we might be looking at a map. These maps are beautiful. We might go along and be like, hey, what were the data we used to run that thing? Or, you know, we have a bunch of scripts like which version of it went with which map? Or even maybe years later, like who even ran this thing? And these are really hard problems. They're not hard problems like prime factorization. These are hard problems about human nature. Like we try to do a good job and we write things down. We make readme files. We make design documents. We make metadata and still, it is fundamentally this is always a hard problem. And so what I'm gonna do here, I'm getting into like, this is the thing that we are working toward and is a metaphor I'd like you to like think about a letter that you send in the mail. So in order to get a letter to work, you gotta put an address on it and you should put a return address on it. And once you do that, it gets sent. And then the remarkable thing is this letter knows where it came from. It's just an inherent system of like, how mail works if you choose to follow the rules. So what if data were like that? What if you could just, after the fact, not like go find a file that you read somewhere, but what if the nature of the data itself remembered where it came from and how it was computed? And if that were the case, all this stuff goes away. So let's do that. We're gonna build a software platform based on all these things we just talked about. We're gonna start here. There's gonna be a couple buckets. One is all the algorithms we're ever gonna write or that might even exist in humanity. This is not an actual bucket. This is just a conceptual thing. And how are we gonna name all those? We're gonna just have unique identifiers for them. We're all that SHA-256 stuff we went through. This is how we're gonna do it. Every algorithm that could ever possibly be produced by the human species can have a unique identifier and we can store that in a table for as long as we wanna do that till we end. Same thing with data. All the data we ever collect, we can have a unique identifier for that with same with this SHA-256. Here's this computational stuff. I told you this was good. You can trust that. And here's what's gonna happen. You're gonna have a user that comes along here and is an author. And this author must have an identification. But as we know, we need not necessarily be public, but it has to be something unique. So this is the same way that Bitcoin uniquely identifies a user. You can have a user and how can be like, yeah, that's actually me. Or do you mean that I need to wanna say it at all? What how's gonna construct here? We're gonna call this thing an EcoShard. And all it's in it, in fact, I'm using this name very specifically to invoke some imagery. It's just gonna say use this particular data and use this particular algorithm. And that's it. That's all how's gonna write. And it goes into a global ledger, just like Bitcoin might. And it gets passed to the EcoShard engine. And with this thing, this chews it up. Goes down and goes, I know what that data is. It's a unique hash and it rips that out. And it says, oh, I know what that software is. It's a unique hash. It rips that out. It's gonna pass this thing over here to the computational core. This thing's gonna come out. You're gonna get this result out and great. And the reason why I'm calling this an EcoShard, I wanted to invoke this imagery of a shard of pottery or something beautiful that's fractured, that's part of a whole that a story maybe isn't totally told on it. And even like this, if you could imagine picking up a piece of pottery and just flicking it with your finger and maybe like hum or it has a certain noise, just the same way we're gonna be able to take the result here. We're gonna be able to do a SHA-256 on it like bong and it's gonna have a unique identifier and it's gonna go back in here. And suddenly there's this complete loop where the map itself doesn't have any information written into it about how it was created, but the nature of using this system, which is not complicated, we'll always go back to this and we'll always know where we came from. I am gonna cruise through this fact and you can read it later and I just want to get to a final point. Can I search the EcoShard engine? I do wanna talk about, you'll be able to search by author, you'll be able to search by data, you'll be able to search by algorithm to see where things go. Okay, so this is an attempt of what we're trying to do. We're trying to unify all this growing body of data and algorithm acknowledge about nature and its value. And this is the last hard problem that I talked about. This is the one that we'll be solving together. Thank you.