The Think 2018 Science Slam Part One

in a hotel room. I've got an air conditioner blasting at me not far away, because I am in Las Vegas, a desert. There's no producer here, there's no studio here. It's just me and a recorder and a microphone and a song in my heart and a dream in my brain, and the A C just shut off so it'll be a little more quiet now. But the Think Conference is

a pretty big deal. It's a hundreds thousands of people all gathered together learning about uh top topics in computing, getting education and workshops involved in all sorts of different things.

And last night I went to a science slam at the conference where we were treated to five different presentations by scientists that are related to IBM Research, and we got to learn about some breakthrough sciences and some scientific work that was pretty pretty interesting, and I wanted to kind of report on that and what I saw and the sort of implications that the presentations have for us,

as you know, human being type folks. Well, the presentation was opened by the Director of IBM Research, Our Krishna, who came up on stage and sort of set the ground explained what in brief the five different science breakthroughs we're going to be about. There was a lot of overlap between them, so expect that when I get into the nitty and the gritty. And he then introduced the actual MC of the evening, Jamie Garcia. Now, Jamie Garcia

is a polymer chemist. She works on lots of different projects, including ways to figure out how to break down long chain polymers so that you can get rid of plastics that would otherwise pollute the environment. So an important work. It was pretty interesting hearing her talk. She talked about how her specific work is very tricky because if you want to create computer models of long chain polymers, it

requires a lot of processing power. The bigger the molecule is, the more processing power it needs to simulate the interaction of molecules properly because the computer is trying to keep track of all the little sub atomic particles. All those little electrons have to be modeled and simulated, and as you add atoms to your molecule, creating these long chains, it creates exponential more work for your computer. So she has a vested interest based upon what she studies in

the advancement of computer sciences. Then she began to introduce the actual presenters, and each presenter had about five minutes to present his or her study her work and explain what they were there about. And it was really interesting

because there was a it was a good variety. There were two men and three women who each got up on stage, and they all were trying to explain to us the importance of their various projects and what what stage they were in, why we should be interested, and they all had very different styles, and it was very interesting to see these different uh approaches of present presenting this information to a general audience, because you know, a lot of people in the crowd were not scientists or

data analysts or anything like that, so it was it is a challenging thing to have to communicate science to an audience where you're not entirely certain what their background happens to be, but they all did a really good job.

The first one to get up was a guy named Andreas Kind, who is an expert on uh cryptography and blockchain, and I mentioned a bit in the preview episode the kind of Overview episode that blockchain was going to be one of the topics that I was going to look into here, and I've got a little bit better grip on what some aspects of blockchain are all about, and

it really blew my mind. So Kind came out and mentioned that almost everything that we encounter, whether it's a product like a designer bag or shoes or car parts, or even drugs like medical drugs, has been copied at some point, and he cited a number that in counterfeit

goods were valued at one eight trillion dollars. So you're looking at a massive problem across multiple industries where you could encounter copies of stuff being passed off as if they were the real thing, and in some cases that just ends up being kind of a a slap in the face to rich people who are buying luxury goods.

You know, it might be hard for you to feel any kind of sympathy for that, because who's going to feel badly about a billionaire walking up to a play a said buying, you know, some sort of ridiculously expensive wallet, and then finding out three months later that it was actually a cheap knockoff that was being passed off as the real thing. Most of us probably wouldn't lose very

much sleep about that. Although that is a legitimate problem, and obviously anyone who has built their business on creating luxury goods of any kind has a really really powerful interest in this. They don't want their products to be copied because that obviously undervalues what they do. If no one trusts that the thing you're selling is actually from you, then you're not going to sell very much of it.

But there are other issues as well that affect, you know, the common folk like myself, where even if we didn't feel sympathy for the rich people who we're buying these luxury goods, we may feel very strongly about these other cases. So Andrea's kind gave another example about car parts, and specifically he was talking about breaks, but it really could

be any car part. He said that in certain parts of the world, depending upon where you are, you might discover that up to forty of auto parts in the aftermarket are copies. They are not the legitimate part that they are, you know, claimed to be, So that could be a huge issue. If perhaps you were getting new brakes put into your vehicle, your old brakes are wearing out.

You go to an auto mechanic shop, you pay a certain amount of money to have brand new brakes put in, but you may not be absolutely certain that the brakes that are being put in are the real deal, like they are actually made by the company that you are told made them. It may be that they were cheap knockoffs, and thus they may not perform at the right level, and they might fail more readily than a real set

of brakes would from the actual manufacturer. That the could be life and death, and that obviously is something that affects anybody. It's not just the people who have money to burn, it's all of us who may rely upon this sort of thing. And even if you aren't someone who drives, and you're you're not too concerned about auto parts and whether or not they're the real thing, or maybe maybe you think you know, I'll risk it, I'll be careful and maybe my skill as a driver will

counteract any downfall of the parts there might be. It's still not that easy to just kind of walk away from this, uh. As Kind pointed out, another big issue is medical drugs. He cited a case of a blood thinning drug and some researchers who discovered that some instances of this blood thinning drug had up to filler material

in them. In other words, there were instances of this drug where it had been replaced with non active ingredients are even harmful ones, and this obviously can cause medical problems. It can lead to uh serious issues, even even the death of the person who is taking the drug potentially, and that of course is absolutely unacceptable. And again depending on what country you are. In some countries, you you are not certain about the origin of the medication that

you're getting. In certain countries, you know, if you're going to a reputable doctor and reputable pharmacist, you're in pretty good shape. You're reasonably certain that the medication you are taking is in fact what you were told it would be. But in other countries where the medical industry isn't as as well established and maybe not as well funded, then

you can't necessarily be certain. You may be taking drugs that were bought at a discount, and it turns out that a lot of them just don't have the active ingredients they're they're supposed to have, and it's not going to be an effective treatment for you. So all of these different examples were problems that kind was pointing out that he said blockchain could help solve, and that raised a question. Guys, They're still more to come about the science Slam, but before we jump into the next segment,

let's take another quick break to thank our sponsor. Blockchain is a digital construct, right. Blockchain is math and code. It's something that exists on computers, but not quote unquote in the real world. There's no physical presence of blockchain.

With cryptocurrency. It's easy to understand more or less at a high level why blockchain is effective because every single trans action that you make with cryptocurrency becomes part of this blockchain that you cannot alter, at least you cannot reasonably alter it without having to rebuild the entire blockchain from that point of transaction moving forward, and since the blockchain is constantly getting added to essentially every ten minutes

in the case of bitcoin. Your computer is never going to be fast and strong enough to build back that chain and uh and beat out the rest of the system so that you can alter a transaction and make

it look legitimate. So anytime you buy something with bitcoin, anytime you transfer bitcoin to someone else, that transaction becomes part of a block of transactions that the overall system will validate, and once validated, it joins the chain, and every future transaction that is made will contain a record that includes your transaction. Though this ledger of transactions is shared across the entire network of computers that are participating

in that blockchain. So if you have ten thousand computers in this network, all ten thousand can see that ledger and they can see the history of transactions that includes the one that you just did. So it makes it

impossible to really alter things. It's a very reliable record, and it makes it very difficult for you to claim that you didn't actually spend that bitcoin, and you in fact still have that bitcoin and you could spend it on something else, well the ledger will say otherwise, and that will mean that you will not have a legitimate argument.

That is fine for bitcoin transactions, but how can you use that same technology to ensure that real physical things are what they say they are, because if you could tag a physical object in such a way that it was related to a point of data within a blockchain, then you could trace the movement of that that physical

component as it moved through a system. As Kind pointed out, manufacturing in the twenty first century is really really complicated because you have lots of different entities that contribute to making all sorts of stuff, whether it's drugs or electronics or car parts. You might have factories and manufacturing facilities that are spread around the world that are all contributing toward this. Even if you're talking about food supply, you might be able to go to your local grocery store.

Let's say that you live someplace like San Francisco, California. You might be able to go to a market in San Francisco and pick up food that was originally grown in Japan. Well, you don't know the pathway that food took from the point of origin where it was grown or caught. Let's say it's fish. You don't know where that fish was caught. You don't know who that fish was sold to, you don't know where that fish was processed. All you know is where you bought the fish. That's

your last point of contact. You don't know anything else

about it. But with blockchain, if you were able to somehow physically link the fish into a blockchain operation, you can actually look at that chain of events and trace back every single point of contact that fish had all the way back to where it was caught, and you would be able to verify that in fact everything was UH was safe and healthy, that it didn't pass through any hands where there might have been contamination, or at least it decreases that, and if there were contamination, you

would be able to trace exactly where that happened because you would be able to look at the chain of events and say, well, according to this, it must have happened at this processing facility, which would make it much easier to UH to do inspections to make certain that everything is on the up and up. But how do

you actually tag that? How do you have some sort of record between the physical object and the digital record, Because if if you don't have some way of verifying that the physical object you're looking at is in fact the same one that's in the digital record, you can't be certain that the blockchain is an effective uh list

of transactions for that specific physical object. If I'm looking at a pair of shoes and the blockchain tells me that the pair of shoes is absolutely legitimate because I can trace everything back, But it turns out the blockchain is a record of a different pair of shoes, I don't I don't know that because there's nothing that's tying the physical object in my hands to this digital record. It doesn't do me any good. So here's the problem.

How do you How do you link those two? Andrea's kind talked about a way of doing that using something called crypto anchors. These are anchor points on physical objects that can link them to digital ledgers like blockchain and uh. The whole idea is that you have to have some sort of encryption, physical encryption that you can attach to those physical objects in some way that would allow you to verify that, in fact, the chain of events is exactly what the digital record says it is. But what

does that even mean? Like, what what sort of physical crypto anchor could you come up with? He gave a couple of different examples. One of them I thought was incredibly interesting. He talked about a malaria test. Now, if you are in another country and you need to take a malaria test, you've been bitten by a mosquito, and you want, obviously you want to make sure that you are testing negative for malaria. You want to also be certain that that test is legitimate, that it came from

a trusted source. So he showed off a type of malaria test this piece of paper that had very tiny, little colorful dots on one part of the paper, and those colorful dots represented an algorithm, a code. So first you would look at those colorful dots and you would make certain that it was the right pattern, the right colors, and you would have a blockchain record of this malaria test that would correspond to the code that you're looking

at the physical code. So as long as the physical code on the malaria test matched the one that was on the blockchain, you then could feel confident that, in fact, this test is a legitimate one. But you say, what about counterfeiting? What would happen if someone was able to look at the blockchain, they were able to see what the code was supposed to be. They take some regular old paper that doesn't have a malaria test on it at all. Then they very very carefully placed those dots

in the right configuration, in the right set of colors. Well, the secret to this is that the actual code itself is the thing you see is only half of the code. When you expose that piece of paper to liquid, then the ink dots, which are on little bitty micro pillars, some of them wash away, some of them are revealed, and you have a new code that will be there in replace. It's usually in this case with the malaria test,

you would actually use a serum to do this. So only with contact with the legitimate serum would this series of dots change and they would change it to a new code. And that new code is also tied to the blockchain, so you would have a before and after. So after adding serum to this malaria test, you could then verify that in fact, the malaria test is a legitimate one, it's a legitimate source. Then you can actually use the malaria test on a patient. So that is

sort of the way to confound counterfeiters. If you think about it, it's not that different from like a special water mark on uh uh a unit of currency, or one of those elements where you have like a transparent panel that's inside a dollar bill or or some other unit of currency. It's one of those things that is

meant to to uh make counterfeiting much more difficult. So by having this dual layer of code that only gets revealed if you add serum to the malaria test, you have helped ensure that the the test that you you have in your possession is in fact an authentic one and it is linked back to this instance of the blockchain. And by having this unique kind of code for every single malaria test, you make it incredibly difficult for anyone to make a legitimate counterfeit or a legitimate a seemingly

legitimate counterfeit. And I thought that was really neat this idea of making sure using these physical uh components, something that you could then link to a digital record. It's really really clever. Now kind also talked very briefly, because he was coming towards the end of his his presentation at this point, about the use of micro chips. I was gonna say microelectronics, and I guess technically it kind

of is, but I mean they're super super small. We're talking about chips that are about the size of a grain of rice that have more than a million discrete components on them, so essentially like a million transistors on this little size of rice chip, or even the size of a grain of salt. These chips can still have very sophisticated components to them, including things like motion sensors

or or a transmitter sensor. And the idea is that these particular chips will cost very little less than ten cents to create a single chip, and that they are able to monitor and analyze and communicate and even act on information. And these little tiny chips would be able to analyze products and measure them in such a way to make absolutely certain that they are in fact um legitimate.

They would concentrate on what are considered to be unique identify irs or whatever the product is, whether it's the specific shape or the chemical makeup or uh any other sort of physical property that is considered to be unique to that product, that it would be almost impossible, if

not outright impossible, for someone to duplicate that exactly. You would have the chip program in such a way to detect that quality, and if the quality is not present, then obviously the chip would indicate that, and the chip would be the thing that's low. That's that's tied to

the blockchain. So as long as the chip is active on this uh, this product, whatever it may be, and that it's verifying that it is in fact authentic, and you can record the process of where the chip was from one point in the supply chain to the next, all the way to its end point, whether it's a consumer or a doctor or a manufacturing facility, whatever it

might be. You could then look at the blockchain verify that in fact, all the right steps were taken, it went to all the right points of contact, all the right processes were uh were performed on this product, and the chip is still verifying that the product that is in your hands is in fact authentic, then you have this ability to say, yes, this is the real thing.

And at least according to kind Uh, he said that using this you could end up cutting counterfeiting in half within the next couple of years, and then moving forward you could reduce it even further. So this again would help people be certain that the thing they got was completely legitimate and they wouldn't have to worry about whether or not the product they had was going to fail on them or be harmful to them or otherwise not be what they thought they were getting. So that was

number one of five. Still have some more people to talk about, some more ideas, more cool notions that were brought up at the IBM Research Science Slam. I think, but first let's take a quick break to thank our sponsor. The second person to come up was a PhD student named Chichilia Boscini and chi Chilia. She talked about lattice

based cryptography and I loved Chichilia's presentation. UH. This young woman got up on stage and began to speak about post quantum cryptography lattice based UH encryption systems, and she made it funny and engaging, which is kind of a magic power. She talked about how once you learn the fundamentals of mathematics, you can build all sorts of things, and that when she was a child, she fell in love with math. She thought math was fun. She liked

coming up with problems and solving them. She talked about a teacher who would use a story of a fictional old old lady who had an electric wheelchair, and a fictional old lady kept on getting into various problems. Various situations and the math problems were all related. To figuring out how this nice old lady in electric wheelchair could get out of a various predicaments. So it was this sort of this real world kind of of scenario to

frame all these very ethereal math problems. And Chilia loved it. So she went into looking UH into math further and studying it, you know, is her primary area of focus. She also talked about things like number theory, where you take integer numbers and you take prime numbers and their properties. And you may remember in the review episode I talked about how cryptography is very much reliant. Our current status of cryptography is very much reliant on things like prime

numbers and factoring. So to give you a quick update on what I said or a quick refresher, I guess on what I said. If you're looking at modern day cryptography, modern day encryption UH strategies, typically what you do is you take an extremely large prime number. That's a number that's only divisible by itself. It doesn't have any other factors, so you can't divide by two or four or seven or anything like that. It's only divisible by itself. But

you take a really really big prime number. I'm talking about a number there's hundreds of digits long, so enormous number, and then you multiply it another equally huge prime number, and then you get a product. You get you get something that is the product of these two enormous prime numbers that becomes sort of your key to encrypting everything. And encryption obviously takes other steps past this. You've got your encryption key, but then you've got to keep on

going and actually encrypt everything. But the point being that someone could get hold of that encryption key, but they don't know the secrets of how things were encrypted because they don't have the actual prime numbers that you used

to create the key. In order to figure out what those prime numbers are, you would have to start setting a computer program to taking that enormous product and starting to divide by various large prime numbers and making certain that the other, uh any any you know, reasonable or real integer number that you would get by dividing by a prime number was also a prime number. And this whole process is very very difficult for a classic computer

can take an incredibly long time to solve. We're talking years or decades or even longer to solve some of these really hard problems because computers are just going to do it sequentially. They're just gonna keep trying, and once they try one prime number, they might try a prime number and the result they get back is not a real number, or it's not a whole integer. Well, that that result has to get tossed out because it cannot

be an answer to your question. Or it might get a prime number and it takes this huge number, divides it by a prime number, and it gets a second whole integer. Everything is cool there, but then the second whole integer turns out not to be a prime number, which the computer also has to test, right. It has to make sure that the result is also a prime number. If the result was not a prime number, then that result is invalid and the computer has to keep going.

So this is a very very long process once you get to truly large numbers. However, we don't know for certain that this particular process is uh is truly a hard problem. There's no mathematical proof that it is a hard problem. There's circumstantial evidence that it's a very hard problem because computers are not very good at solving it, but that that alone is not a mathematical proof. Mathematical

proofs and evidence are two different things. I think I actually listened to a gentleman talk with this young woman yesterday, and I think he was getting proof. The word proof was was causing him to hang up a bit because he wasn't thinking mathematical proofs. He was thinking proof. I said, hey, I have evidence here, I have proof that this is a hard problem. It's not exactly the same thing. They're related,

but not the same thing. But she explained that with quantum computers, if you're able to quantize the information, and you're able to use quantum computers, because quantum computers use cubits, and if you have a sufficient number of cubits, if you have a quantum computer that's powerful enough, it could take minute it's to solve a problem that would take a classical computer years and years and years to solve

a classical computer to solve a problem like this. Like I mentioned in the preview episode, quantum computers are not necessarily great for all computational problems, but for a certain set of computational problems, they would be far more efficient assuming you had a powerful enough quantum computer. Now, she did say, there is some good news. We don't have

to worry about our encryption schemes falling apart overnight. That it took years and years and years to create a quantum computer with just a single cubit, and these days you're talking about quantum computers that are only at around the fifty cubit range, and you would need many, many, many more cubits in order to be a serious threat

to cryptography. That before we get to that point where there's gonna be more years of work and development to make quantum computers reliable and make sure they don't just decohere easily and decoherences when the quantum states all begin to collapse in on themselves. So if that happens, then a quantum computer turns into a classical computer, and then you're just back to where you were before. UH. You have a computer that's going to go through the old

process of trying to solve this problem. Quantum computers sort of solve problems in parallel. They kind of test all possible UH answers simultaneously because cubits can act in superposition and be zero one and all technically all values in between. And because of that, you can run multiple problems and each bit is acting as either a zero or a one. That means you get every single potential combination depending on

how many cubits you have at your disposal. Right, so if you have fifty cubits, that's the equivalent of fifty bits, except the bits can be both zero and one at the same time. You can run any problem that would require fifty bits or fewer, and it will run all the potential solution simultaneously, and then will analyze those results, and it then assigns sort of a threshold of confidence for each of those results at the end of the process.

And technically usually whichever result has the highest level of confidence is most likely the correct one. Uh, this is kind of it sounds like I'm dancing around things, but that's because quantum is weird. Quantum does not answer in certainties.

Quantum answers in probabilities. So it may say that I'm confident that this particular result is in fact the solution to your problem, and you have to decide where's your threshold, where is the what is the cut off for certainty that you need before you go forward and say, all right,

this is our result. But if you do have a quantum computer with a sufficient number of cubits and you're presented encrypted information, then you could potentially do the equivalent of a brute force attack of figuring out what that encryption scheme, what what numbers it uses if you have this this quantum computer, because it can run all those calculations I talked about earlier essentially simultaneously and then give you the most likely answer, and then you would have

essentially the keys to the kingdom. And it's it's kind of like thinking about someone has found the perfect way to create a skeleton key that fits every single lock that's ever been made. Once you do that, then locks are useless because someone could get hold of one of those skeleton keys, they can get access to everything. So this is why people talk about quantum computers as being the end of our current encryption strategy, because it won't

do you any good. If someone actually has access to a quantum computer, they can decrypt anything that's out there that's using that particular approach. So she said, well, eventually we're gonna get there. We're not there yet. We're at

fifty cubits. And in order to get a quantum computer stable, you have to keep it at an incredibly low temperature to maintain that quantum state, and you cannot interfere with that machine at all, because if you do, then the coherence dissipates and you're back to having a less powerful classical computer. But eventually we're gonna get to a point

where they are going to be powerful enough. So we have to start thinking about what the next evolution in cryptography must be in order to protect our data from the quantum world. And this brings us to the lattice based approach. I had no idea this was a thing, and it blew my mind when she talked about it. Classic cryptography, we're using large prime numbers. With lattice based approaches, you create a a applotting system. Imagine a two dimensional grid,

so just grid paper. You've got that grid in front of you, and you take a point. You make a point on that grid, and the lattice problem requires that you find the points in a grid that are closest to that fixed point you've chosen. That fixed point you've chosen, it's called the origin. And so your job as a computer is, all right, let me find the points that are closest to the origin, and that will be the basis of my cryptography. If it's two dimensional. You could

do it yourself. You can just look at a piece of grid paper. You see a point that's on there already. You can quickly identify which points are the closest to that central point, and you would have the answer right in front of you. It's it's it's a it's not a difficult problem. What Chilia said was, imagine that you don't use just two dimensions. Imagine you use use more than two dimensions. Imagine you use one hundred dimensions. Here's the thing. We can't really imagine that we live in

a world of three dimensions that we observe directly. For if you argue about time being a dimension, although you cannot physically see time, you see the effects of time through the causality. But if you were to add extra dimensions, which you can do computationally, even though we as human beings are not capable of really imagining that, it would it makes this problem far, far, far more difficult to solve, perhaps too difficult even for quantum computers to solve in

an arbitrarily easy fashion. So if you do make such a problem a lattice based problem, where you know the answer and the person that you're sending information to also knows the answer, but no one else knows the answer. Then, even if they have a quantum computer, they can't just brute force figure out the solution to your cryptography strategy

because it's too difficult to map out. Uh. There's no mathematical proof that can simplify this nos, no sort of shortcut to the in destination, and so a lattice based strategy for cryptography might be the few sure of cryptography in general. I think that's amazing, and honestly I still only have kind of a semi basic grasp on the concept. But the presentation was fantastic and I highly recommend you look into it if this sort of stuff sounds interesting

to you. To Chilia made mathematics sound really entertaining and fun. Uh I liked her approach. I don't think I would find it nearly as fun. I think I would quickly run up against the very limits no pun intended for all my calculus fans out there, but the very limits of my understanding of mathematics, and then I would get frustrated. Because I loved math up through triggonometry, and then once you got to calculus, I I I hit a wall and I just suddenly could not get my mind wrapped

around those concepts. So I have a huge admiration for people who not only understand the concepts, but they are capable of communicating not just the general uh way that this sort of stuff works, but also their own interest

and enthusiasm for the subject. Alright, guys, I recorded the full episode about the Science Slam, and it turned out to be a little long, So rather than do a full epic length episode about the Science Slam, I figured out end this episode here and we'll pick up in the next one to talk about the other presenters who took the stage at Think Conference two thousand eighteen IBMS conference to chat about the science they've done and the work that they're looking at and how that has the

potential to really change our world. I am incredibly thankful that I got a chance to see these brilliant people speak and to to get some inspiration from them, because it was really cool to hear about stuff that usually I just read like a press release about, but to actually hear the people who are doing the research talk about it and to kind of convey their excitement about

their fields of study was really invigorating. So make certain you tune into that next episode to hear the conclusion of the science slam and what else I learned, because it was really cool stuff and I hope you guys find it interesting too. If you guys have suggestions for topics I should tackle in future episodes of tech Stuff, you know, once I get through with this little mini series over at the Think Conference, I'm gonna be right back to my normal recording schedule and uh, you know,

good old reliable Jonathan talking about technology. If you have anything that you would like me to talk about, maybe there's a particular company or a technology or a person that you think I should profile, or maybe there's someone I should interview or have on as a guest host. Get in touch with me let me know what you think. The address for the show is tech Stuff at how stuff works dot com, or you can always drop me

a line on Twitter or Facebook. The handle for both of those is text Stuff h s W. Don't forget to over to Instagram. Check out tech Stuff over there. We've got lots of behind the scenes photos and other goodies for you to look at. And also I live stream when I'm recording in the normal studio on Wednesdays and Fridays. If you would like to see me record this show live, you can go to twitch dot tv

slash tech Stuff on a Wednesday or Friday. The schedule is up on the page there, and you can watch as I record these shows, as I stumble my way through mistakes, as I restart, and actually doesn't happen that often. I'm mostly mostly what you hear is what I record, and you can chat with me. There's a chat room in there, and whenever I get to a break, I'm happy to chat with all the people in there. Sometimes you guys help shape the show and I really appreciate that,

So join in. Come on over to twitch dot tv slash tech Stuff, be part of the group. I look forward to seeing you and I'll talk to you again really soon for more on this and Bathos of other topics. Is that how stuff works dot com wh