#231 Ajay Prakash & Gavin Brennen: Qubit Protocol – Quantum Computing & The Coming Threat to Crypto

Transcript

[Music] this episode of epicenter is brought to you by shape-shift oh the easiest fastest and most secure way to swap your digital assets don't run the risk of leaving your funds on a centralized exchange visit shapeshifter IO to get started hi welcome to epicenter the show which talks about the technologies projects and startups decentralization and the global blockchain revolution my name is Sylvia central kochira and today we are gonna talk about a topic that I

think most people will have heard of but perhaps you know need a better understanding in order to fully grasp the potential of this technology and that is quantum technologies and specifically quantum computing you know we've we've been thinking about this for quite a while at epicenter and didn't really know how to address it or when it from what angle but while it's here in Singapore I was fortunate enough to meet someone who is doing work in in this field and told me about a white paper that had been written by his team about about the quantum attacks that were possible in Bitcoin and how we can

mitigate against those attacks and so our guest today our guests are a jape Rakesh and he is the CEO of the qubit protocol and Gavin Brennan and Gavin is CTO chief community officer at qubit and Gavin is also an associate professor in physics at Macquarie University in Sydney and so qubit is a governance protocol that is used to select and fund the best quantum technology startups in the world and so we'll talk about the qubit protocol but well we'll spend most of the discussion dissecting quantum technologies in quantum computing so getting a

high-level understanding of what these technologies are and how different they are from classical computing or binary computing and also we'll learn about how quantum technologies and quantum computing can affect the technologies and the specifically the algorithms that are used in Bitcoin and a lot of blockchain so hashing algorithms digital signatures lip the curve signatures etc so guys super excited to have you on today and to dive into this topic yeah thanks for having us - yes thanks for having us on alright so before we dive into the topic of Bitcoin and how Bitcoin can be vulnerable to certain

types of quantum attacks let's let's perhaps start with describing and and you know keep in mind here that some of our listeners are not necessarily engineers or mathematicians or physicists so you know think of it as like a explain like I'm five exercise here like how would you explain quantum computing to a five-year-old sure yeah so it's it's always good to go back to what you know like we think we know classical computing I mean everyone is used to using laptops or cell phones and those are all doing classical computing and if you were to zoom in on the transistors there it would be you know

involving voltages which you know some amount of voltage would represent a1 and some different amount of voltage would be a zero and all this stuff could be measured and you would just be doing lots and lots of operations flipping ones and zeroes doing things like adding that and doing NAND gates and or gates and things like that and of course all this happens incredibly fast billion operations per second and at the end you get out some some answer to a problem and I mean a quantum computer is also doing that kind of thing it has an input and an output an output is that you something you measure and you find

whether you've got an answer to your problem but using quantum bits so if you were to zoom down to really small sizes like the size of an individual atom or a single photon of light then the laws of physics really look different than they do in our everyday experience and you can have like instead of a bit which would just be up or down zero or one a quantum bit a qubit can be in a rotated superposition they call it of zero and one and it's really you can just view it as like a spin that is rotated away from

being straight up and quantum systems can do this they can you can have it with electrons with atoms with pieces of lights and now the exciting thing is that they're building technologies with superconductors where you have current that flows in one direction or another representing a zero or a one and you can amazingly get currents that has zero and a one so it's going some amplitude to be going in one direction and amplitude to go in another direction and that's a qubit and when you get a bunch of these quantum bits together you get a quantum register which is part of a quantum

computer and you often hear that quantum computers get their magic from the ability to do parallel processing so they can act on all the possible bit strings at once like 0 0 0 0 0 1 0 1 0 etc and that's true but at the ends you only measure one thing and so if you were just measured if you're just operating at all these things at once and just measuring one thing you would have an exponentially small probability of getting the right result so the real power of quantum computers comes in because you can in clever design of an algorithm you can make it so that the correct solutions to a problem are

reinforced adding up probability amplitudes and the wrong solutions are canceled out because in quantum mechanics you can have positive and negative probabilities amplitudes so you cleverly design an algorithm so that the right solutions are reinforced the bad ones cancel out and at the ends you get an answer that's much much faster than any supercomputer in existence today could ever do so that's a very brief description of how these things work just the way that I understand this is that in in binary computing a single piece of information can be either 0 or 1 or on or off in

quantum computing a single piece of information or qubit can be anywhere on an XYZ axis so essentially I think you know we would some people would describe that as a sphere and when you line these qubits up you can perform computations on them but like the the the thing that's unique about quantum computing is that because you're measuring quantum particles as soon as you measure them you essentially disrupt their state and so you can't effectively just look at it and measure it because the light that you used to measure it will disrupt the the quantum particle and so what you

need to do is you need to perform many measurements on this and then with this with probabilistic functions one can extract actual meaning or an answer to a function or an algorithm yeah that's right so if you imagined it like a spinning top a classical bit would be like the top can only be pointing up and down so it's like you know vertically oriented a quantum qubit as you say it could be pointing like anywhere on a sphere so it could be rotated but when you measure it you can only measure if it's upward down and and once you measure it it's stuck there it's it's in it's either up or down state not in a

rotated state and so the point is you want to design the algorithm so that you force all the the right solutions to a problem to be say pointing all up and then you'll know that those were the ones that had the solution to your problem okay and so why is it that this is so much more powerful than binary computing what what are is it about these quantum properties that make it so much more powerful yeah well so it to tell you the truth we don't know the full answer to that and actually proving that quantum computers are faster than classical computers is very hard in many

cases you can't so you know the big example that got everyone excited about quantum computers with Shor's algorithm for factoring and it it solves you know taking a big number and finding the prime factors quickly which it does this in polynomial time I think it's n cubed time where n is the number of bits and people think that the classical algorithm is is hard so super polynomial but we don't have a proof of that it could be that you know in five years someone comes up with a fast

classical logarithm or it could be it's already been found in the NSA is keeping it secret we don't know we do know that there are provably some cases where quantum algorithms do better but having those proofs is is it's actually kind of rare but we can say generally that quantum computers make use of some properties that you just can't get with classical physics and one of the key properties is entanglement which gives you correlations between qubits in lots of different directions that you use you can't get with classical physics and

another one is superposition and interference this is where you can have state of a system in a combination of being up and down with lots of these qubits so you can have something like you know 0 0 and 0 1 and 1 0 and 1 1 at the same time so yeah these these do the elements of entanglement and superposition and interference are all really new things that show up in quantum physics that give the power so give us an idea then of like where we are in terms of technology maturity with regards to quantum computing like if you were to compare it to it's like the

advent of binary computing like take maybe like the 50s and like the first computers the first like electrical computers electronic computers that were built like in terms of timeline like what's what's the maturity level yeah we're like yet in the 1940s so we're you know okay we're talking like you'll see press releases coming out from Google and IBM at talk about 52 qubits and there's a 72 qubit device that Google has announced however that device is actually now completely wired up and

working they do have the qubits but as of today it's it's not a completely working device so basically it's like having a regular computer with like a transistor that has 72 bits of of storage sort of memory or is yeah so you could use it for storage yeah mmm it but it does do some some quantum things that a classical computer couldn't do and the big push right now is to really prove that they can build a device that can do things that would be extremely hard for a classical computer

to do and there's a hope that we'll actually see evidence of that this year by the end of 2018 there should be firm evidence that they've got a device that does things that are beyond what our supercomputer can do right now and so one of the applications well yeah so in terms of advantages over binary computing quantum computing as is isn't necessarily better than regular computing for anything there are specific applications and specific types of computations that are better can you describe what those are at the moment there are some really cool applications of looking and I think it's

also important to understand that quantum computing is not just the only field in quantum technology so yes quantum computing I think some of the biggest things we're going to see some improvements in is gonna be things like AI and machine learning and I think Gavin can go into more detail about it but essentially quantum computers are really good at multiplying matrices upon each other which is essentially what AI and machine learning is doing and it will see disruption in all these other fields like supply chain management you know financial services anything that really is to do with AI and I think the way you can sort of think about quantum technologies is sort of like a backbone to all of these other fields and

industries that will be using it going forward not necessarily you are will it be sort of a forefront technology will be sort of like the Internet where people don't even realize they're using it but it has impacted significantly what they're doing and then there's other fields such as like quantum secure communication which we're sort of seeing a bit more maturity in where you can use qubits to essentially transfer information and if someone sort of eavesdrops on that qubit stream then you know simply because as you explained before when you try to measure these qubits they collapse into a state and as a result you can sort of see if someone's eavesdropping or looking into the information that you're sending so

that's the interesting field and then finally something like quantum sensing so some of the applications of quantum sensing could be using quantum tunneling so used to the position entanglement in quantum computers but you can use quantum tunneling where you can sort of measure the potential difference across a small small space and you can feed nucleotides through it which is what you know genes are made up of and as a result you could essentially read an entire a gene sequence using quantum tunneling in quantum sensors so there's a huge array of different applications but I think what we're going to see is it being sort of a backbone player to all of these different industries

interesting so we shouldn't expect in a quantum smart phone at any point it from what you're saying these technologies were sort of Oprah operating the back end feeding you know B state these models and so like improving our AI and machine learning and this sort of thing but not necessarily like in consumer products yeah I think unless sort of we come up with a way perhaps there will be startup that we fund that comes up with a way to apply quantum computing to universal Turing machines but until then we're not really going to see it being applied wide scale in laptops in in mobile

phones and things like that yeah and some of the examples of algorithms for quantum computers include quantum chemistry so finding the the best configuration of atoms for drug design or another big one is understanding the Haber process which is the describing the the chemical activation process behind fertilizer design so I think can't remember I think it's something like 5% of the world's energy is devoted to a fertilizer production agriculture feeding people and having a more efficient way to design this catalysis would be a huge

benefit to the world's economy and then there's of course code-cracking and another big one is materials design so like a very basic thing you might learn in high school is you know solving a linear system equation like matrix a times vector X equals vector Y and you want to find out what X is and quantum computers can do this very fast and that has tons of applications like you know designing our radar and routing of systems so it's it's actually quite a quite a few applications which I'm sure

we haven't thought of yet what about brute forcing so classical computer computer science problems like the Traveling Salesman problem yeah good question so you might think ok this is something we know is hard that's like like one of the first things you learn if you take a computer science course but actually classical algorithms are quite good at solving Traveling Salesman problem it's only like pathological examples in the hardest case that are exponentially hard in the number of cities the number of nodes and the good heuristics exist in fact you can get I think within factor of 3 halves of an

optimal solution using good class cool heuristics and it turns out that that kind of problem doesn't have enough structure that we know of to make quantum computers much faster so it's always good to you know tame your expectations not not all problems see a big speed-up with quantum and the Traveling Salesman problem is one where we generically would not expect some big speed-up okay so now that I think we've got a good grasp on quantum computing how it works what are the applications and how it's better

than binary computing for certain types of applications let's let's let's dive into this paper this research paper room so we'll link to those shown in the show notes and so the paper again is titled quantum attacks on Bitcoin and how to protect against them so this paper describes primarily two types of attacks so one attack on the hashing algorithms that are used in mining for Bitcoin mining and of course a hashing algorithm is an algorithm where one inputs data and the ashing hashing hashing algorithm will generate a hash or a unique set ru string representing the state as this is used in in Bitcoin mining of course ASIC processors used for Bitcoin mining do

nothing but create hashes or generate hashes and the other attack that you describe in this paper is on digital signatures so specifically the elliptic curve algorithm and the specific curve to Bitcoin so before we dive into the this this topic and and these two types of attacks could you tell us why did you choose to write this paper what was interesting to you about you know how quantum computers could potentially attack Bitcoin or other cryptocurrencies yeah it was the news so we just kept reading these news articles about you know quantum computers are gonna be the end of cryptocurrency and you know these these apocalyptic story is about how the

new quantum machines coming out are gonna ruin all the investments in Bitcoin and we decided no this is bunk we need to get some reality to this and so so yeah we took we took it to task and just you know put in some numbers and some thinking to it so before you wrote this paper there's no prior work on this topic like real scientific work not much no I mean there have been some papers that talked about alternative blockchains using quantum technology but actually there are things you can do that don't even require a quantum technology that can make sure that cryptocurrency can be

made safe against clone computers okay so let's get into it so first let's let's talk about the the hashing algorithm section of a paper so before we we dive into how hashing can be affected by quantum technologies let's just maybe dissect a hashing algorithm at a high level how does the hashing algorithm work yeah so it takes as input message and the message can be huge actually you know it could be an encoding of an image or you know audio even pieces of a video and it it applies a bunch of operations like you know you might represent that that input and

hexadecimal some big string of hex and the hashing algorithm will do additions and multiplications and bit shifts and permutations and it's it's actually a nonlinear function and at the output you get a string also in hexadecimal and that's called the hash and hashing is what's referred to as a one-way function so it's it's very easy to perform the hash like there's actually a funny video on YouTube if a guy just in pen and paper doing it applying the sha-256 algorithm with the hash it's very low hashing rate so it's very easy

to do but it's very hard to invert so you know given the output working out what the input is is very hard so it's it's called one-way what what's unique about this algorithm that makes it hard to go the other way at least with with a binary computer or with kind of paper yeah it's just it's just these these compositions of nonlinear operations that's are don't have enough structure for you to invert and I mean we get these things happening a lot I mean even in physics like you

have a chaotic system if you change the initial input a little bit the output is completely different and and that's that's the key to it it's just it's a physical process here a computation that's if you were to change like your input string by one bit the output is completely different so what is he what is the sort of classical attack on Bitcoin then in terms of cracking the hashing algorithm with a quantum computer how would one like walk walk us through the attack right so yeah so the the mining problem at least for Bitcoin is you applied two hashes to an input and the input would

include transaction information and timestamp and so forth about the block you want to verify and what you want to do is you want to find a nonce that is a certain string that you add on to that the transaction data such that when you do the double hash you get a certain number of leading zeros in the output and the number of leading zeros you want is determined by the difficulty in the Bitcoin mining available at that particular time so of course this computing power gets better the difficulty increases and Bitcoin mining demands that you have more leading zeros

that you need to find and since there's very little structure in this what classical ASIC devices do is they just try lots of different nonces they just you know iterate over all the possible announces they could do until they get the double hash that has the right number of leading zeros so it's just it's a brute force trial and error and the quantum attack would be to try to find a faster way to do that using quantum mechanics but the thing is because hashing has no structure there's no real structure that a quantum computer that can take advantage of to

get a big speed-up so the best attack we know of is just to use what's called a grover search so it's just a fancier accelerated quantum search where you're doing everything in parallel and so you're basically acting on all possible nonce values at the same time and it turns out that a quantum computer can search an N item database in root end time whereas the classical one would take n time so there's a speed up but it's only a square-root speed up which is not a tremendous amount and especially when

you compare the speeds of quantum devices versus Asics so there's like an ASIC you can buy it's does 14 tera hash per seconds and that's so much faster than what a quantum computer can do even for a small scale now like you know typical speeds for the computers that are being used at Google and IBM or like you know 66 megahertz and even if you imagine improvements in the future to the speed it's still gonna take quite a long time for them to be competitive even with ASIC machines today so the

attack then is not necessarily to reverse the hash that that a miner would have discovered that matches the number of leading zeroes defined by the difficulty but to really to out speed to outperform the miner in getting more hashes per second than a standard ASIC yeah so like an ASIC will just just do this brute force thing they'll try one nonce do the double hash do you get a good answer no try again other nonce etc what a quantum computer does is it acts on all the nonce values at once so you you store in the initial state the data

of the transaction for the block and all the possible nonce values at once and you act with your quantum computer to process that state - and it will reinforce the nonces which give you a good result that is giving you the leading zeros you want and it will cancel out the strings which give you bad results and at the ends you get an answer which is a correct nonce and that whole process takes root and time on a quantum computer but that's that's the best that we we know of to do for a quantum attack so your conclusion on on

this particular attack is that Asics would always be faster at doing this brute forcing then a quantum computer would be at attempting to do this massively parallel processing operation on you know all the nonces at once well know so eventually if you get a big enough quantum computer and if the difficulty is hard enough it will it will take over Asics it would be faster but it's just that the timescale for us to get those machines is pretty far so we did a kind of quantum Moore's Law analysis yeah where we we looked at how

the number of qubits has been increasing over the past few years and also how good those qubits are that's called the gate Fidelity's how well they're doing those gates and also how fast those gates are and you can find estimates for an exponential improvement in all three of those things and then we just did a forecasting given the current rates of growth with possibly improvements or slightly worse how long would it take you know a quantum computer to be a real threat to Bitcoin mining and yeah we don't see any any real threat coming

until maybe you know 2040 okay so it's it's not an imminent thing and you know of course these these crypto currencies that use use this kind of proof of work mining they could once quantum computers become ubiquitous what they should do is increase the difficulty just to make it so that you know it compensates for the speed-up okay so then in in in that scenario one would just need and it's like a lot more Asics attempting to to find the right hash but for presumably like a much much larger difficulty and then the quantum computer would still

wouldn't still be unable to outperform these eight-six yeah and eventually I mean if if quantum technologies really take off everything will be quantum like all the computers will be quantum and then like quantum Asics yeah exactly yeah okay and I mean I know the paper doesn't address this but are there any types of attacks that one could perform on alternative consensus algorithms like proof of steak for instance is proof of steak quantum resistant yes this is a good question and I consider this still research I don't think we have an answer to that there's been some work so the

coin qrl has some statements about having a quantum resistant proof of steak protocol but I haven't I haven't gone through that in detail or I haven't yeah I haven't seen the arguments that's convincing for that and yeah I think this is a very interesting research area also Cardno is using the aura Boreas protocol for proof of stake which is claimed to be secure against classical attacks so this would be interesting to understand the security

of that against quantum attacks this episode is brought to you by shape-shift the world's leading trustless digital assets change quickly swap between dozens of leading cryptocurrencies cluding bitcoin ethers e cash gnosis manera golem walker and so many more when you go to shape-shift dot io you simply select your currency pair give me receiving adders sent the coins and boom shape-shift is not your traditional cryptocurrency change you don't need to create an account you don't need to give them your personal information and they don't hold your coins so you are never at risk from a hacker or other malicious actor shape-shift has competitive rates

and is even integrated in some of your favorite wallet apps like Jack's so you can swap your digital assets directly within your wallet just as easily as putting on your slippers whenever you see that good-looking fox you know that's where shape-shift is so to get started visit shape-shift oh oh and start trading and we'd like to thank shape-shift for their support of epicenter so the paper describes this alternative consensus algorithm through proof of work which is momentum can can you describe momentum and why it's quantum resistant yeah so you know we were imagining well okay maybe you want to see if you can come up with something which is even more secure

against quantum attacks on mining imagine that somehow you got really fast quantum computers like a whole new technology showed up which made them work at tens of gigahertz clock speeds or 100 gigahertz so that it really really did become a significant threat to mining then you there's a way to modify the mining using momentum where you demand that not just that the the hash or the double hash of some string is below a certain number but also that the hashes of two subsets are equal to each other so your demands that it's the

hash of a string that includes like the transaction data of the block 1 nonce and a second nonce and you also demands that the hash of those that the first nonce and the second nonce is equal so you've sort of forcing some kind of structure to the the nonces and it turns out that there's some provable statements in quantum mechanics that say you can't do better than solving that kind of problem in time n to the 2/3 so you've you've gone from a time which was n to the 1/2 to a time which is n to the

2/3 and that's a lot slower so you've done some clever modification of your mining to make quantum computers have less of an advantage so are there any block chains or cryptocurrencies that are using this this algorithm today I I don't know the answer to that so this is something that you guys came up with no so I mean I think momentum has been discussed before as as something to you know make mining more difficult but yeah we have a we have a proof that this is the speed that what like classical or

what a quantum computer could do against that kind of counter measure okay so just to sort of close this topic on the hashing algorithm part of the paper so in conclusion your findings are that in the in with current proof of work hashing algorithms are quantum resistant to a certain threshold and we shouldn't get to that threshold for at least another 20 years presumably a lot will have changed before then so you know if your bitcoins are stored somewhere right now and you're wondering whether or not a quantum computer will be able to steal them or if you're sending Bitcoin

transactions and afraid that maybe an attacker will be able to come in with a quantum computer and double spend those transactions or revoke those transactions you know you should not worry at least not for that attack vector for the moment because quantum computers are not don't outperform Asics in terms of their ability to solve the mining problem in in proof of work and right if if we do get there and if you know Bitcoin still uses proof-of-work mining in 20 years or 25 years or whatever we can reinforce mining by by using this this new

algorithm that you described called momentum essentially makes it exponentially difficult for a quantum computer to solve the to solve the mining problem yeah you can you can modify the mining problem to make it harder for quantum yeah that's okay that's reassuring yeah cuz you know I wouldn't want quantum computers stealing my bitcoins right great okay so now that the net that we've been reassured on that front let's let's move on to the second part of the paper which is about digital signatures and specifically paper describes how the elliptic the elliptic curves the signature or algorithm and the specific

curve that is used in Bitcoin could potentially be attacked by a quantum computer so as primer let's first sort of dissect the elliptic curve signature algorithm and how digital signatures work yeah so the elliptic curve signature is taking some input and it applies this function an elliptic curve function to the input lots and lots of times and then you get an output so you can like you know go on Wikipedia and look up what this looks like but it basically involves you have

some funny-looking elliptic curve in a 2d plane and you're given some points and you can go to another point on the curve and doing that is calling doing applying this curve function once and if then you do it again to another point in the curve and again and again and again and at the ends you'll get to some final point and that's that's called the output of the elliptic curve function and this is thought to be hard for classical computers to work out how many times you applied that function it's called a discrete log function so you know the output is equal to a function

operated on x times and you want to find out what X is it's like taking the log of a to the X well the log would give you X so you're trying to find out what that is and this is thought to be really hard for classical computers to do it's it's you know kind of like the hashing problem where it's it's easy to do it but to find out what how many times you had to apply it is hard so in 1993 Peter shor came up with this brilliant quantum algorithm for solving the discrete log problem and since that's the basis for you know

having security of decontrol signatures based on elliptic curves this got everyone really you know standing up the NSA in the u.s. got very excited about this and this really made the field of quantum computing take off in fact if Shor's algorithm didn't exist it would have taken much longer for quantum technologies to be where they are today just one clever idea like that made a big difference yeah so what happens then is because quantum computers can solve this this discrete log problem efficiently and by efficient I mean it

goes like order n cubed where n is the number of bits and is typically 256 bits then there's a real threat to digital signatures used on a platform like Bitcoin and the way it would work is that say after a transaction has been broadcast to the network but before it's placed on the blockchain then a quantum computer attacks this so if if the secret key can be derived from the broadcast public key before the transaction is placed on the blockchain

then an attacker can use this secret key to blog broadcast a new transaction from the same address to say her own she could basically steal money and the attacker ensures that a new transaction is placed on the blockchain first I say offering a high transaction fee and there you've got a theft so yeah that the threat is if you can do this quantum computer attack within a typical block time but for Bitcoin is ten minutes then you're able to steal so let's let's just back up a little bit and and and sort of describe this this attack factor on

elliptic curve signatures so with with the quantum computer were able to so a public/private key pair is derived from from entropy so we generate this random data from that random data we apply function that function generates a private key and from that private key we apply another function and we get the public key and in a Bitcoin transaction whence when we broadcast the transaction we broadcast the public key and and the transaction which is signed by private key yeah and the the attack that you're describing is one in which with the

private key we would be able to use a quantum computer to sort of reverse engineer the the function and obtain the private key yeah so that's right so I mean of course if you're if you're worried about quantum attacks you would want to use you'll generate new public keys you shouldn't keep you know public key sticking around but if you you know stick something on the blockchain and you're waiting for it to be verified you've announced your public key and a quantum computer can infer what the private key is from that public key just like what you were saying and and then once they do that then then they can generate their own

transaction because they have your private key but a Bitcoin address is actually a hash of the private key sorry oh the public key it's not actually the public key itself correct yeah but they could basically spoof you because they have your private key so it's still you can still derive the private key from the discrete log problem given the public key did you describe that works I don't understand what you mean but they could spoof you well so yeah it's just that if you know you can you can derive the secret key from the broadcast public key so from the half of the yeah public yeah and then so then before the the transaction was placed on the blockchain

then an attacker with the access to the secret key could broadcast a new transaction from the same address - you know her own account for example okay but if it's if it's a hash of the on the unless just refresh my memory or unless the the actual public key is also broadcast in the transaction in addition to the hash of the public key isn't it true that with the hash of the public key given what we know about hashing algorithms we can't get the actual public key but we only have the hash ya

know so you have to have the public key yeah okay I see so you need that you're right if you'd only had the hash of it that wouldn't you wouldn't be able to get it okay and so in the paper you described three types of attacks so you've talked about one so one one type of attack is that when we broadcast a transaction we can derive the private key from the public key that we've obtained which is in the men pool and while that transaction is waiting to be to be added to a block someone could presumably within ten minutes derive the private key from the public key the public key and issue a new transaction

with a higher transaction fee and that would presumably get picked up by a miner and essentially you know it it looks like like two transactions and the miner would would pick the one that has the higher transaction fee right but that would we would need to assume that one could do this attack within ten minutes can you give us some sort of an idea of like like giving current maturity of quantum technologies is this something that can be done today or again is there sort of a timeline here where yeah so that's that's exactly right so basically we we're this we consider this the most dangerous attack because even if you're paranoid as users

so you don't like you're not reusing the same public key over and over use you know a new one for each transaction even then this is still a threat and so yeah we really wanted to know how long do we have to wait before this is a viable threat and this is where you have to really get into the physics because you know we can estimate the growth of qubits quantum Moore's law and how good those qubits are but because the gates aren't perfect you have to use what's called quantum error correction and I mean our correction already occurs in classical computers all the time in your

cell phone we're using error correction all the time and quantum computers have to do the same and it turns out there's a big overhead so what we found is using the quantum Moore's law and the number of gates and cubits you need to solve this discrete log problem to crack the signature we estimate optimistically quantum computers could do this by 2027 that's a you know a significant threat I mean of course that assumes that there's a lot of progress and but you know China is throwing a lot of money into quantum

computing Europe in the US Australia so I mean it if you want to be conscientious then this is something to be worried about so what counterman so that's that's actually in less than 10 years so you know presumably he could be before it could be it could be after but how can what kind of mind of ministers do we have to protect against this type of attack yeah so fortunately the computer scientists have been working on this for a long time and the big best countermeasure is using what's called post quantum cryptography post quantum crypto and in fact there's a whole field

like this there's conferences devoted to this topic and the idea is to to change the the digital signature to use a different kind of algorithm not not elliptic curve but there's a whole host of other algorithms some of them are hashing based other than use what are called lattice based they go by names of like things like the lithium bliss rainbow X MSS and these to our best knowledge are secured against quantum attacks basically because they they don't have the kind of structure that

the elliptic curve signatures do so what one thing that we haven't really addressed here is that public private key cryptography and elliptic curve algorithms are used to secure all kinds of stuff not only Bitcoin I mean these are used in banking in encryption in military I mean these these technologies are used to secure any type of transaction online you know HTTP TLS so you know if if if one finds a way and if we get maturity in quantum computing to crack public/private key cryptography bitcoin

is is one of the things we should be worried about but it's it's one of many that power the world economy I believe you when when you say that you know people are working hard on this problem because there's much more at stake than simply you know that a couple hundred billion dollars in market cap in Bitcoin and other cryptocurrencies sure I mean you know maybe the location of nuclear silos has been encoded using RSA or something like that and yeah people have to be worried about that so this quantum cryptography these new algorithms that you know could presumably protect us against CT types of attacks you need a quantum computer in order to generate like a quantum public/private key pair or can you do this with a regular

computer will there be sort of a discrepancy and you know who has access to this technology or is this something that you'll be able to do with any type of computer yeah so for the the post quantum crypto that can be done with any type of computer but the downside is that the transactions race will be slower so there's a lot of work now and trying to speed up the transaction rates what do you mean by transaction rates well just because the the key links end up being longer then would be used for like elliptic curve so it just you know things would be slower so you mean it takes more time to generate a private

key pair yeah I see and then there's all kinds of fun things you'll look like when you look at the what's going on in the conferences people talk about side-channel attacks where you can try to learn about the key a secret key by monitoring the CPU usage of the computer that is that is generating the the key using post quantum crypto so there's all kinds of fun science about really proving that these things are secure okay not to be confused with a blockchain side channels this is another type of side channel yeah that's right right okay that was one attack vector so that the paper also

describes this other attack vector that you briefly mentioned which is that you know since we reuse addresses once you've so say you you you you you issue a transaction to the network you've issued the public key say for instance there's a there's a there's change that goes to an address if we know the public key of of the of the address that has this change then we can again sort of derive the the private key and spend unspent outputs after coins have been sent to a new address and so the way they made against the mitigate

against this would be to just keep using new addresses all the time and never reuse addresses yeah that's right okay and there's a third type of attack that you described in the paper you talked about that one yeah this is the processed transaction attack so yeah if it if a transactions been made from address which has not been spent from before then yes as you say this transaction would be placed on the blockchain with several blocks following it and one could then you know try to use the quantum attack to get the private key but then you have to catch up to the current block which requires

that you solve this hashing problem we described above which is still hard for a quantum computer so that one we don't see is a really serious threat so if basically if users use due diligence and always using new private public keys we were continued to assume that quantum computers are not going to get a big advantage with hashing and the only real significant threat is being able to crack this private key during and under a one of these transaction times for Bitcoin ten minutes ok and and presumably by then we we will have made enough advances advances in quantum

cryptography to be able to mitigate against this type of attack and you know the the entire world of public private key cryptography at some point will hug be overhauled from the ground up in order to protect against quantum attacks yeah yeah good employment for coders ok so so in this case bit short timeline on the attack vector that we described but you know if you're worrying about your bitcoins being stolen you should probably also worry about everything else including the funds of your bank account the military and you know the nuclear bombs that are protecting your country against mutually assured

destruction exactly so we got bigger problems we've got a bigger problems than than the Bitcoin sitting on your ledger nano right ok so that's I guess reassuring you know that if if Bitcoin is gonna be secure then you know other more important things perhaps would also need to be secured so the paper doesn't address any other block chains but you know these algorithms specifically the hashing algorithms used in mining and hashing algorithms that are used just in at multiple areas sort of in in a way block scheme works and these signature algorithms are used in other blockchain so you know have you thought about how

this can also affect other block chains just etherium you ripple or any other the other blockchain networks and blockchain stacks that we see today yeah so I mean at least from this signature part there was some cryptocurrencies that have taken this issue pretty seriously so for example H cash is using the bliss algorithm the post quantum crypto digital signature algorithm Cardno is using a quantum secure signature quantum resistant ledger is also I think they're using an XM s s version but yes in this paper we were really focusing on the the

proof-of-work mining and the threats to proof of stake is a very interesting one which is ongoing research can't say now that we know that proof of stake would be quantum secure but there are people working on it okay great well you know those who are more interested in this topic should definitely read the white paper or the the research paper that was written will have links that show notes also will have links to an article in in the the Technology Review that describes this you know this white paper this research paper and the the attack

vectors is in in a in a much more understandable and sort of layman's terms kind of way so what will have the links to that in the show notes before we before we wrap up here also like to talk about qubit and spent some time on on this new governess mechanism this new I think AJ you described as a as a Dao too in that will invest in quantum technology startups so can you tell us what is qubit and why did you choose to start this company yeah so it actually spawned off the back of this paper when we sort of thought started thinking about how quantum technologies were going to impact the world and how we

could play a bigger role in where this was hearing and I actually went to an event where there's a few VCS and investors there and it was called should-should VC's get entangled with quantum technologies and the biggest question that everyone was asking is how do we get involved you know I don't understand the field how do we vet these startups and make sure that actually I'm going to have viable commercial outcomes and then under startup and it was difficult for them to find funding when their commercial outcomes we know 10 15 years away so we started thinking about how could we come up with a way that allows these thoughts get funded allows VCS to get exposure to quantum

technologies and how do we ensure that we picked the best startups and then Quinn came on board who was a chief governance officer and he was really into the governance and blockchain space and he wrote a few papers on it and together we sort of came up with this governance protocol to select the best startups and get a layer of 50 experts across the walls such as Gavin to help us make these decisions in where we should fund so what is it about the the quantum technology space that is is so unique and that you know maybe doesn't apply to other industries in terms of how VCS and investors make decisions about what companies they should invest

in my personal feel is that it's like the internet all over again when the internet first came out you had all these people that didn't understand it or really had you know adverse to it and they didn't really know how to go about it I think it's a very similar landscape right now with quantum where quantum technologies coming out and a lot of these fees a lot of people aren't really sure what to do with it or how to you know go about this opportunity and that's that's essentially because you know quantum is going to be like a backbone to all these other technologies like I said before so like the internet spawned a bunch of different companies in a web applications mobile applications etc quantum is really gonna

spawn a lot of other fields like in medical industry AI etc so give us a sense of the sort of quantum startup ecosystem because I mean of course we hear about like AI and blockchain and like machine learning and like there are these sort of hot hot topics and like hot startup fields at the moment that a lot of investors are interested in but I mean personally I I don't of many quantum startups or certainly not heard of very many and to me the only actors that are really doing a lot of working for this are you know like we're Spencer and earlier IBM and Google so give us a sense of the ecosystem at the moment like what are people working on yeah so it's cool because we get to sort of interview all of these

interesting startups are doing really cool projects in in quantum technologies so some of them could be around quantum random number generation then some of them looking at middleware so looking at the layer between the qubits and how we operate algorithms on it medical applications so I think I mentioned before about quantum tunneling and how you can use it for genetic mapping then you go into hardware cloud computing so Righetti is a cloud computing company as it's done quite well and then also things like secure communication so how do we take telegram to a completely new level of secure communication and how do we take photography and security to a

new level as well so these are kind of the startups that we've been talking to over the last couple of months I mean these startups don't have the the hardware to actually build anything right now I mean that they don't actually have quantum computers what are their mostly just at the research stage for the moment so generally what we're seeing is a lot of these startups being backed by academics so for example even qubit protocol we have you know four quantum professors on our team so a lot of these guys do have a very strong academic background and also on top of that we now see all these systems like IBM q rickety forest where you can use simulations and cloud computers to write

software and algorithms which is quite interesting so which we're trying to see how we can really help the ecosystem as well with these resources and the expertise to help them really cultivate these companies for someone who wants to start a company in in the quantum technology field like say for instance you wanted to build a product that does secure messaging using quantum technologies so you mention this example of telegram at at an entire new level given the uncertainties and given the fact that you don't actually have any actual hardware what's what's the risk involved here for an entrepreneur

but also for an investor I think it's important with these new companies coming up that they do have an academic presence I think it's a bit difficult for companies without that academic backing or someone who's you know research this for years to go out and say hey we're gonna start a quantum messaging service however with that said there are a lot of quantum cloud computing services out there like Righetti forest and IBM q so there are starting to pop up some tools that people can start using but I think at this stage at this early stage just like the internet like you know back when the internet was first starting you really

needed a research background or something very inherently technical to start these things nowadays when it comes to website so you can just go Squarespace and make it but really early days was very difficult so I think it's a very similar ecosystem right now where you do need someone with a high level of technical expertise research background to help you out if you are going to go into this endeavor but I think going forward we're gonna see more of these platforms like rickety forest and IBM q to help entrepreneurs get into the quantum space so one of your opinion are the areas that we should see so the innovation occur first like are there

any specific types of applications of quantum that you know we should see sort of mature in the next few years so just looking at the sort of start ups of we're talking to right now we're sort of seeing more software companies I think there's gonna be a lot of innovation in the offshoots so helping quantum hardware become more efficient so I think Gavin sort of touched upon the fact that right now one of the biggest limitations is this gate era and creating stabilized qubits so looking at surrounding technologies and how we can help improve that as it was - directly impacting it and also just applications of quantum floud computing

now that it's available what kind of algorithms can we design what kind of specific applications maybe for genetic um genetic composition of people like how to read that may be looking into how we can impact AI so there's this sort of fields that up-and-coming yeah and also like with quantum sensing is actually a really big area that AJ brought up miniaturized atomic clocks there are actually companies already making these and that's really great especially in the case of a GPS spoofing attack if you have a very accurate clock you can do navigation also gravimeters which can be used for mining and

construction to find cavities under the earth surface and and magnetometers we there's a company based in Europe that is envision that is working with polarized diamond spins that are quantum spins inside of diamonds and you can use this for enhanced MRI so you get much better resolution of MRI imaging and so there's interest in the medical community in that and this kind of stuff is is a lot more near-term than quantum computing fascinating so tell us a bit more about this Dao how is it gonna work and sort of what are you expecting in

terms of the types of startups that will sort of apply for investment and who you're targeting in terms of investors yeah so the way it works is the governance protocols there's two layers we'll have 50 experts across the world in various different fields so academic background quantum formalization etc and then you have your public voters so essentially just like a normal government except in this case public voters will also be able to vote on the proposals themselves as kind of like a liquid democracy as the person of purely representative one so that's really interesting how we're pushing the boundaries of governance through Quinn's input because he's written a few papers

on it he has analyzed the Dow from 2016 an academic paper and then that's connected to it Dow which is obviously storing all the funds that we then invest into the quantum startups that are selected through this governance protocol and the kind of startups are going for is in we're sort of seeing it two types one is sort of the accelerator standard type where we're seeing a lot of PhD students sort of graduate from university and sort of say hey we want to start a company in quantum we've done a lot of research in this area so that's where we can come in with the expertise and the money with an accelerator program so a more formulated program that something like 500 startups might

do and then we're sort of seeing a more establish to the research direction sort of becoming a company later so a professor might do some research in you know five years into a quantum the random number generation and then decide to incorporate it and then you know try to commercialize the product and as soon as the second half of it and that's also something we're looking to helping out as well and so these funds so the CEO is sort of an investment fund so investors will well invest money in this Dao and then these these sort of advisors will curate the the projects that sort of bubble to the top and which investors

will place their money what's the structure or what's the legal structure then for a company is it is it like an icy it or they'd sort of doing an IC o---- on this platform or they giving way I said to clarify um it won't be investors putting money into the fund wolf will put money into the fund through our SEO so when we sell our tokens and you buy to take part in their governance vertical will use sixty percent of the funds to put into the Dow and then that will be used to invest into companies any profit that derives back to the fund is we're looking at buyback schemes such as wahby and Finance to help derive value back to our token holders now in terms of the

startups we will get equity from them for a certain amount of money so traditional investments into quantum and the idea is that the value always flows back to everyone who's in the community so it's not just the money that we're giving it's just that with this whole token design now you're sort of vested into the portfolio companies we get this quantum community everyone gets behind it it must trying to help out including the 50 experts that we have okay so tell us about the roadmap when when will this launch yep so we're currently in the Private Eye Co so we're raising at 20,000 eath hard cap at the moment we're getting onboard experts so we're at

about 10 confirmed and about 20 that we're sort of still negotiating things with with a goal of 50 in the next month or so and then we're looking to do a pilot program which is really great in the next two to three months so we've already started application so we've interviewing startups and funneling them through the process with a main platform launch at the end of the year and even the conference that we're planning a world quantum conference great well this has been a fascinating topic I'm really happy I was able to speak to two experts on the topic to better understand quantum computing quantum technologies and how it relates to cryptocurrencies I hope that all the listeners now have a

also a better understanding and grasp this concept with a lot more a lot more certainty so thanks for coming on guys thanks for having us excellent thanks for having us thanks for all the good questions and thank you for to our listeners for once again tuning in you can get new episodes of epicentre every week and so you can find epicenter on iTunes SoundCloud or wherever you get your podcast you can also watch the video version of the show on youtube if you want to support the show a great way to do that is to leave us an iTunes review it helps people find the show and we're always happy to see your reviews if you want to join our community you can do so by going to epicenter TV slash Gitter to join our theater community

so thanks so much and we look forward to being back next week [Music] [Music]