Quantum Computing vs Web3 Cryptography: Risks, Challenges andOpportunities – Ian Smith

code is for our telegram community and these slides are linked from the telegram community so you don't have to take a picture of everyone and try to OCR the the URLs you can just go to the telegram community and then click the links this is going to be uh somewhere between the academic and Engineering that the last few presentations have been so these are the the companies that are currently working on quantum computers in the retail space uh slide comes from my adviser who had a a lot of contributors to make this so this is not including the University or research labs this is not including uh the governments like China who have basically nationalized all research uh this is just the commercial space uh sorted by the the kind of uh quantum computers that they're making there's a lot of them and they have massively varied approaches even in the photonic section they're using completely different sections of physics uh they're using different methods of observing different methods of calculating and there's lots of variety inside each one of these these systems this is River's road map uh somewhat these are these people are very Advanced they've solved some problems that other people haven't solved but this is kind of like a typical road map to compete with IBM's road map they are talking about getting a a million cubits in 2026 and then from in 2027 they might have like five million they might have 10 million they might have still just a million but this is a very kind of like a standard road map that you're seeing today where people are targeting very large machines very very soon now there's I should say that this speech should have around a 100 papers as footnotes there it's very hard to say anything clearly in Quantum because everything is uh complicated as some of these U presentations today may have shown you there's a lot of complications here so I'm just going to try and say some some things that should probably have many many many many many footnotes and maybe later versions of my presentation will include many many many footnotes in the notes but I'm just trying to make some some broad Strokes here some understandable statements and and point at the direction of Technology the risks the advantages and the opportunities courtesy of uh verum this was the original IBM road map overlaid with the improvements in the algorithms overy year and the expectation this is an old one you see from uh 2022 is when this was published there's an expectation that every year the hardware requirements go down as the algorithms get better and better and better every year the hardware gets better and better and better and IBM hit their road map but they removed it they deleted their road map they said okay we have to start over with error correction this time and they have a completely new architecture now but this was the architecture that they actually stuck with and hit all of their targets um the big problem here is that this was the where the algorithms were predicting uh a single large machine and the reality is that we we touched our Hardware performance in terms of just raw cubits by improving the algorithms to a network mode that networked mode doesn't use impossible to build Hardware it uses Hardware that's currently in mass production so that's a that's a huge shift not necessarily in the the machines themselves but what we can do with them because now we're using Hardware that's of that's on the edge of production or being produced depending on the company and what their architecture is uh some of these companies are still you know five or 10 years away from making something that hits the minimum 1152 cubits in the chart some of them are doing it right now by modules the minimum kind of threshold here that we're talking about danger is 8 million cubits and at that point it's possible to break Bitcoin and just do it a bit slowly that number used to be billions of cubits if you check Microsoft's uh Quantum resource estimated at Q sharp they're still saying billions of cubits now the algorithms have improved Microsoft is still publishing the old information it's very important to read a lot of papers to figure out what the current status is Daniel latinsky works for squantum and he has been publishing Quantum resource and Quantum papers since 2016 I read some of his early papers I looked him up and I found that he's been very busy active volume takes and says instead of copying the state of every Cubit to every other Cubit we're just going to take the part of the algorithm that you need to calculate and copy it where it needs to go which is what we do in Silicon all the time um almost all the time the current state-of-the-art of AI compute clusters is where they copy the memory of every single video video card to every other video card in the cluster which is very similar to the way that the quantum computers were working with mutual entanglement now we're just copying the part of the information that has to be copied in order to perform the calculations required one other thing about this is that it's 40 times faster to use a lot of Little Machines instead of one big one that that computation performance increase shows up in the the chart to the right uh there is a lot of performance that can be gained by changing the architecture and then improving the algorithms especially with Quantum memory uh over and over and over this is an iterative process latinsky uh in his presentations has said that there is a lot of performance gain that can still be had both in the active volume architecture and in his architecture targeting uh Bitcoin specifically so active volume was published with RSA as a Target and he's targeting uh in this paper 10 minutes for a 256bit elliptic curve the block time for Bitcoin is 10 minutes so they would be able to read the memory pool of the pending transactions check the public keys and convert that to the private keys that's what he's specifically talking about getting the private keys of pending transactions and then spending the money before the block has been mined and before the money has been uh delivered it's just taking it right off the wire this isn't satoshi's old coins this is your pending transaction so on the right hand side we have slope machines with slow memory and on the left we have fast machines and he did these calculations for different architectures but he he picked samples that he had data for um so on the very far right you're seeing trapped ions which is Oxford ionics and on the very far left you're seeing superconductors which includes you know IBM um to a some extent d-wave uh but also Google so there's there's different platforms have different capabilities some of them make memory easy um as previous uh panel mentioned and some of them make uh the calculation very fast but at a trade-off that you need more memory Linsky included calculations where there is a 20 km fiber optic cable used as a delay for the photons to make them slower in order to need less memory which which means he the the algorithm actually runs with fewer cubits but it runs longer so this is kind of their uh sample uh this isn't a picture but this is the diagram that they proposed and they they suggested as their device these are currently in mass production at Global foundaries in uh several facilities the main facility is in Upstate New York uh another component is being produced in Dres Germany and they're trying to launch production in several other facilities um squantum is currently in production with a device that's horizontally scalable it can just keep making more of the same device and be able to solve bigger and bigger problems as time goes on and it scales faster than one for one for every tripling of cubits they get a 7x performance increase that's huge that means that we're going to be focused on large data centers for computation that are running quantum computers maybe only one per company and we're going to see a lot of computation being done very Central fashion uh Oxford ionics they have a system that runs at room temperature it's 256 cubits they're talking about putting it on just being able to manufacture these on you know the same Wafers that motherboards are printed on using the same facilities that motherboards are printed on and this runs with magnets and lasers it's not a simplified production whereas squantum is a bit head in the on the manufacturing in the production they're only making a few cubits per individual device and so Oxford ionics if let's say they make an 1152 Cubit uh device they'd be able to manufacture a lot of these things and be able to use all of them for cracking Bitcoin and ethereum and other other systems there's that that threshold of 1152 cubits is important because if it doesn't meet that threshold it can't maintain the network connections and do useful work so you can imagine that some video game has or some AI says that each graphics card needs to have 32 G of ram but it will run slowly on some models and faster on others it's a similar kind of like minimum requirement as long as you meet that minimum requirement you can join the cluster and do useful work Intel is ah head in manufacturing they are producing these very large Wafers um blinks include pictures of them these large Wafers use electron spin they don't have networking and they don't have the algorithms on the Wafers so it's like a a processor that can't run any code at all but when they do figure those things out you're looking at probably the number one manufacturer for Mass scale production of cubits maybe a million cubits a day and that's why it's so scary and so dangerous we don't know how many cubits they're producing on the Wafers we don't know what the uh yield is for error correction all we know is that they have the ability to produce a lot of these things so when Q day comes when Quantum day comes for for blockchain and elliptic curve cryptography it's going to alter the way that we do things in blockchain specifically the public Keys get converted into the private Keys the private key is your money people will be able ble to send your money to other places uh this also means that smart contracts that have an administrator will be able to have that administrator privileges stolen and then the contract updated or replaced or managed they'll be able to drain Bridges they'll be able to drain dexes liquidity and staking pools will be able to be drained uh if someone's you you know has their private key as their for their validation service that validator will be able to have its currency removed and sent other places it's just Mayhem all over um Bitcoin says that they're protected by utxo from coron computers what that means is they're protected against slow quantum computers the previous slide went from you know 16 days or 160 days down to 8.3 seconds utxo doesn't protect against a 10minute attack they could shrink their block time to one minute and that might buy them a few months or a few years depending on how quickly the mass production works and it's just trying to outpace the mass production it's not trying to uh protect against the underlying math or the underlying science it's just trying to be faster than the mass production so the utxo is only a little bit helpful it just means that it takes a faster one to take the out of the account Bitcoin has a bunch of old uh addresses um they've done a migration before and that migration moved from the public key to a hash of the public key satoshi's coins are public key storage it just directly uses the public key as the address and about 40% of the old Bitcoins use that old address Style so we can prove that qay has occurred when satoshi's coins get moved we'll know that someone has a quantum computer of sufficient size and then it'll be proven to the Bitcoin and you know blockchain Community if satoshi's going to move there's a bunch of trapdoor wallets that are out there a lot of money sitting on a on a wallet that uh has a public key exposed um and people are watching that in case of a breakthrough but this is just when you can prove it we know it's going to occur the only thing we don't really know is is when one of the problems here is that let's say that someone takes satoshi's old coins now there's doubled the number of Bitcoins in circulation because there's a bunch of public keys that are available so there's an economic issue and if you check uh coin market cap there's $1.7 trillion in blockchain web 3 uh on layer ones on layer twos there's about 80 billion or so 70 billion in Defi and just all together it's about $ 1.7 trillion in terms of stable coins usdc and usdt there's $150 billion but that's only 8% of the total value lock which is a bad calculation but just scale of the problem only 8% has onchain Fiat but you can't actually get that Fiat to your wallet you can't get it to your bank the only you can do is say well it's onchain so how do you get the money well you have to find someone who's willing to pay dollars for crypto and that those dollars will become more valuable relative to crypto when everyone realizes that the crypto is going to to become worthless and the closer to get to qay the more worthless and the more people will understand that some people will just wait for the end they'll wait for for the the blockchain to be provably insecure before they say that it's it's actually doesn't have value anymore I try to say that blockchain is a system of trust you're trusting the cryptography to to prove that the money is correct that the value is there but if they've been about the cryptography being Bitcoin forever money well if it's not forever and the narrative is wrong how valuable is it and so we're facing a problem where misinformation to the average person may poison them against blockchain which is a valuable technology having the wrong narrative will hurt us in in the near term or medium term and then when they sell there's just not going to be enough buyers we're going to experience a bank run one way or the other there's just no way that we can't that we can really avoid a bank run what we can possibly do is prepare problem here is that we have to prepare ahead of time if you do the migration and you do the preparation in the middle of the attack then the attacker can just take the money that you're trying to migrate so this is I had to say something I would rather say nothing of course just not be possibly wrong but I've I I think we need to consider the aggregate risk not just beside Quantum but all of the companies that are working in quantum computers collectively and the collective risk becomes very significant in 2026 and than very large considering the road maps in 2027 I don't know if the any road map is going to to come to fruition IBM had a good track record but then they had to throw it away the track record for riverlane well they've got some good successes but that's one data point we don't know if they're going to be able to maintain their road map uh timeline if they can maintain that momentum well then riverlane or you know side Quantum if they produce a bunch of machines and they actually work that gives credibility to those timelines and so this is I'm not saying that their timelines are credible I'm just saying this is what the timelines are saying this is what the timelines are predicting now some companies like IBM have a really really poor timeline for the future they they're not going to get anywhere close for a large machine for a long time which is want Daniel latinsky let's use a lot of little machines and let's use them to solve one big problem that's a that's a huge shift because now you don't need IBM's super super super quantum computer to actually work now you just need a lot of little ones to turn on so using that methodology I have kind of a a napkin calculation of 2027 or 2026 there is a very low risk in 2 but the risk is still present but it's based off what we don't know not what we do know the 2027 risk comes from what we do know um latinsky uh this is a link to his crypto presentation where he spoke to cryptographers about his algorithms and using active volume and the the 50 million tofly gates to to break Bitcoin in 10 minutes uh he presented that paper he explains it and he includes in the Q&A section as well as other sections in in the presentation further optimizations that can be made I don't think that we're going to come up with the the current algorithm as the as the one that's going to actually break Bitcoin I think we're going to have a better algorithm that actually breaks Bitcoin where is that well 2026 is more viable when you consider a lot of Innovations in the algorithm space and some Innovations ahead of schedule or for some schedules on schedule for quantum computers it's not just how good are the machines it's how good is the code versus the machine and that's the big takeaway I want everyone to get every year that goes by the hardware requirements go down for cryptography every year cryptography gets weaker crypto toog graphy never gets stronger it just gets weaker as the NSA and other you know mathematicians have more and more attacks and the the computers get better and better why are people picking on bitcoin it's the next Target it's just what like what we can't do in emulation uh and when we have enough uh enough machine power what about RSA which is the RSA 2048 is actually a really common SSL certificate Why not pick on that well that's a harder Target it's not that Bitcoin is what people hate it's what they can do it's just the next Target ethereum is the next Target uh you know salana is the next Target sana's got slightly weaker along with cardano slightly weaker cryptography so if they just miss the mark on bitcoin well then they'd be able to break cardano and uh uh salana ed25519 is slightly smaller sorry I adjusted my slides uh last night so here is uh a possibility for web 3 and that is to migrate um this is true for Bitcoin and eum as well and it's just to change the cryptography and there's there's a user experience here that's very very important everybody has to sign a transaction from the old cryptography targeting the new cryptography the cryptography has to be exist and has to work before they can go to the new new system now there's proposals uh emergency proposals of going to cryptography that doesn't work and that was one of these like six Monon migration plans for Bitcoin and the idea there is that you sign it you have a calculation but the network just ignores the calculations like you've just locked your money for the next six months or nine months while they actually build the next system out and make that cryptography work but in every situation it's going to be a migration now there's several ways to do those migrations um it's actually harder to upgrade a blockchain in place than to make a new one when you're making a new one with a blank slate you don't have to worry about the old cryptography you just need to provide the ability to migrate to that cryptography you don't have to have a double consensus mechanism you don't have to have double validation signatures you don't need to have double wallets you just need a new wallet in terms of migration um signal Apple Google Chrome have already done this migration partly for the you know capture now decrypt later attacks which people have said oh we should have we should have migrated last year or the year before uh 2022 would have been a good time to prevent the capture now decrypt later attacks so anytime you have a central key distribution it's actually pretty easy to upgrade SSL has already you know begun that process and they're just waiting for keys to expire and then they'll give them new keys and the new keys are quantum safe um some companies are going to be massively affected by this especially if they do uh you know content distribution uh any kind of content distribution networks are going to have cdns are going to have a a big big penalty because the cryptography is really really expensive and so they're you know casting forecasting tens of billions of dollars in just Hardware upgrades there's s some examples of systems that are actually hard to migrate and those will be like cars because the public private key guarantees the the firmware is correct the USB bus hasn't been tampered with um so all of the keys use public private key cryptography um all the cars do U missiles for the front and FID identification systems and some of the military grade uh encryption this is actually done over the wire between like the helicopter and and the base or between the missile versus the plane and when you have lots and lots and lots of Hardware that has your shaped keys that becomes harder to migrate um iot devices uh and the fips chips for certifying uh to banks have a harder problem because they don't have enough RAM in the devices to use these massive keys that are required for post qu photography so they're facing a different struggle it's a different and hard problem also uh but a lot of the iot has upgradable firmware it's just that they don't have enough RAM on these little devices so your refrigerator might need an upgrade but do you really care like someone's going to make the ice machine turn on so what so when people are talking about uh well bitcoin's going to suffer but so is everyone else not really there are some places where it matters but usually there's not a public record of all the a bunch of public Keys you actually have to get close to the the jet with the to get the key challenge to to do the fer identification system if they're not broadcasting it how do you have the public key when we have a ledger of public Keys especially under the account model that becomes a much more critical issue upgrading becomes more important so the advantage here is that we can make a clean slate we can just start off with no old cryptography and I'm I'm telling everybody hey look uh I don't want to make a blockchain with old cryptography at all no zase uh snarks at all just have clean cryptography nothing old no elliptic curves anywhere in the ecosystem and that's much much much safer than having a hybrid system with old cryptography dangling around there might be a bunch of money trapped in a ZK snark if someone can fake the ZK snark then they get the the secrets as well as the ability to spend some value that was sitting inside that CK or the ability to forge a new one there's a lot of possibilities when you start using elliptic curve cryptography lots of places and I think that the only safe ecosystem is a new ecosystem so we're working on that we're in testet now so the weakest link should metric should be applied it doesn't always happen sometimes you can have cryptography combined and you have toing break both systems before you can break the the complete system but usually it's based on the weakest link and if you can take someone's private keys well then you can just be them and spend all of their money regardless of what put is put on top of that no matter what you're gluing or dressing it up as if you can prove provably be them according to the system of trust that you're working with then you are them so hybrid systems mixed systems I consider highly dangerous there are some possible exceptions to this but rule of thumb just trust by default and then you know verify experts really really need to weigh in on any kind of hybrid system possibly using a ZK snark and then a ZK Stark on top or ZK Starks and then ZK snarks these are things that are done but they have to be very carefully checked so the opportunity here is with a clean slate we need new members um if you wanted to become the new number one Unis swap if you want to replace Unis swap no amount of marketing dollars is probably going to get you there you probably just cannot spend enough money to replace Unis swap but if you're the new number one on the new empty ecosystem well then you start getting as the users migrate people start understanding that Quantum is a risk then you're the new number one by by volume and you become that new unattainable number one nobody else will be able to replace you with a significant amount of user adoption I'm just recommending that people deploy their contracts to our chain as a backup and run both copies the users will come we'll get users but in the meantime at least have a backup for your project at least have a backup of your d uh if you're layer one it might be good to have a backup layer one that runs in a cluster if you have an EV might as well just make a copy of your ecosystem into a safe Quantum safe system so we have to disc discard most everything this is kind of a a little bit of a bunch of buzzwords if you don't see someone talking about lates they're probably not Quantum safe there is a Quantum Randomness and that was covered in previous panels but that doesn't make a Quantum safe network if you have random numbers that doesn't give you suddenly immune elliptic curves now you your public key still gets turned to the private key even if it came from a Quantum random number generator it's still using the same mathematics and quantum computers don't work the same way when you have silicon computers you're doing only two operations addition and comparison when you're doing quantum computers physics does the math physics spits out the the numbers physics generates that public key into the private key so it's really important to use the the N postquantum cryptography because we don't have physics that breaks those math problems so this is the last slide here's the telegram link QR code again and I'm recommending my competition and other projects that are quantum safe if you know of anybody who's working on Quantum safety I'd love us to just join together to try to save the ecosystem even though um you know we may have competing projects I think that if we don't work together we have a lower chance of making web 3 survive period we have to break the correlation with Bitcoin we have to convince the exchanges to stick around and not just sail off on their Yachts we have to work together to actually make uh you know web 3 stick around do have time for questions one question anybody thank you so much for for your amazing speech s [Applause]