So hello everyone as you just heard my name is Yamar Zakov and I'm a journalist but today I will be moderating this panel with this easy to understand and easy to remember name hacks quantum threats and blockchain based finance. What you need to know the audience is that we're going to split this uh conversation into first of all the dangers today. We're talking about hacks of course and second the threat of tomorrow uh quantum threats. What do you need to know dear panelists is that the topic is complex and not everyone in this room will understand everything you said. So I'm going to interrupt you if I think that's needed.
Okay, let's start. Mr. Swolf, another easy to remember name on the topic. Give us a concrete example of actual loss of funds in crypto in crypto business due to comp due to compromise security. Can you give us this example?
All right. I have a question for the audience who has heard about the breach which happened a few hours ago 350 million. All right. So almost all of you unfortunately it happens all the time. So before yesterday I wanted to speak about a different breach but since yesterday I want to speak about generic breaches because it turns out that uh almost all security breaches crypto exchanges and even hardware wallets because the hardware wallet breach was also a recent one and uh the sum was very close to the yesterday's sum.
So yesterday it was $351 million. Uh the crypto wallet was around $400 million. Now the crypto wallet uh there was also a liquid network breach around the same amount also. Uh and there are other breaches at banks for example a swift system. If you are a bank, the objective of the hackers would be to break into you and compromise the Swift system.
You see, the Swift system prints out large transactions on paper and you verify it, but the hackers can make the printers print one thing and the transactions underneath do another thing. In yesterday's breach, the hackers compromised the transaction uh a piece internal an internal piece which was uh linked to the hot wallet of the company and they compromised the signing mechanism where they could issue transaction records as if they came from inside the system. Now the system was designed to operate efficiently as almost all systems are and it complied because it came from inside and issued nine or 10 different uh let's say uh asset types and a total of $351 million. The challenge is that they have monitoring. They probably had monitoring from a cyber security perspective on their IT assets, but they obviously didn't have enough monitoring of transactions and they didn't have enough controls of transactions.
So 351 million gone. But in general when is the uh the cryptography is broken and when we're talking about uh uh human mistake stolen case uh social engineering compromised software. The interesting thing about cryptography is that cryptography is almost always except in the rare cases on very stable. So the the hardware wallet breach for example which was also a recent one the hardware wallet employed signing mechanisms. So the cryptography was solid but the random number generator which they used in the wallet 5 years ago someone from the development team decided to change it to a softwarebased random number generator which was much much less random.
Nobody paid attention. Nobody did like threat modeling of the change. And it it went through five years silent and suddenly 12 or 15 different hacking groups discovered or shared most likely shared the information among themselves and wiped out the wallets of many people. So the cryptography it's about the implementation. It's not about the algorithm itself.
And that's where the weakness is uh in the cryptography and in the smart contracts as well. It's the same thing most in most cases.
Okay. To understand that uh the difference, let's leave let's leave the present now and uh we're going to tomorrow. Mr. Buan Toros is with us. Hello
Mr. Toros. Let's assume that I am the demi that I am. And please can you explain to me and to the audience what exactly quantum computer means?
Well, quantum computing is um an alternative to classical computer uh to classical computing but kind of much more powerful in a sense uh because instead of using classical bits to store information where we know classical bits can have like states of zero and one uh we use uh quantum bits. So quantum bits uh are the quantum analog uh where we have quantum mechanical system which can have two states again zero and one but since this is quantum mechanics we can also have any superposition between zero and one and this makes it much more powerful for for particular u types of problems and u that's why uh quantum computing is so interesting and people have been investing in uh recent there's uh a lot of money and attention into it because there are certain kind of uh problems that can be speeded up exponentially using using these machines. Can you explain to us uh but simply please uh what's the difference between uh I need to help physical and logical cubits and why this difference is important for us to understand how far the uh the the crypto uh the the how far we are from the point of breaking the cryptography
right so yeah that's that's a little bit tricky but when we talk about physical cubits typically we mean the physical system itself that's that's used to represent present the the quantum bit. So it can be like atoms, ions, photons, some kind of superconducting circuits, whatever. So this is the physical system. We we use it to store the information and we can once we give these two states of the cubit, we give it labels like zero and one. We can we can think of it also as a logical cubit.
But but the problem with this is that uh these systems, they're very fragile. So uh what I mean is that when they interact with the environment they lose these quantum properties that are important uh for the algorithm to run. So the ability to store these superpositions of zero and one is is lost very rapidly and what we what we need to do is uh to find a way to uh to fight with with this noise. So people have developed what's called quantum error correction and in quantum error correction uh what we do is we use a lot of these these physical cubits uh to encode a single logical cubit. So uh there is this redundancy of physical cubits uh and we get a logical cubit which which is uh much more resilient to noise but the price to pay is that you need more more physical cubits basically.
A little bit technical question but important one. Google quantum AI mentioned that a figure of under 500,000 physical cubits back in March this year uh as the threshold for a breaking in in uh cryptography. Do you agree with that estimate or would you correct it?
Well yeah obviously I agree. I mean this is uh just uh yeah they they ran some resource estimates Google recently and came up with this with this number and this number is uh based on certain assumptions that that Google did so it's like um physical platform that they use in their case it's superconducting cubits a certain architecture uh like connectivity nearest neighbor between the cubits uh uh a certain noise model and under these assumptions this this this number is is uh correct. But we shouldn't look at this at some as as some fixed universal constant or whatever. When we change the assumptions uh if we turn up if we use a different uh physical system, if we use uh a different uh error correction protocol, different uh noise values, whatever these numbers can change a lot. A question that came to me as a journalist, Google, IBM, they're telling us uh about their progress in quantum computing, but what do you think?
Is this progress real or is uh more of a PR action?
Mhm.
Because of all the investments that were made,
right? Well, yeah. Now first of all I should say that all these big players like Google, IBM, Cera, yeah all the others they they are all like uh serious uh serious companies with very very talented talented scientific teams. So what they do is really remarkable and all all their progress is uh is real. But uh one thing we should distinguish is what they publish in the scientific uh journals which is something that they are very careful about.
And then there is these press releases for the general audience which are sometimes uh I would say very speculative and um yeah sometimes they say things that I'm not sure it's good to be said like uh I I can give an example for example like Google when they announced uh recently their new willow chip uh They uh they achieved uh error correction below threshold which was a remarkable result and everything. They ran some circuit sampling uh which is an algorithm to show quantum supremacy on a problem that actually no one uh cares but doesn't matter. But then there was this press release. Oh look our new quantum chip can has access to parallel universes and something like this which is basically BS. So yeah one if one wants to really find what what's happening should look at at the scientific uh journals and I would say that even if if uh one's not an expert nowadays with AI it's much easier to say okay can you help me with these papers what's happening here and something like that
okay uh a question uh which answer will be speculation but if you have you if you must um specify a decade not just a range a decade when uh do you personally expect to see uh quantum computing that will break cryptography when it will be
yeah well that's that's even harder I mean these predictions never turn out to to be uh true so yeah I I think it's reasonable to expect that uh within the next decade we can have a full tolerant quantum computer that can run shor's algorithm at scale and yeah probably the end of next decade I would say so that no one remembers what I've said. So yeah.
Okay. Thank you. So when we know what exactly uh quantum computing means I will ask Mr. Bab. Hello.
Thank you. U harvest now decrypt later. It sounds alarming but as we know on the blockchain every uh blockchain everything has been public for years. So is the risk really a new interception of a data or or a longstanding exposure?
Um so there is this theoretical attack of harvest now decrypt later that um that is um basically someone is tapping your wire and is listening to everything that is coming in and out and you can theoretically listen to every sort of communication like this. Um so basically someone would gather this data which is currently encrypted. It is useless to you cannot do anything with it. But with the hopes of acquiring a quantum computer one day they can decrypt this data and they can see the secrets that you are um transmitting whether for in whether for financial gain or for influence gains uh political secrets and uh intelligence and whatnot. On the other hand, we have the blockchain where uh all of the data is currently uh sitting uh encrypted and and it's visible for everyone.
Um I I think that this harvest now the crypt later approach um is uh quite um far-fetched. I don't think there is any real risk because if a quantum computer would exist um um it would require um a lot of uh resources to decrypt all of this random data that you have collected which is in the pabytes millions of it is too much data to to be useful on the blockchain. On the other hand, we have um the encrypted um um UTXOs. So the the basically the encrypted data which um everyone uh could see using a quantum computer, they would be able to silently work on this problem. Um but on a follow-up question, we're going to discuss um right which how many bitcoins are vulnerable exactly what what this means to the risk.
But um I I think there is not much reason to uh to be worried about uh um the harvest now decrypt later attacks and there is more to be worried about about yes this publicly exposed keys that are sitting on the blockchain. There's an estimate five to seven years to the industry to to migrate to the postquantum security. where does that estimate came from and do you think it is it realistic from today's point of view?
Yeah, so this is uh the estimate that's coming from NIST um five to seven years also from Google. So there is a lot of research um you know that that has shown that quantum computer is possible. There is a lot of experiments that prove that computation is there. Um it is um developing exponentially. So in many uh ways it is a question of when a quantum when a such a computer would arrive not a question of if um given a long enough timeline it will arrive and and and and this is this is basically what all the lines say.
Um I still wouldn't be too worried 5 to seven years is an extremely conservative um uh timeline. Of course, it is advisable to migrate to to secure algorithms as soon as possible. But in any case, we have the technology that has been proven. Yes, quantum computation exists. It will develop further.
But we also have the the signatures which are which can protect against this computation. Um it is a good thing to follow. Um it is very conservative. If it takes longer, it's most likely not a problem.
That leads us to Mr. Serof. Uh Mr. Sedov, which is the bigger vulnerability uh for the real institution today uh the blockchain wallet case or uh the regular IT infrastructure. Where is the bigger problem?
Actually the biggest vulnerability today is the lack of awareness that uh hackers are already using unconstrained AI models or LM models. uh with all their brakes removed and all their constraints removed to attack IT systems and from that perspective the IT system is the vulnerability but it's actually that you are but are fighting against systems you might be attacked by 10 different hacking groups at the same time all of them might be using AI and you have two three security engineers 10 security engineers 100 security engineers all of them will be inadequate when you're fighting against someone with uh a thousand times the speed and a thou a lot of intelligence and tools that can rewrite exploits in such a way that they would not be detectable by by your antivirus or other security systems. So it's actually uh yeah there were many breaches with uh wrongly created smart contracts but much more order of magnitude more breaches happened because uh the hackers were able to compromise a system inside the company handling transactions. They were there for months. They were learning the system.
They in eventually sometimes they know the systems better than the developers of the system to find flows that they can use for fraud in a fraudulent way or in a way that they com they gain access and issue transactions and steal hundreds of millions of dollars.
Mr. Bab this postquantum cryptography um Ethereum's hashbased signatures are these actually usable today or are still at theoretical stage?
Um yes so as we established in the previous question the quantum technology exists the the computer um as well as um the the signature schemes um for postquantum cryptography. So those schemes are very well established. They have been documented for for many years. They have been heavily researched. They have been tested.
Um so the technology fully exists. Uh it is applicable to to current blockchains. It is safe. There is a different technical considerations, right? So the transactions become a bit larger.
But this is all a technical question. So I think that the technical side uh of the question is fully established. What I want to actually emphasize on is the government the governance side of of of this um question and exactly how are we actually going to migrate those blockchains uh to be quantum secure. So on the one hand you have like uh you have the small coins that have a small community a founder a leader those can easily migrate because the leader can just say okay now there is a hard fork now we switch. Then we have something like Ethereum which has the Ethereum foundation.
It is um it it serves as a very uh wellestablished governance unit. But on Bitcoin and and and this uh this scheme that you mentioned BIP 360 I believe it was um it is a lot more difficult. Uh so Bitcoin is extremely decentralized. Um there is no governance body. There is no CEO of Bitcoin.
there is no founder who can come and say okay now we do the fork and we do it in this way um I think uh it will be a very messy process once a quantum computer arrives and and it and we can see it we can we can see that uh that it's working um and this is the the the question we should be worried about exactly how are we going to going to communicate and how are we going to organize the updates for Bitcoin um and the vulnerable coins Let's imagine that you're advising a Bulgarian bank uh that's about to hold her her first digital assets. What's the one thing you're going to insist on before they hold a single coin? The one thing
uh can you repeat the question please?
If you advising Bulgarian bank about their
Mhm. uh first digital assets if you're going to incited for one thing to uh on before they hold a single coin what's your advice the the first the one and only so so the one advice regarding the the quantum threat is that they should start experimenting with with with quantum um secure uh technology as soon as possible something that exists again it's established um and yeah it it's ready to be implemented as we had said in the timeline five to seven years. Um it's it's we there is enough time to experiment.
Okay. Uh Mr. Ferdov, uh similar questions to you. Uh when you're hired, what are the first three things you check on the system of your clients?
Well, as I said, the designs usually are amazing from a per performance point of view and from a security point of view. the design is you know following best practices in almost all cases otherwise nobody would accept the designs for crypto exchanges and uh digital assets but the thing is that we see breaches on a daily basis and I I can assure you all of them were amazing on paper and uh the challenge is that uh they didn't do proper threat modeling as if you know looking at the architecture diagrams of the system that they are designing and thinking okay what would happen if a hacker compromised this point or what would happen if a hacker gains developer credentials or the credentials of the CTO or the hacker obtains access as system admin and has access to all machines all servers and all desktops in the whole company what would happen then and uh there is a lack of threat modeling there is a lack of simulation of things that are unexpected. So you're waiting for a black swan event. Usually you cannot predict the black swan event, but you can attempt to attack the mental models and the digital models of what you did. Like for example, most crypto exchanges, they have a hot wallet, warm wallet, cold wallet.
Yesterday's breach emptied entirely the hot wallet. All designs say you should have a hot wallet. Why did you store all your money in a single hot wallet? Why didn't you have 50 hot wallets, 100 hot wallets and a system which shuts down entirely the moment it sees that you had a certain level of activity across the wallets and suddenly the activity went, you know, out of the ordinary. Shut it down.
No, they they had the design in a entirely you know operational perspective not an under attack perspective.
If all this is known for years how can you explain why why so little was changed on the ground through over the years especially when we're talking about institutions. Oh
why so if the problems you're talking about are known for for years why so little was changed over the years when you're talking about institutions do you have your answer
it's a very difficult question to answer why hasn't it changed because people follow what's written in the standards they follow what's written in the documentation and they're very rarely they're like hackers are creative people they're, you know, getting incredibly wealthy, stealing millions and hundreds of millions of dollars and remaining out of jail depending depends on breaking rules and breaking expectations from from the system. The job of the IT guys that are designing the system and developers is to make sure everything works. So they're completely different spectre of mindsets. So for something to change, we need to start paying people to think creatively about the problem and to break the system. Okay, you have developers, hire a few people who will day and night attack your system from inside.
It's not I'm not talking about penetration testing because penetration testing from a cyber security perspective also most of the times use documented tools, documented procedures, methods. You need to hire people who are used to think outside of the box and attack the mental models that created this architecture.
Okay, one last uh question for all of you, the same question. If our audience need to remember one thing from your speeches today that is very very important Mr. Bab let's start with you one thing that they should remember
for today and for tomorrow.
Right. Right. So, so on the institutional side, I think one thing they should remember is that they should really start looking into uh the quantum technology that exists today, the quantum signatures and they should start um planning uh a migration um for securing their their data with postquantum cryptography. Mr.
AI is coming for all of you and us. [laughter] Uh and just to elaborate a little bit more what I said in the beginning of this conversation, unconstrained LM models are getting incredibly good. And until today, all the breaches that happened in all the systems happened with hackers without access to something that made them 100 or a thousand times more powerful. Now they have this tool and it's urgent. yesterday.
You should buy the hardware to run your own LM systems. The unconstrained defense just like they are doing unconstrained attack against your infrastructure. You shouldn't do it in a year. You shouldn't do it in two months, in three months. You should be doing this yesterday.
Dr. Toros.
Well, I would add AI is coming. Quantum computers are coming too. So, yeah.
Thank you very much. Do we have an uh do we have an additional questions? No. Okay. Thank you very much.
It was very interesting and thank to the audience.
Automatic transcript — names and jargon may be misspelled.