Banks Are Breaking📉Algorand CEO Explains AI Threat

0:00 Could quantum be a problem for banks and crypto? Today, we're going to be breaking that down with a very good guest today. It's going to be a good one.

0:06 Let's just get started.

0:07 I want to thank our sponsor today, and that is Tangem, where you guys can get into full self custody, very easy to.

0:13 Do, your secure crypto hardware wallet.

0:16 And one of the things about Tangem that I think a lot of people misunderstand.

0:20 With self custody is the.

0:22 Ease of use and also the ease of being able to get this set up for you. It's like a credit card, works with your phone, always present. Make sure to get a three-card set so you've got a backup. But you guys, of course, can use our link down below. It's going to give you a 10% discount. And they also have several offers from time to time. So always check what's going on over there on the website. This is the self-custody to use. So make sure and check it out. Tell them we sent you. I want to bring in William Herkelrath, who is the new CEO over at Algorand. It has been a while since we've had someone on from Algorand.

0:55 So William, welcome to the show. How are you?

0:58 I'm quite well. Thank you very much. Appreciate it. All right.

1:01 Well, let's get into a couple of things.

1:03 We're going to learn a lot about what Algorand is doing around Q day and what kind of quantum threats you guys are seeing in the market that are kind of helping you that. But I want to lead off with the first story here that kind of sets the tone here. And that is, is the U.S. banking system really ready for a post-quantum future? As you guys know, there is a lot of situations around quantum and the potential threat here When you look at the current banking industry, how far away are they, do you think, from full preparedness?

1:36 Oh, I think they themselves would admit that many of them haven't even started. So very far away is the easy way to answer that question. And there's been, I think, a lot of big debate as to when they have to be ready. So I think a lot of them are still trying to ask the question as to what the timings are here. It wasn't until recently... that people thought that what might be a 2032 problem or a 2033 problem is now probably more likely to be a 2028 or 2029 problem. So it's been very recent that people have really escalated that topic.

2:10 In the article we just showed, this was kind of an interesting point, and that is the result of an unusual contradiction. The federal government has imposed a deadline on itself while leaving the financial system largely to an advisory timetable.

2:24 So it appears.

2:26 The government's trying to prep, they've got a timeline, but they're not putting anything out there for one of the major industries in the US. And if we had a financial system shutdown or even the threat of a major bank shutdown, that would be pretty significant in the markets, much less just in general, I think to kind of call a lot of the fears of consumers. First of all, do you feel like the banks can really accelerate to get ready for this? Or is this something that is going to just happen and they're gonna have to work through it? Or do you think AI, may aid in being able to prep these banks.

3:00 I think ai is both a threat to having these things be broken sooner and it's also the aid that we'll probably be using to help us get ready for those threats at the same time i think ai is going to be a big part of this story regardless i i think my only guess on why they are focusing on their own uh systems before they decide what to issue to the banking industry as a standard There was some pushback. In Europe, they have issued standards and they have issued hard deadlines. And I think the banking industry was a little bit like, hey, listen, this is still a little unknown. You can't be giving us requirements just yet. It's still early days. So they might just be reacting to that. But the reality is that once

3:40 people decide what they want to do here, then the regulations will come. We'll get very specific rules, I suspect.

3:47 Yeah, I'm a little bit concerned because I don't think we see the kind of technology prowess in this administration that can kind of prep for this. I know that's what Sachs was supposedly going to be doing, David Sachs. But we don't really have a lead out there that's kind of pushing this forward. And I think we need it.

4:04 So this is going to be something to watch for sure. Q-Day could arrive before quantum computers are commercially useful. This is coming, obviously, in from the EU. So it brings up a topic that I think is important for our audience, and that is When you look at hacks that are happening today, many of the vectors that are out there are being hacked by both black hat and white hat hackers.

4:27 They're getting access to some of these vectors that are kind of honeypots of data waiting on quantum or even advanced AI to be able to actually pierce that data and utilize it. Is that something people should be worried about right now, even before quantum potentially becomes a problem?

4:48 I mean, that has a very common perspective. That's my perspective, yes. Any data that's stolen, even if it's encrypted today, and you can't decrypt it today, can be decrypted in the future. So that is exactly what that harvest now, decrypt later references. And people are acquiring huge amounts of data with the notion that they'll be able to get at it later. Absolutely.

5:11 I know that Algorand has worked a lot to try to advance your preparation for QDAE.

5:17 But how close is it?

5:19 This is kind of isolating down in three papers, three months, now rewriting the whole quantum threat timeline. And there's a few points right here that I wanted to hit on, especially around crypto, which is talking about major crypto, including Bitcoin, Ethereum, virtually every major cryptocurrency could be broken with fewer than half a million physical qubits. So when you look at that, And the potential here, how close do you think Q Day is? I know you guys are preparing for this.

5:47 So what is your target zone?

5:52 So we are going to be fully, I think, completely ready by the end of 2027 to be perfect. Honestly, we'll be 90% ready much sooner than that. There's things around the edges that we feel like we're going to really, really need to do to harden it. But we are already, all of the accounts on Algorand today are already quantum resistant.

6:08 The wallet infrastructure is quantum resistant. So the most important pieces we're already ready today with. And around the edges, there's work to be done in the next few quarters.

6:19 I'm kind of curious, where do you find developers that can develop against quantum risk?

6:28 Actually, the math is the hard part. It's not to downplay the development effort, but it's actually understanding how the math works and making sure that you can do that correctly is where a lot of the challenges. So it's the cryptography space rather than the developers. Once the cryptographers have worked it all out, the developers can bring it to production in a reasonable timeframe. But we've got some great folks that have been with the Algorand for four or five years. We've been doing this for a long time.

6:55 This is not a new strategy for us. We did our first state proof, quantum state proof in 2022. So we're two or three years into our journey already.

7:04 Other issues that are happening is we've got banks now being hacked all over the world, really, open source AI agent hacks in South Korean banks. This has exposed a whole bunch of records. When you look at that, first of all, are you guys examining how some of these hacks are taking place so you can use them as learnings?

7:22 What are you watching right now?

7:27 To be perfectly honest, the most important thing that we're working on is the math, is the cryptography, is the protocol level to make sure that it's functioning. And it's also things like compression sizes, bit sizes. There's a lot of work here that people I think are probably overlooking. All of these calculations are much more complicated, higher level math.

7:47 And a lot of the systems that we're used to, we're used to instant finality on everything we do in our lives, right? As we do these new types of handshakes and this new type of cryptography, we have to prepare ourselves that if we don't do the right things, things are going to go slower. So that's what we're focused on more is making sure that our thing works really well and it works really, really quickly. And we haven't spent too much time. I think the attack vectors are pretty well known. So we're more in the solution space than focusing on the attack vectors.

8:16 For our audience, I think people need to be aware that there's a lot of fourth party and even third party incidents that are occurring. What we mean by that is it's not necessarily the banks themselves in some cases that are being exploited. What you're dealing with is some of the supply chain that's happening here. So that gets into a whole new layer because now you've got supply chain vendors, people that maybe don't necessarily deal with this on an everyday basis, especially around security. So it is a challenge right now. Do you think that is the weak link right now across both quantum and advanced AI?

8:53 Well, absolutely. First of all, and you mentioned at the beginning that the government is saying that they need to be quantum resistant by some sort of date. Guess that means that everybody who sells to the government, everybody that has them as a client is probably going to be asked that very hard question as to what extent they have solved all those problems if they want to get those government contracts. So This is how it tends to work through the supply chain is that the vendors start to become basically demanding of it. The banks will do the same thing. The banks will require all of their vendors to be quantum resistant and that's how it'll force everyone to get to that level.

9:28 That's what concerns me because every project I've seen the government launch, 99.99% is much further out in terms of their target date. even if they're trying, which they don't really have a target date right now, even if they're trying to get there, this seems to be just comfort speak trying to get people to say, okay, don't worry about it. But in reality, this could be much bigger overall.

9:54 So that's a concern I have right now that I think is going to be an issue going forward.

9:59 What about post-quantum security?

10:01 Once we do get to a point where we have a Q-day occur, the kind of security mechanisms that could be put in place, even with companies like yourself, what do you guys, do you have kind of a plan after this occurs?

10:16 Well, so basically what we're doing is making sure that everything that we build right now is already quantum secure, right? So as we bring new features to market, each and every one of those new features is going to be quantum secure, which means that we don't need to worry about exactly when Q day comes. The day it comes won't impact us because our systems won't be vulnerable. That is the same recommendation we're giving to our partners. That is pretty much the thread that's running through our entire organization at this point. And that is, again, all those little edge cases that I mentioned that aren't yet PQ secure, that's what we're harping to make sure that we do meet those timelines.

10:51 Are you guys at all worried about the social engineering aspect of it?

10:56 Revolut, this is a good example of it.

10:58 They exposed a lot of KYC data after they had a fraudulent government email that was passing their security checks. So basically what it was was social engineering that worked its way through. It wasn't really a hack. It was basically just social engineering, which has been around for 30 years. So That seems to be almost a human element that could end up piercing some of these veils that could even get past this quantum secure layer. So are you guys trying to deal with that?

11:29 There's only so much we can do. Our specialty happens to be in the post-quantum area. That's the AI threat that people are talking about. The fact that AIs are getting better and better and better at figuring out. The one thing I think how where these overlap is that if you... AIs are finding... tremendous vulnerabilities in existing 32-bit encryption, which is the most common encryption that's out there today.

11:53 And not because they're brute forcing it, because they have powerful computers, just because the AI is finding these ends, social engineering and other ends into that. And so people are asking themselves, well, we have to fix those holes anyhow. If we're going to fix those holes today, are we really going to do it first on the old cryptography and then do it again three years from now in a post-quantum way? So, so I think what we're hearing is some demand is like, well, we might as well go, let's, let's look at post-quantum right now. If we're going to do this anyhow, let's look at it. Let's look at the full solution as to what we need to do over the, you know, and get that done today. So that's,

12:26 it's escalating, I think, demand for, for what we can bring to the table, which is, you know, I guess positive for us, but, but no, the reality is social engineering is something that, that no, no piece of math can get around. If an AI figures out how to get your passwords, there's not much I can do about it, unfortunately.

12:42 I'm looking here at your resume.

12:44 And it's got two decades of AI enterprise, institutional finance, blockchain.

12:52 You were a co-founder.

12:53 You also did some leadership roles at Chainlink and Curve. So when you look at the overall industry, you know what's up. When you look at the overall industry right now, who would you say are the blockchains out there that are closest to a solution outside of what you guys are doing at Algorand? If you look at just the industry in general.

13:15 So there's not that many. I mean, we are really proud to be the forefront. That's kind of why you said I was newly in this role. This is the one I wanted, right? What was my next job? What is that I wanted to do? I wanted to be in the post-quantum space. And this was kind of the only play. The NEAR protocol has got some great stuff.

13:32 They're also working on some really important things. There's other parties out there that have some things working in labs, in a labs environment. But as far as Being fully post-quantum, working in production, doing millions of post-quantum transactions, there really isn't anyone else just quite yet. That's where we are.

13:49 We're far ahead of the game. Again, we've been working on this for three years. I guess the one thing that I get nervous about is that everyone thinks they're going to be able to make this transition quickly and easily. It's just not that true. It's taken a long time to get us where we are. Other people will be able to do it faster because they'll be able to learn from what we've done. But it's not a two-day process.

14:10 It's a very long process to work yourself down this road.

14:13 Okay, so you look at a lot of these models that are being advanced right now in AI. So devs, as well as the math that's going into this kind of development, is being advanced pretty significantly. And I would say by this time next year, you could probably take that.

14:30 On an order of magnitude.

14:32 In terms of growth and opportunity, wouldn't that allow a lot of these projects to be able to catch up as long as AI is actually advancing fast enough?

14:42 I think that that is absolutely their belief, and there are many people who have that belief. I actually don't agree with that philosophy, personally. I don't think that... Because my experience, while I think you can... Doing the math and actually knowing how to put that math in practice are completely different things. Building an algorithm that works, but versus building an algorithm that can actually achieve finality in less than one second are totally different things. There's a lot of art to this. It's not just hardcore brute force math, and it's not just hardcore brute force engineering, which is what AIs specialize in doing. Obviously, they're getting better and better. I do think they're going to be helpful.

15:23 There's no question. But there are real challenges, particularly with blockchains and decentralized situations that that require, there's a lot more work here than I think people are aware.

15:34 Do you think it's being overblown at all? Because you hear some analysts out there and researchers that are claiming that quantum is a very narrow vector and in many cases it could not necessarily advance to where it's going to cause problems in all these different areas. It's designed to really cause, or excuse me, answer very narrow, explicit questions particular challenges, do you think it is being overblown or do you feel like this is, no, this is the thing we have to watch out for?

16:03 I think both things can be true at the same time. I think that the problem with it is that even if it's only a 1% chance, it's complete and utter failure. I mean, like I think it's 1% of a complete and utter disaster. Like when we were talking about the AI, one of the AI guys said, well, look, there's only a 10% chance that AI is going to wipe out the human race. Even if that's a 1% chance, I'm not too happy with that. It doesn't have to be a large percentage chance for us to still need to take this seriously. Look, I think it is probably coming faster than people think. Every time I read a new paper, someone has figured out how to bring this Q date forward an extra year or something. To not prepare for it,

16:49 I think, is pretty naive. From what I can tell, every Fortune 500 CISO has to have this on their 2027 roadmap. They have to be spending money in this area in order to keep their boards happy. So we know that there's major investment coming this way.

17:04 You guys have announced that Algorand says that your mainnet has passed more than 2.7 million post-quantum transactions.

17:12 So you said.

17:13 Earlier in the interview that you're 90% there right now.

17:15 Is that right?

17:20 What I meant to say is, so that is absolutely true, the 2.7 million post-quantum transactions. I'm saying that 90% of all of the things that you would need to do on a blockchain are already secured by post-quantum. To get us to the full, to get us to that end, that's what we continue to work on. There's other little areas in and around digital assets that we're working on to get us the full spirit. But we can already do what most of these institutions want to do with prospective digital assets. We've got the minting and the burning of the ability to hold those in custody, to interact with those in a fully post-quantum manner, that is already live and in production, and that's what we're pushing out

17:56 to the market right now.

17:58 So, William, you've been in the business for a while.

18:00 You've seen kind of all sides of it from the finance, the tech, and obviously the.

18:03 Advanced narrative around blockchain.

18:06 What would you say to banks that are out there trying to prepare for this right now?

18:10 What do they do first?

18:13 Well, there's a lot of complicated things that they need to do just with their internal systems that's going to be solved by the Microsofts and the Oracles of the world. That's not problems that we're looking at. But I will say to them with respect to digital assets themselves, if you are wanting to work in the digital asset industry, if you're tokenizing your assets, you're creating stable coins, you're working in that area whatsoever, and you're doing so all on the old cryptography stack, you're going to have to redo all of that. And so our message to them We would love to get in that office right now and help show you how to get this done in a post-quantum way today. So make sure you handle this.

18:52 Get this done. There's no reason to be incurring all this tech debt or putting your clients' assets at risk using the old stack when the new stack already exists. That much we can solve today. The other issues in their systems are not things we're focused on just yet.

19:07 We've had a chance to interview a lot of different people on this very subject, and especially around AI and some of the vector exposure that's happening around these hacks. Do you foresee a major AI hack on a major bank in the U.S.

19:22 Before the end of this year?

19:25 I don't know the timing, but it's going to happen. I can't see the timing, but I think so. I mean, look, these things happen all the time. We've We're both old enough to have seen them happen multiple times.

19:37 We haven't seen one of those happen, though.

19:39 A major AI hack where you see real Americans lose access to their accounts, possibly maybe just lose the entire amount. Who knows how agents could get in there and and potentially work. But this is the concern I have. And it's, I think a lot of experts out there have kind of voiced this is that the banking systems are not ready for this.

20:00 Yeah, no, I think you're right. I don't know the timing of it. I will admit, I don't think it's a quiet, you don't think that's not a quantum threat. That's just, that's just the generic social engineering threat, the AI threat. Um, I think about that a lot and there's no doubt in my mind that there's going to be, that's why I actually, I don't like the word Q day because that implies that there's some magic day where all of a sudden everything's at risk. I think we're just going to see these incremental risks are just going to grow. It's more of a spectrum rather than a single point in time.

20:29 All right.

20:29 So you guys have unveiled your roadmap for resilience by 2027. Does that mean you're going to be full quantum resistant? by the end of this year or is it going to happen next year?

20:41 So as I said, all of the digital asset work, the vast majority of things that banks need to do with digital assets today, we already are quantum resistant with. Full quantum resistance, there's things around the edges again, like the way that consensus is exactly calculated, the way that randomness is assigned to validators. These are highly technical things.

21:02 Those items are going to be coming in the first few quarters of 2027. So that's why we are giving ourselves a little bit more of a buffer, but most of what people need to do today, we already are prepared and ready for providing them a fully post-quantum solution.

21:17 So how would, okay, so explain to our audience for Algorand to be able to solve this problem and let's go with it, you guys fix this, or you have something that's in place for 2027.

21:31 What would be the effect on the market.

21:34 Especially the financial industry, How would you guys go out and assist the rest of the market to kind of prepare for this?

21:42 So we provide a tech stack, first and foremost, that the financial industry can use. Specifically today, that tech stack is focused on digital assets only. We are thinking about other markets we can get and be helpful with throughout 2027 once we kind of nail down the digital assets side of the equation fully and completely. And that tech stack is such that those companies can be working with it effectively right away. So just like they would be working with the tech stack on Ethereum, or they might be working on one of the other blockchains, ours is one that they can work with. And we have a multi-chain solution so that it's easy to be interoperable with ourselves and some of the other blockchains out on the market.

22:23 But that's effectively what we offer. And so when we say we're going to help them, we can do consulting, but we can also just bring them in, physically give them this tech stack, give them a custody, give them a wallet solution, give them what they need and a key solution to get to market.

22:38 So it sounds like you guys are prepared and maybe digital assets really are in more of a good position to be protected first. Beyond that, is that correct?

22:49 Is that what you're saying?

22:51 Well, that's what we've solved for first. I agree that there's another big market out there that we're excited to go help with, but that happens to be what we've solved for first, and that's what we're focusing on first.

23:02 All right, with that being said, obviously you guys know who your target zones are in terms of web, we'll call it web two companies that are out there, whether it's the banking, the medical industry, the real estate, hospitality. I mean, there's all sorts of companies sectors out there. Is banking the number one target for you guys once you start to move outside of the blockchain element?

23:25 Yeah, that's our current expertise. That's where we know the best. And I think that's where the threat is the largest, to your point. Are you anticipating any deals in 2027 with major banks? I think so, yeah, absolutely. Can't announce anything, but conversations are very positive as we stand right now with how we can be helpful to them. They need to be learning. They need to be working on this.

23:55 They have budgets in 2027 to bring in partners and bring in experts and be testing stacks and getting POCs ready. All of that activity is going to happen, whether it's with us or whether it's with other people. These deals will be announced. Everyone wants to say that they have something on that roadmap next year.

24:12 Well, you know, this is going to be such an interesting flip of the coin if a crypto company is the savior to.

24:22 The banking industry. I mean, that to me is just perfect. It's the way the story should end.

24:28 I like to say that the purpose of us is to bring the word cryptography back into crypto. I think crypto, unfortunately, has that kind of a negative connotation, but it did come from the word cryptography. The world's, many of the world's cryptographic experts are obviously working in our industry. So it shouldn't be too surprising that the crypto industry, which is a cryptography industry, would have some expertise in this area.

24:50 Yeah, I love it.

24:51 William, thanks so much for stopping in today. We're going to keep a close eye on Algorand and what you guys are up to because this could become something big and very fast. So thanks for coming in.

25:00 Amazing. Thank you so much.

25:09 If you like this video, hit like and subscribe.

25:11 Drop a comment down below and also join our free private member group.

25:15 The link is in the description.