We recently heard the argument that it's simply too expensive to develop a cryptographically relevant quantum computer. We vehemently disagree. In this episode we explain why.
[Tim Callan] (00:00.08) So, Jason, I recently had a conversation with someone. This was about post-quantum cryptography, and this person made an assertion that I did not buy, and I thought we could talk about it here. Let's talk about it. So the assertion goes as follows. Basically, the argument goes as follows, which is: hey, listen, all this excitement about post-quantum cryptography is unnecessary, and the reason it's unnecessary is that it is a very hard task — a very hard engineering task — to get all the way to a crypto-relevant quantum computer in terms of the number of stable qubits we require. And for considerably fewer stable qubits, we can realize other very real benefits of quantum computers on other problems that exist in the world. And therefore, what we will do is — there will be a race, as there is now, to get to that number of stable qubits to solve those problems, the medical applications and things along those lines — and then at that point, progress will halt, because there will not be economic incentive to continue the very expensive task of expanding the number of stable qubits. So that's the argument someone made to me. And therefore — the 'therefore' is — and therefore we don't need to do PQC. So, do you have any reactions to that?
[Jason Soroko] (01:37.80) Yeah. I kind of have two, but go ahead. I think we're probably the same, Tim. But if you're a nation state and you live in a world where it's so strategically, tactically important to be able to encrypt and be able to keep a secret from your adversaries in a strong way — and also, if you're a nation state that belongs to an economy that's based off of being able to have privacy and secrets and transactions that have integrity and uptime — because don't forget, this isn't just document signing, this is also authentication.
[Tim Callan] (02:32.75) It's also keeping the identity of all these things straight. In the long term, it's not just harvest and decrypt. It's much more than harvest and decrypt — and just the harvest and decrypt problem is bad enough. We have had multiple podcasts on all of these topics.
[Jason Soroko] (02:47.00) So to think that it's not economically viable to push quantum computing to the point where we have a cryptographically relevant quantum computer is, I think, almost bizarrely naive.
[Tim Callan] (03:07.44) Yeah, I think if you even were to make the case to say the private tech companies that are working on this — the Googles and Amazons and Microsofts and IBMs of the world — even hypothetically, if those companies would hit a point where they said, 'We've got enough qubits, we're done investing' (which, by the way, I don't even buy that), but even if that happened, that still wouldn't solve the problem, because there are 20 nations in the world who view this as an essential military component, and they're not gonna stop doing the work. If there was no commercial industry involved and it was just military — oh, we're going to have a cryptographically relevant quantum computer. Absolutely. Right? Just from the military application, it's going to happen. And that military application will not be directed only at military targets. Right? The point will be: I want to be able to use this against other nations in every way I can think of, and that absolutely will include stealing secrets from individual companies. That absolutely will include, you know, the potential for attacks on OT and infrastructure and electric vehicles and all of that. Right? All of that will be in the playbook.
[Jason Soroko] (04:35.52) Tim, let's give this idea — you know, it's so bizarre and so naive — let's give it a chance. You know, let's do a mind experiment.
[Tim Callan] (04:46.00) Sure.
[Jason Soroko] (04:48.00) Okay. Tim, I just invented a quantum computer that can fold proteins, and I've almost cured cancer. We're done, right?
[Tim Callan] (04:54.43) Yeah, yeah. No, of course not. Absolutely not. And then, for a couple other reasons, right? One is — and even if we look at those, let's say, above-board tech companies we know about, the ones, the big tech giants, the ones I listed — do we really think we're not gonna find new useful applications for qubits that are beyond breaking cryptography, that will require more qubits?
[Jason Soroko] (05:21.51) Of course we will. Tim, I'm going to offer you this. So I offered you, like, the minimum case, which is like a standard healthcare neurology problem that has a finite set of qubits that it would require. You're talking about cryptographically relevant quantum computers, which is maybe higher than that, but not a ton in the grand scheme. I would say that the thing that makes this even more naive of an argument is that some of the problems that science is trying to consider require so much Fourier-transform stable qubits — when you're trying to model universe-level stuff...
[Tim Callan] (06:13.04) You need a colossal computer for that — humanity wants that. And in addition to that, there also is the fact that nobody can predict and understand exactly where science and knowledge and tech are going, and need to go. Right? So you can't sit here today in 2026 and know what questions we will be asking in 2035. You can't. And so, for all those reasons, I think that's just not a safe assumption. I replied with a thought experiment. I said, as follows. I said: let's pretend that you — because the original objection was like, 'Look, this is gonna cost billions and billions of dollars, and nobody's gonna spend billions and billions of dollars' — and so I said, let's pretend that you were this unmitigated genius cryptographer, and you had, in your own office, invented a way to reliably defeat RSA and ECC, and let's say you had it in your hot little hands today. Let's furthermore say that you were prepared to sell it to the highest bidder, and you had no concerns about ethics, criminality, national identity, harm that came to people — you know, any of that — and you were just plain gonna get the most money you conceivably could get for it. Let's further pretend that you had an efficient way to connect to all the potential buyers who might pay money for this, so that you could get exactly what they were willing to part with to have the secret. Do we think there is any way that you don't have dozens of buyers who are prepared to spend $10 billion and up to be the one who has this magical capability?
[Jason Soroko] (08:12.77) Tim, I don't think a trillion dollars is even a whiff. Absolutely.
[Tim Callan] (08:20.08) If you had that level of dominance over everyone else — again, just from a military perspective — that's a cheap trillion dollars. It's a cheap trillion dollars. So I think it's important to keep these things in mind. And let me just say — since the person who did challenge me on this is a listener of the podcast and might be listening to this episode — I think this may have been posed as a little bit of a thought-experiment-level challenge. I don't know that this is somebody who honestly believes it in his heart; I think it's somebody who's testing my ideas. So that's fine — I'm happy to have my ideas tested. But it was a new angle on it, and it was a creative enough question that I thought we should surface that for all the listeners to think about. And I think it's a valid question to ask. But my response to that is that it's a resounding no — absolutely they won't stop. Absolutely they won't.
[Jason Soroko] (09:19.00) There is no economic reason why we should not achieve a cryptographically relevant quantum computer, Tim.
[Tim Callan] (09:25.00) Yeah, absolutely. I agree. The economics of it are not going to be what prevents us.
[Jason Soroko] (09:31.86) Great thought experiment, and great thinking. Thank you, Tim.
[Tim Callan] (09:35.06) Thanks, Jay.