For use in teaching asymmetric cryptography to b-school students via an injection of snake oil. A transcription of audio from a recording (part 1, part 2) of Robert Grant’s BlackHat 2019 presentation about Time AI. Read Bruce Schneier’s coverage and links therein for context.
All not-italicized text is Robert Grant.
Please submit a PR if you can make out the parts that were unintelligible to me!
[Part 1 26:22, start of Q&A session]
So this morning undersecretary Mike Griffin said on his twitter feed, “what I want to bet on is the ability of the united states and our western allies to out-innovate our adversaries.” I think that’s really important and it’s not just a national level but I’m excited for Time[.AI]’s timeliness to be able to impact [… unintelligible …] so we will go after the new market but the market is so large already that nobody knows encryption… rant… like, nobody knew botox but everyone wanted a botox grant but this is something that I think is so important about sovereignty. So with that I’ll open the time to just a couple of questions real quick. Yes, go ahead,
Speaker 1: […] I understand correctly that all encryption is based on factorization [yeah]… including Diffie-Hellmann…
Diffie-Hellman exchange, that’s… more of the exchange methodology and I think a lot of people refer to that but yes it’s all based on factorization of prime numbers. including Diffie-Hellman, including RSA, including elliptic curve.
Speaker 1: AES too?
AES is not! I’m talking about all asymmetric key, um, encryption.
Speaker 1: So are you aware this is wrong. [Sorry?] Diffie-Hellmann is based on discrete logarithm. [Sorry I can’t – parle france?] Oui oui oui.
Okay [chuckles]. So, but you can use Diffie-Hellmann exchange for public key cryptography too. I’m talking about public key cryptography, and asymmetric… key. Right? So that’s different. Symmetric key is great. AES encryption, fabulous, no issues there. In fact, Time.AI converts… ah, all of the data that’s transferred into an AES encryption automatically so it’s embedded within it. Yes.
Speaker 2: So, can you uh … [unintelligible] which is, you can determine whether a number is composite or prime using your new method. [Yes.] Um, how does that compare to AKS?
Um, I haven’t used AKS so I can’t say that I can compare it. But I can tell you that we don’t have to go from the other way around, which is you don’t have to start from the number and work backwards, we can predict from the beginning, outwards. Right. So, we know its position. Now we’ve also done other analysis and found that there’s a mirror reflection on when quasi primes show up versus prime numbers. And it’s around the number… six.
Speaker 2: I may have misspoke. I mean if I were to give you a number and ask you if the number was prime, you’re able to use your method to determine that. [Oh absolutely] And do you have like Big O notation like the worst-case for how long it would take to determine the number…[interrupted]
We haven’t done any speed tests on that yet. Again, once we knew that an index could be created, ah, for this, we immediately went to developing the new encryption. Ah, but we have actually worked with the amount of time that it takes, so we know, ah we also want to be really responsible about this, it’s not really about factorizing anymore, I believe there’s going to be – very shortly – equations. Not algorithms, but equations – that actually geometrically solve the solution of this entanglement problem between sum and product. Yes.
Speaker 3: On the day that those equations are found, will you release those publicly?
That’s a good question, but I can tell you that there’s a lot of people that would be very sensitive to this particular topic.
Speaker 3: Not the equations, your technology. If there was an equation which breaks all encryption, will you release your encryption.
Well, not all encryption – public key, right, anything factor based, right. Would we release our encryption’s equation? It depends. I don’t know the answer to that question. We have considered doing white papers on our encryption. There are certain… parties… that we’re in talks with right now that may not want us to do that. I can tell you that Microsoft is working on a non-factor-based encryption I’ve heard about. I’ve heard Google working on it too. I don’t think they’ve been publishing that. So I think that that whole concept of, you know, Kerchoff principle etcetera, I think is going to be challenged very soon, because, people are really starting to look at this as trade secret, more than anything else. So I can’t answer that question right now, and I don’t want to venture it right now, because I can’t say what our partners would want. We’re looking at doing partnership with several companies right now, many of whom you would know.
Speaker 4: [unintelligible] Microsoft and google are working on that, [unintelligible] poster page in 1976, [unintelligible]
Actually they just went back to one, probably because they know about this prime factor issue, they went back to the 60’s.
Speaker 4: [unintelligible] 1978, asymmetric encryption based on things other than factorization… [unintelligible]… maybe there is a bit of pure art. Possibly, [unintelligible] prime numbers, there is a bit of pure art, [unintelligible] so it was described and published 23 centuries ago. So maybe you should… Google some of…
Yeah yeah that’s part of it, but the quasi prime was not. So the prediction of prime numbers was not. Actually, the first time that anyone published the issue of the 24 hour clock, which was not our invention, we cite in our paper, Peter Plichta who actually published that, I think in the 90’s, he’s an old chemist from Germany, but the advent of the quasi prime, and being able to predict primes was only in this year.
One last question then we gotta go.
Speaker 5: So during your talk, you made allegations that RSA might not have a problem in fact.
Not just RSA, right, I’m not talking about the company RSA, I’m only talking about general factor based encryption. [Right.] Public key.
Speaker 5: Right. Can you substantiate that will some numbers, like, the security level – how significantly has it dropped?
I cannot, and I will not here. I can just say to you that… there was a great article that came out in the Economist last October, I can’t remember the name of the guy that wrote the article, but it basically was called, “Prime factorization… Quantum computers will break Prime Factorization.” But in the second paragraph it says, “in addition to quantum computers, any one of the millions of the professional and amateur mathematicians might have a stroke of inspiration tomorrow, publish an equation that unravels it.” I think that is actually more likely than what we’re talking about from a timeline perspective for when quantum computers are going to come out.
Speaker 5: How come you’re not following the standard procedure in academia where when there are findings in cryptanalysis, they are published?
Because I don’t have to.
Audience murmur: Come on, man.
Dan Guido, CEO of Trail of Bits: You uploaded that document to a website so you could get it reviewed though. You said you put it on a website that was Cornell-affiliated. And that’s Arxiv. That’s like me saying I put information on Google Docs, so Google reviewed it. You’re lying!
No I’m not.
Dan: This is a scam!
No it’s not.
Dan: It’s a scam, and you’re just using marketing material from this to trick a bunch of people months later, you’ll say you’re legitimate because you spoke at Black Hat. You should be ashamed of yourself, and all of you working for him should quit your…
[video break, start Part 2]
Dan: …somebody with a moderate knowledge of mathematics, […] every single thing you just said.
No, no. I’m sorry. I did not say, that all encryption is. I said factor-based encryptions. Yeah! Don’t lie, no. No, please. Please. So I’m just saying this, I’m just saying, you know what…
Dan: You shouldn’t be here. You shouldn’t be here, you should be ashamed of yourself.
You can say whatever you like, you’re entitled, you know what…
Audience member: Well let’s not be rude, alright, let’s just not be rude, let’s be respectable.
Audience member 2, addressing Dan: Who do you work for? Who do you work for!
Dan: Who do I work for?! I run a cryptography team!
So. Please have him removed.
[A burly man, who does not appear to be affiliated with Black Hat, removes him, without laying hands on him.]
So, the point is that, we don’t claim to know everything. Okay. All I can tell you is that we strongly believe that this new encryption is a great approach to be able to benefit consumers, it’s a novel approach, it’s real, it works as an encryption, and we will subject ourselves to regular testing. Right, we absolutely! And if it makes sense for us to publish it, right, because our partners who are reviewing it right now want us to, I’m more than happy to do that. But I’m just saying I want to handle this in a stepwise fashion. And as it relates to an equation that comes out for factorization, right, or speed of factorization tests etcetera, we’ll subject ourselves to all that too. But, we want to do it in a stepwise fashion. You have to be responsible. I wouldn’t even have gotten into this encryption field if I didn’t feel like it was absolutely necessary. The last thing you want to do is come up with some discovery, then all the sudden have everyone tell you, dude, you just like screwed up this stuff. So we’ve been relatively quiet about it on purpose. But now we’ve got an encryption that’s working, that’s great, we’d like you guys to try it, and it’s gonna have to stand on its own merit, just like every other encryption. So of course there’s gonna be people that are going to have their own skeptical position [gesturing towards door], it is what it is. But you know what, those are usually people who don’t create anything.
Thank you very much.