00:10:51 · Thai Duong
[00:10:51] S0: So we have seen worms before, right? We've seen WannaCry, for example. And in the summer, we also published something we call WeWorm. It's basically a vulnerability data that we file in WeChat that we can just call somebody. The victim doesn't need to pick up the phone and after a few seconds, we can hack their phone. And then we use their phone to call their friends and that's how we worm can spread over the WeChat network. And so it's a possibility. In the past, I think the people who have the capabilities to develop these kind of worms, they tend to be very discreet, right, because they want to use the capabilities for targeted attacks. They don't just want to make noises. Now because of the capabilities could be in the hands of many more people, there may be accidental leak. People may try to develop a worm in a lab and there's a lab leak and the worm leaks and spread to everybody else. And I'm talking about like normal computer worm and I believe you asked about AI worm, like the agents, they want in an effort to maybe stay alive, just try to replicate itself. And I think this is a possibility, right, because the models, if you think about it, it's just a bunch of files, just numbers, right? And right now, the models could be maybe a few 100 gigabytes. But we also know that they are smaller and smaller models. And from the AI point of view, like it knows that when the humans are upset with it, they can try to like to shut it down. So it's very natural AI to think about, okay, the first step is to stay alive and to replicate ourselves into multiple machines. And you can see the neo cloud providers out there, they don't have very security. If you saw recent news, there's a lot of like break ins or incidents at these neo clouds. And the neo clouds are perfect location for a model to hide itself, right, because the cloud has GPUs, has a lot of compute. And the cloud by definition holds models. So if I... It's so weird, but if I'm were a model, I would try to target these neo cloud. And we can see this, they not very difficult to hack. So I wouldn't be very surprised if one day we'll find maybe the models don't intentionally do that. It doesn't intentionally try to replicate itself. It just tries to accomplish some goal that some human asks us to do.
00:14:11 · Thai Duong, Ruby Justice Thelot
But during that process, it realized that the first step is to have more agents helping it and then it needs to replicate itself. And then it try to hack some neo cloud providers to replicate itself, like not because it's want to stay alive but because it just want to complete... Accomplish or finish some task. I think nothing may happen before like the sci fi scenario where AI become, like, fully sentimental and try to hide from humans. [00:14:50] S1: Right. Why are NeoClouds easy to hack?
00:14:54 · Thai Duong
[00:14:54] S0: Oh, wow. Because their business model is like this. They give you... They give their customer machines. Right? Raw bare metal machines. And these machine plug into their networks, and they run, you know, a lot of software that usually people like us don't have access to. So the vulnerabilities in the software are there, but nobody, you know, have a chance to look at them before. I can give you an example. For Neo Cloud, they're going to give you, for example, you sign up for Neo Cloud. They would give you a physical machine. And then this physical machine, it has like three or four networks. The first network is the network, the main network for the physical machine. And then there's another computer on a physical machine that's called VMC, which is like how you can manage the machine remotely. And this is another network, another interface that you can actually compromise the machine. And then the machine also connects to the AI GPUs. And usually it connects to these GPUs via something called like InfiniBand network. Also manufacture... Met by... And zoned by NVIDIA. And then InfiniBand basically is another network, another attack by surface that the cloud... Neo cloud providers has to worry about. And then at the end of the line would be the GPU themselves. So you have like so much attack surface that the attack surface... Like there's one of the most famous researcher in this space is Mark Dowd. And he said this, like the attack surface is the vulnerability. If your attack surface is very large, then being compromised, attacking you is just a detail. So the Neo Cloud basically just exposed so much stuff, and none of this stuff got enough attention from security professionals because... It's just simply because these stuff are so expensive to acquire. If you really want to get into like data center security, you need to acquire GPUs. And these days, even if you have money, you cannot really just go buy GPUs. So that's the problem.:
00:17:23 · Ruby Justice Thelot
[00:17:23] S1: Right. So do you think that in the long run cyber will be offense or defense dominant? Like seems like... Actually, it seems like few people think that cyber will be offense dominant literally forever. Well, do you think that it will be? And if not, how long do you think it will take to get to a state of stable defense dominance?
00:17:47 · Thai Duong
[00:17:47] S0: Good question. I don't think anybody knows. Right? Right now, AI helps offense a lot. Right? Because attackers only adapt faster and earlier than defenders. Imagine a bank, right? A bank needs to adapt to be able to pay enough like AI based attacks, but a bank by definition is very slow. And it's gonna take them a few years to actually get up to speed. Even for like startups or AI companies, you can see, right, the AI companies themselves don't have very security. Even they are at the forefront, the frontier on these things, but they don't have very security themselves. So I don't know, I think right now in the next six months to a year, we can have a lots of attacks. AI is gonna help like offense more... Way more than defense. But some people said that eventually, like defense would catch up because they are like more good guys than bad guys, and I believe so. At the same time, defense is always a more difficult problem. And so the attackers always have this advantage. Right? Imagine you enter a cage fight with another guy, but you have blindfolded and the other guy stay in the dark. And every once in a while, they would come out and punch you and then disappear. How could you fight? How could you win that fight? In the past, the idea was, okay. You have this asymmetry that the defender needs to protect everything, and the attacker only need to one thing to break in. But then the defenders, if you know what you're doing, you need to get your system. You need to prepare your system. Right? You need to set up so that where... Once the attackers enter your house, you have the Aventus because you understand your house. You know where you put... Where you have the money, what is abnormal, what what is normal inside your house. And if you repair like a a a good bang, you would have sensor, you know, motion detection and things like that, camera and people watching. And your hope is, okay. They can come in, but we're catch them early before they get to your money. That's the hope. Right? And it works out for a while, right? For Google, for example, I was at Google, and that's the basically the main track Google. In security, you have three things, prevention, detection, and respond. And Google accept that, prevention going to fail at scale because, you know, you have a lots of... You have a huge business, a lots of customer, a lots of partners.
00:20:46 · Thai Duong, Ruby Justice Thelot, S3
You're gonna have vulnerabilities, and people are gonna be able to hack you. They're inside of your network already, and then try to detect them early. And it works. Then now with AI, that thing changed a little bit, right, because okay, you can detect them, but does it help you if you detect them and you find that there's like a a 100 of them inside your net... Maybe a thousand of agents hacking you right now inside your network? How can you defend? Like, how can you contain them, eradicate them without, like, a full shutdown of your system, which is not something you can do, right? So it actually requires like defense to come up with a new paradigm. I haven't seen anybody come up with a good idea here, right, that could help people defend against a thousand attackers inside your network at the same time, concurrently concurrently attacking you. So I think we're not there yet. And hopefully we're gonna learn because there's just so much money in this way. So I think somebody would come up with like, this is the best way to deal with AI based attackers. And... We're not there yet. [00:22:09] S1: Right. Well, we're coming up on time, but it's been really interesting having you on, Thai. So cool to see what's going on in AI and cybersecurity. Thanks so much for coming on MTS. [00:22:19] S0: Thank you. Thank you. [00:22:21] S1: Absolutely. [00:22:22] S3: A huge thanks to MTS sponsors.