Today's number: 62. That's the percentage of influencers who report feeling burned out, according to a new content creator mental health study. Today's other number is zero. That's the percentage of Americans who care.
Welcome to Prof G Markets. I'm Ed Elson. It is April 23rd. Let's check in on yesterday's market vitals. The S&P 500 and Nasdaq hit record closing highs on the ceasefire extension. The Dow also rose. Meanwhile, the price of Brent crude rose back above $100 per barrel after Iran seized container ships. The dollar and Treasury yields climbed. And finally, cannabis stocks soared on news that the DOJ is moving to reclassify marijuana as a less restricted substance.
Okay, what else is happening? Anthropic built an AI model so dangerous they refused to release it to the public. We discussed that last week. However, we just learned that unauthorized users got a hold of it anyway. Bloomberg reported that a small group of unknown, unapproved users have had access to the model Mythos for two weeks. They broke into the model the same day Anthropic launched Project Glasswing, a controlled rollout giving a select group of companies exclusive access to the model. Anthropic says it is investigating the security breach from these unidentified individuals, but has found no evidence of malicious use so far. The question is, if they can access Mythos without permission, then who else can, and how bad could it get? Here to help us answer those questions, we're speaking with Sherri Davidoff, founder of LMG Security and co-host of the podcast Cyberside Chats. Sherri, thanks for joining us on the show here. So, we discussed Anthropic's Mythos model the other week. It's extremely powerful. How powerful? I don't know. But I know that it was so powerful that they refused to release it to the public. I also know it was so powerful that we saw meetings at the White House where they convened among some of the most powerful people in the nation to figure out what to do about this thing. And now we learned that someone accessed it, I guess, and it was unauthorized. I can't tell how much of this is real and how important it really is. So, that's why I have you. What do you make of this?
Well, I mean, this is something we've been expecting from day one. They say three can keep a secret if two of them are dead. Mythos preview has been released to over 40 tech companies and Lord knows how many people at those tech companies. So, from day one, I'm sure there was unauthorized access. The question is who is accessing it when they shouldn't be. Is it just people at those companies? Is it someone who has wormed their way in from the outside and has actually hacked in? But honestly, the part that I didn't expect was that we would actually hear about it. So, bravo to Anthropic for investigating and for whoever it is that figured this out and has started to publicize it.
Well, it wasn't them. It was Bloomberg. It was the reporters over at Bloomberg who figured this out. I mean, I guess if we could just go back a little bit, from my understanding, the problem with Mythos is that it can basically hack any cybersecurity infrastructure anywhere in the world, and you are an expert in cybersecurity. So could you just take us through, like, what is the danger of Mythos exactly, and how powerful is it? What are the capabilities?
Sure. Mythos is completely changing the security risk landscape right now. Now, I was just saying to somebody, I have been in this industry for 25 years and I have never seen something that completely altered the security landscape to this extent. So absolutely, we need to be paying attention and it is very serious. The capabilities of Mythos preview, according to Anthropic and those that have used it, are it's basically like a hacking ray. You can point it at software and even if you don't have the source code, it can still detect vulnerabilities and write code that will exploit it. So, a human doesn't need to do anything to break into that software. And that means that already researchers have discovered thousands and upon thousands of bugs and there are working exploits to go with that. So what that means is that literally anyone who has access to this could potentially just point it at wherever they want and break into that system before the vendor has a chance to even think about patching it. And so there are some defensive measures that we all need to be taking right now. And the goal by releasing it to some tech companies was hopefully to give them a leg up, but let's be honest, that's a lot of people that have access. And when that many people have access, it's going to get abused.
So are you concerned that this will get into the wrong people's hands? It sounds like your view right now is that people who aren't supposed to have it have it. But since we haven't seen a mass data breach or at least something that is extremely concerning yet, then it's not a problem yet.
Well, first of all, unauthorized access is absolutely inevitable when this many organizations already have access. And I've actually been wondering, for those companies that do have access, how are they being vetted? How are their security procedures being reviewed? How are they limiting this access? Because there's only so much you can do to secure any one company. But I'm hoping that Anthropic is maybe vetting them in this process. The challenge is you have to weigh that with the desperate need for these companies to start patching bugs. When you think about it, some of this software is used on millions and millions of systems, whether it's antivirus software or operating systems or systems in the cloud. And so when, not if, an adversary is able to find and exploit these vulnerabilities, it's going to have a massive impact and we need to give these tech companies a leg up.
So what does it mean for some of the largest tech companies? I mean, these are big names in the world of markets, companies like Cloudflare and CrowdStrike, all of these cybersecurity firms. What does this actually mean for them? I think, and I guess this is a debate, whether it's a bad thing or perhaps even a good thing, because I guess now you need cybersecurity more than ever.
Well, in the long term it means we're going to completely change how we manage software vulnerabilities because up until this point, at least in theory, you find out about a vulnerability, you work really hard, you patch it, maybe it'll take a few weeks, maybe it'll take a few months, and you hope that the bad guys don't manage to figure out the same thing and exploit it first. Right? So, we've had that lag and now I'm sure this is a complete emergency for all the major tech companies. And honestly, I'm concerned about every organization in the software supply chain because frankly your small businesses, your smaller suppliers also need access to this to patch their bugs as well. So, I'm sure they're scrambling. They're trying to figure out how on earth are they going to patch these thousands of vulnerabilities? How are they going to prioritize? They've got to be using AI to help counteract this. But AI can also introduce bugs. My guess is right now they're scrambling to work on the system to make it so they can patch things faster than we've ever been able to before. And we're going to need to be shifting to more continuous software updates instead of like, oh, you get an update once a month or something like that.
Do you think the government should have a role in this? I mean, some people have made the comparison of AI to things like nuclear bombs, and I mean, I'm not sure if that's the analogy to be making, but I think you can compare it to a weapon at least.
That's very insightful, Ed. And when I first heard about this, I thought it reminds me of the cyber disarmament movement where there was talk that like the CIA and the NSA were stockpiling vulnerabilities and that creates risk for everybody. And certainly AI with these capabilities creates risk for everybody as well. And so one of my questions is what other companies beyond Anthropic have access to these capabilities? They may not be the only ones that have invented an AI that is capable of doing this. And so where I see the government potentially being involved is maybe setting some standards for disclosure. If you have an AI that has certain capabilities, do we need to track that? Do we need to notify anybody? And how is this going to be coordinated?
Do you think that, I mean, some would say, or I think some people think or suspect that maybe it's all being overhyped, that part of this is sort of like a fear-mongering marketing tactic to say, look how powerful, look how scary our technology is, to kind of get investors excited about, wow, this is going to be really transformative. And I guess it's sort of tough to understand because I think that's a pretty compelling point. Like, it would make sense for Anthropic to say, look how crazy and bad this is, and get a lot of headlines and generate a lot of heat, and then people like you and I come on podcasts and we talk about it. But then at the same time it also seems very dangerous and frankly stupid to just assume that it's marketing and sort of cover our eyes and say, no, this isn't going to be a problem, because it could very well be a problem. I guess where do you stand on this? Like, to what extent is this hype and fear-mongering to get everyone kind of excited and also scared versus, no, this is very legitimate and we actually should be scared.
Yeah, I wish I thought it was hype. But I myself have been using Claude Code. I've been looking at the capabilities of Opus. Last year I actually did a research project with my colleague Matt Duran and we researched the dark web and looked at tools like WormGPT and FraudGPT that were already pretty darn good at finding vulnerabilities and writing exploits without all of the ethical guardrails that we have on legitimate tools. So absolutely, I think that there are some real capabilities here. I do think that the future will be different than any of us are predicting. The other thing to remember in security is that frankly cyber criminals have had the ability to get into most organizations for a long, long time. And it doesn't always mean there's some disaster that comes out of it. The same thing as it's possible to burn down a building. That doesn't mean all buildings will burn down. Even adversaries have limited resources. And so there's always going to be that balance. But yes, absolutely. This is a huge change to the cyber risk landscape. I'm sure Anthropic is capitalizing on that. They put out a press release. I also suspect that there are probably many other organizations and groups that have similar capabilities that are not putting out press releases, but we should absolutely be concerned.
Do you think that China, this is the kind of thing that China would be focused on, or Russia, or any other, perhaps Iran? I mean, to what extent is this like a real national security risk?
Well, I'm wondering if China already has capabilities approaching this. If not, I'm sure they're trying to get their hands on it. Same thing with Russia and other nation states and organized crime groups. I also thought about when Tesla was almost hacked. This was several years ago, and I don't know if you caught this news, but Russian agents tried to bribe a Tesla employee into installing malware on Tesla's network. And they offered this employee a million dollars. And amazingly, that employee turned around and reported this. But what if the bad guys offered any employee at one of these 40 tech companies a million bucks? You know, is somebody going to take them up on it and say, 'Oh yeah, you can come into my computer and access it.' So, we really need to be thinking about how are nation states going to get these capabilities if they don't already have them.
