OpenAI, Google, and Anthropic Call for Defense Against Rogue AI - Criminals Victims of Their Crimes

Eli the Computer Guy (0:00)

Oh lordy, these are the types of stories that just make me want to giggle a little bit. OpenAI, Anthropic, Google, and 100 other companies call for action to defend against rogue AI coming from TechCrunch. The AI companies come out, and with one voice, they say, you know, we should probably take cybersecurity a little serious now, don't you think? So I think this is just a funny, stupid story, because it's taken a problem that we have had for decades at this point, and then trying to pretend like with artificial intelligence now is when we should be taking things seriously. Right, so we've been having ransomware attacks. Ransomware attacks have taken down hospitals, ransomware attacks have taken down insurance companies, ransomware attacks have taken down schools and universities and gone out after government organizations and the whole nine yards. And we still haven't had much better cybersecurity policy, right? Information has been stolen for decades at this point in time, and we still don't have a real national cybersecurity program, right? This is one of the things I've been arguing about for a long time is that we need real national cybersecurity strategy. One of the funniest things is I swear, like every two years, one of these big organizations, one of these big institutions will come out, and they will give their security recommendations. And I'm a good guy. I'm a good geek. I'm like, oh wow, they're coming out with new security recommendations. Let me read this so I understand how to secure myself in the modern world. And I swear to crap, every time they do, it's like, yeah, make sure you have good passwords, make sure you have good patch management. Like they'll have like 10 things, and it's all like the crap that should have been done in 2005, literally, Five Eyes. I did a video about this probably four or five months ago. Five Eyes, so Five Eyes is the national security organization between the United States, Canada, England, Australia, and New Zealand, right? Supposed to be like, ooh, our intelligence organizations come together, right? Anyways, they came out with a cybersecurity memorandum a number of months ago, and I about lost by fucking shit reading that dumbass piece of thing. Because I swear to God in it, it was things like, update the software on your computer. In 2026, Five Eyes' recommendations for cybersecurity is patch fucking management. We dealt with patch management in 2005 And this is the problem, right? I have a lot of young people out there, those little 18-year-olds, 19-year-olds, right? Those snot-nosed little brats. They want to do something real in technology. And they're like, well, obviously, Eli, we need more cyber defenders. We need more cybersecurity. Look at all these attacks that are happening. I'm like, oh, dude, here's some Kleenex. Wipe that shit up. And two, two, if these organizations, if these companies, if the government gave two shits about cybersecurity, none of those attacks would have happened from the get-go. Again, when we talk about things like ransomware attacks, so many of these ransomware attacks, what you have to understand when you're dealing with security is that the database system, the database system has its own internal security, something called privileges. So the ability to add tables, drop tables, alter tables, insert records, remove records, whatever else.

There's a security system in the database. But what you also have to understand is that the database itself is a file. The database itself is a file. So here's the thing. When you hear about these ransomware attacks, see, you will hear this countless times of, there's been a ransomware attack, the database is no longer usable, but don't worry, none of your data has been exfiltrated. None of your data has left their systems. And that seems weird. It's like, well, okay, so the database system has its own security. So how are the hackers able to get the database system, encrypt the data in the database system itself? But they were able to do that, but they didn't have access to that? That seems weird. Well, the reason is, the reason is, is because of gross fucking negligence by the system's administrators. Because remember, the database file is a file, is a file, just like a JPEG file, or a movie file, or whatever else, right? So if you can come through the directory tree, right?

If a virus or a hacker is able to access that file through the directory tree with the appropriate system level privileges, then they can do something like run the software to encrypt the database to make the database unusable, right? And so, right, when you have a system administrator that logs into their computer as a global administrator and using gross negligence that should send them to jail for a year, they then check their e-mail. In their e-mail is a spear phishing attack. When the spear phishing attack is triggered, right, that virus or whatever else, remote access tool or whatever else is going on, that will go into the system, right, at the privilege level of the user. And if the user has escalated privileges on the overall domain, right, all the computers and the servers and the network, then that remote access tool or that virus will be able to do whatever it, basically whatever the fuck it wants in that entire system, right?