30 minSeason 3, episode 18

The Debate: Is the CIA Triad Truly Dead?

With Chris Romeo, Izar Tarandach, and Matt Coles

Threat modeling & security architecture

We’re debating an online article claiming that the CIA Triad (Confidentiality, Integrity, Availability) is a relic and needs to be updated for 21st-century threats. The discussion includes whether new properties like authenticity, accountability, and resilience should be incorporated into modern security models.

Episode title image: The Debate: Is the CIA Triad Truly Dead?

Listen to the conversation

About this episode

We’re debating an online article claiming that the CIA Triad (Confidentiality, Integrity, Availability) is a relic and needs to be updated for 21st-century threats. The discussion includes whether new properties like authenticity, accountability, and resilience should be incorporated into modern security models. And we delve into the use of analogies, system properties versus values, and the role of ethical considerations in cybersecurity. Listen along to our discussion on whether the foundational elements of security need a refresh.

The CIA Triad is Dead 




Follow AI Security Table:

➜ Home: https://securitytable.ai/

➜ X: https://x.com/AISecurityTable

➜ LinkedIn: https://www.linkedin.com/company/ai-security-table/

➜ YouTube: https://www.youtube.com/@AISecurityTable

Transcript · 30 min conversation

0:11Matt Coles Off a cliff, you mean?

0:14Chris Romeo Hey folks, welcome to the Security Table. And that was an odd place to start the conversation. When Matt said off a cliff, no, no, it was in reference to what we're going to discuss. So joined by Matt Kohls, he's our Taryn Das, I'm Chris Romeo. We are looking at an article today called The CIA Triad Is Dead: Stop Using a Cold War Relic to Fight 21st Century Threats.

0:41Matt Coles And of course we don't mean the 3-letter agency, we mean the 3-letter.

0:45Izar Tarandach Yeah, we're like, okay.

0:47Chris Romeo Yes. Question. What you would refer to as question 1 on the CISSP test. What is the CIA Triad?

0:54Izar Tarandach What is the CIA Triad?

0:56Chris Romeo But they've changed it now to, why does the CIA Triad exist? Confusion.

1:02Izar Tarandach Very philosophical.

1:05Matt Coles Indecision.

1:07Izar Tarandach Awesomeness.

1:08Matt Coles No, that's not it. Parole.

1:12Chris Romeo Parole.

1:12Matt Coles Snip. Parsley. Bird.

1:17Chris Romeo All right, so. Where do we go from here?

1:20Izar Tarandach What is the central proposition of this article?

1:26Matt Coles Uh, that a, that a fundamental property, the fundamental properties of security no longer exist or are outdated.

1:33Chris Romeo Outdated for modern cybersecurity, which—

1:36Izar Tarandach But first I have to go meta here for a second and understand one thing. When will people understand that they cannot say blah, blah, blah is dead and then write something saying, let's extend blah, blah, blah. Because if dead, you don't extend it. It's dead. You do something new.

1:56Matt Coles You have to think of the analogy. The tree in the forest dies, lays on the forest floor, and then from it springs new roots.

2:07Izar Tarandach We have to have a conversation about how you do an analogy.

2:09Chris Romeo I'm a tree farmer, not how trees work.

2:11Izar Tarandach Remind me. Remind me. Analogies don't work like that.

2:17Matt Coles Okay. The animal, the animal dies in the forest, lays there for a while.

2:21Chris Romeo Another animal, another animal comes out, explodes out of it. Is that what you're saying?

2:27Izar Tarandach Well, wait, isn't that how it works? All right.

2:31Chris Romeo So anyway, welcome to the birds and bees lessons of the Security Table.

2:37Izar Tarandach Hey, Cyberdykes.

2:39Chris Romeo No offense, but anyway, I just don't understand.

2:42Izar Tarandach I don't understand.

2:44Matt Coles The AI agent dies in the container and from it springs— springs the human.

2:51Izar Tarandach An AI agent dies in the forest. Does anybody hear a sound?

2:55Chris Romeo Does anyone hear it scream?

2:56Izar Tarandach Yeah. In space, nobody can hear a scream. Anyway, so let's go back to the CIA is not that dead. Mori Bond.

3:05Matt Coles All right. Does everyone understand what the CIA triad is?

3:10Izar Tarandach Confidentiality, integrity, availability.

3:12Matt Coles Yeah. Core fundamental properties, right? Yeah. Everything's built on it.

3:16Izar Tarandach Ever been. And it's one of those things like Stride. You can always like force things down to that acronym.

3:25Matt Coles Uh, let's not use that analogy because I think that there are some things you can't, but that's okay.

3:31Izar Tarandach Always for deferring values of always.

3:34Chris Romeo Matt, did you write this article? Are you saying that the things that don't line up with it are AI, cloud, ransomware, deepfakes, and supply chain attacks?

3:42Matt Coles No, that's not what I'm saying.

3:44Chris Romeo I thought you might have been the author of the article and just didn't realize it.

3:47Izar Tarandach No.

3:47Matt Coles Yeah. No. No. Uh, I was just disagreeing with his article that, well, let's see, privacy. Doesn't fit straight?

3:59Izar Tarandach No, no, I said always for changing values of 'always.' Always— it's "always" with an asterisk!

4:06Matt Coles Yeah, like yesterday it's fine but today not so much...

4:10Izar Tarandach Right?! But—but-but-but-but…but—the thing here: The article here is saying that really there isn't that—it's just let's extend this. So we have to do something about it.

4:22Chris Romeo Somehow. It needs to be refactored.

4:24Izar Tarandach Yeah. Is their argument. It needs to be refactored with authenticity, accountability, and resilience. That's part of the core. That's core. I'm just going to throw this one out there. Close your eyes and see the typical security team operating in the wild. Now, on top of what they do every day, day in, day out, And accountability and resilience. Notice that I left authenticity off. Accountability and resilience. But my only question is, seriously?

5:03Matt Coles We need to take a step back. There's security principles and there's business principles.

5:10Izar Tarandach And there's things that you have to get done on a day of work. So, yeah.

5:15Chris Romeo So, What is accountability in this context? Like, hold on. Accountability is like, Matt, did you finish your TPS reports?

5:24Matt Coles I think I'm going to have to have you come in on a Saturday. Yeah.

5:28Chris Romeo Like, how do you, how does a technical feature or a system have accountability?

5:38Izar Tarandach Clearly, clearly you can use tickets that book people back.

5:41Matt Coles And why is it a security, why is it a security problem?

5:45Chris Romeo Okay, so I mean, I get resilience. I get resilience. That's been a big topic. That's been a big topic for the last couple of years. Yeah, but that's a security general.

5:54Matt Coles But that's a quality of a system.

5:57Chris Romeo Oh, it's very challenging to reach, to say that you have resilience system.

6:02Matt Coles It's a goal, it's an objective. It is not a fundamental security property.

6:07Izar Tarandach Wait, resilience of what?

6:11Matt Coles When you have a system that is resilient, it is capable of sustaining attack and continuing to operate. Right. Yes. Which means it's capable of having its CI and A compromised and still be able to operate. Yeah.

6:28Izar Tarandach But when you say, when a system has the property of resilience, What is that system exactly? Are we talking, uh, product security? Are we talking, uh, operational security?

6:44Matt Coles Are we talking— well, it doesn't really matter. It doesn't really matter, does it? I mean, if you have a, if you have a, let's take ransomware. So they use ransomware in the article. So if we have a system, for instance, that is doing data storage, a resilient system would be able to survive an attack against its availability of data. So if you have a single database on a single file system, that file system gets attacked by ransomware, the system falls over. You've lost your data. It's unavailable. A resilient system would have a ready, ready hot, a hot backup available to switch over to that isn't affected by the ransomware to continue serving services, even though it's been compromised, originally been compromised by ransomware.

7:26Izar Tarandach What a bit cheeky. Just say that in our 3 people experience among the years, this resilience thing has always been something that surfaced in threat models and we never gave it a label because it was there, you know, hey, you got a single point of failure, you shouldn't have a single point of failure. Let's do something about the single point of failure.

7:47Matt Coles Yeah, I mean, when what can go wrong, if you look at that and go, well, what happens if your data gets hit by ransomware? Is a reasonable question to ask, right? And then a threat of, I have a single point of failure because I only have one database and no backups. Yeah, sure. That comes out in threat modeling.

8:06Izar Tarandach But we've been covering that with availability. We have been doing resilience and we've been calling it availability. So why do I need a new name?

8:14Matt Coles No, no, we've been calling— So it's resilience. We've been calling availability the problem we want to avoid. We want to avoid a lack of availability. By building resilience.

8:24Chris Romeo So they're complementary to each other.

8:26Matt Coles They're complementary, but that one's of, one's of them's about the principle.

8:29Izar Tarandach That's not what a complement, that's not how a complement works. Uh, if you, if you're trying to solve the lack of availability with resiliency, it means that availability, having availability is having resilience because you just added the lack of.

8:44Matt Coles So you inverted that. Maintaining availability is resilience. But availability is still the fundamental property that you want to preserve. Right.

8:52Chris Romeo So we're back to the original being better than the copy.

8:54Izar Tarandach So why, why, why does the article need a new name for it?

8:58Chris Romeo Well, cause that's how you get people to read your article.

9:00Izar Tarandach Oh, that's why. Sorry. Now I got it. That was like— Now I got it.

9:06Chris Romeo I got, I drank some truth serum this morning. No, no, no, no, no. Apparently it just kicked in. I'm like, whoa.

9:11Izar Tarandach No, you just, you just revealed, you just revealed the fourth value. Marketing.

9:17Chris Romeo Let me drink a little more truth serum.

9:21Izar Tarandach It's the fourth value. All right.

9:25Chris Romeo So Matt, they did mention privacy though. That was your argument about the CIA. Stride doesn't cover everything. No. They're complementary.

9:32Matt Coles No, no. Stride doesn't cover everything.

9:35Chris Romeo Oh, no, you, oh, you said Stride? I said Stride.

9:37Izar Tarandach Yeah, because I, I, I added that much like Stride. Okay.

9:41Chris Romeo I was, you know, reading things on the internet. Complementary privacy provenance compliance. I like that word provenance. It sounds very like suave. Yeah.

9:53Matt Coles It's where data comes from. It's an integrity problem.

9:57Izar Tarandach Isn't integrity a technical difficulty?

10:00Matt Coles Well, yes, but authenticity is an integrity problem.

10:04Izar Tarandach Right. But no, he is just like pining stuff on We know how fundamental principles work, right?

10:10Matt Coles You could derive back to them. Right. They're usually at the root there. If you factor them, if you get the factors of something, you can—

10:21Izar Tarandach But my point is, we also know how keep it simple works.

10:25Matt Coles Wait, hold on. Wait, wait, time out. Time out. You're talking to a math major and a computer science major.

10:31Chris Romeo Sorry. And— Sorry that you guys had that Bad experience in college.

10:36Izar Tarandach Dude, I just squeaked by. I don't count.

10:39Matt Coles And you know that computer science is like, all goes back to the math, right? All of the computer science stuff is all like math-based.

10:50Izar Tarandach No, it goes down to turtles. It's turtles all the way down.

10:54Chris Romeo Like Ninja Turtles or regular turtles? If it was Ninja Turtles, that would've been cool.

10:59Matt Coles Regular turtles.

11:00Chris Romeo Okay.

11:01Matt Coles But they didn't have Ninja Turtles in the early days of computing, so, you know, according to the article.

11:08Izar Tarandach Or did they, and they were very warheathen. That's right.

11:11Chris Romeo Okay, so Matt, you got privacy listed under complementary. That was your argument against, you know, how Stride didn't cover everything. Is that a good thing, that privacy is listed under complementary here? Do we, is there a shining light amidst this article that we're just ripping on.

11:28Matt Coles I'm gonna, I'm gonna say something bad and I'm gonna say something and it's probably gonna be bad, but privacy is not security.

11:36Izar Tarandach Right.

11:37Matt Coles But no, no, no, no, but yeah, no, no, no, no, no, no.

11:41Izar Tarandach You're right. You're right. You're totally right. You're totally right. I agree with your conclusion.

11:43Matt Coles And so fundamental security properties versus fundamental privacy properties are related, but not the same.

11:50Chris Romeo So somebody didn't do their homework.

11:53Izar Tarandach But wait, if you're going, I'm going to throw a different one in here. If you're going to extend the CIA triad and make it, what's a foreign, I don't know, 50-ender?

12:07Matt Coles Quad? Quad?

12:10Izar Tarandach I don't know, I'm sure there is.

12:11Matt Coles Quadratics?

12:12Izar Tarandach Anyway. Quadratics? If you want to add things to the CIA triad, I would vote flipping for privacy and safety. Those are my two goals. But hold on. Hold. Oh, hold on. And what? Wait, wait, wait. Stop the presses. Why, why, why, why are we adding? Why are we adding things? Let's add ethics too. Yeah. Yeah. Let's put ethics in there too.

12:34Chris Romeo Don't we have that?

12:36Izar Tarandach No. Under contextualism. Yeah. Yeah. Yeah. Yeah. Let's break the whole thing. Yeah. All right. All right. Back up. Matt's head is—

12:47Chris Romeo you ever seen the movie Scanners? Matt's head is about to explode.

12:51Matt Coles He's like, I feel, I feel like an episode. I feel like the— what is it? Um, it's not Terminator. It's, uh, he goes to the red planet. Schwarzenegger. Um, my mission here is Running Man. No, not Running Man. The one he goes to the red planet, but he's really in a dream. Oh, but total recall, total recall, feature.

13:12Izar Tarandach Spoiler alert. He was dreaming the whole time. I never finished the movie.

13:19Matt Coles Oh. Did you see that?

13:20Izar Tarandach Did you see that they're remaking Running Man?

13:24Matt Coles I hope it's better than the remake of Blade Runner. Anyway, it's not called— so, so, all right. So why, why am, why am I about to have an aneurysm this morning? You know that I've talked about the three-legged stool for a long time, right? This three-legged stool of system properties, security, privacy, and safety being the three parts of the stool. Right, that was more than three.

13:53Chris Romeo That was more than three pieces.

13:55Matt Coles No. Security, safety, and privacy. I can count the three, right? And then— I can't. And then within that, Security has its own properties. There happen to be 3 or more as we, as we're discussing here, right? And so, privacy up, security has confidentiality, integrity, and availability as its core. Privacy has other properties and safety has yet other properties that are at its fundamental core, right? And so, the, the 3-legged stool of, it's kind of like the Iron Triangle, but it's really just the 3 legs of the stool that holds up a system: security, privacy, and safety. Security is about protecting the system and keeping it functional. Privacy is maintaining the integrity and confidentiality of data for a user, or about a user, I should say. And safety is making sure that the system causes no harm to individuals. They have, they are all interrelated, they rely on each other, but they do not are not the same.

14:59Izar Tarandach I am going to offer you, uh, what was it? Complementary view. What if, what if we got to the conclusion, while I agree completely with what you just described, my problem is with the word property. Okay. Because we keep saying that systems have to exhibit the property of. And that brings us to how do we prove the property being there? How do we prove the system exists? Is it enough for the system to show it's 90% of its systeminess to have the property, right?

15:39Matt Coles Well, the CIA tried to try it on principles.

15:46Izar Tarandach They are also called properties in NIST. Okay. So the step that, the jump that I wanted to make was exactly to take what you just described as the tripod, which I think it's a great analogy. One like some tree dying in the forest.

16:07Matt Coles The tripod, there you go. All right, I was thinking of the three-legged stool, the tripod makes it a lot easier to set up.

16:12Izar Tarandach No, no, no, sorry, sorry. I meant, I meant, no, no, no, I meant the three-legged stool, 'cause you can actually, put things on top of it and build more.

16:19Matt Coles But, but tripod works too, and it's a lot easier to say.

16:22Izar Tarandach Yeah, but it goes to a point. It's, it's not— you put, you put things on top of it, you put— no, no, the, the, the stool is better.

16:30Chris Romeo The stool— maybe you're not using tripods right.

16:33Izar Tarandach Yeah, plus, you know, stool, we can take it a lot of places anyway. But my, my point is, what if instead of properties we looked at things as values, right? So exhibiting a property to, in my head, could be totally wrong, and I probably am, goes straight into implementation. But if you bring things back to values, when I'm designing something, I'm keeping the value of CIA, safety, privacy, and ethics. That's the one that I wanted to add to my to my soup, my letter soup, because we know that there's a lot of system designing going out there that is taking into consideration the best interest of its users or whatnot. And I hear Brooke talking a lot about the ethics of stuff and it really opens my ear. He has something very strong there. So my whole point is that by bringing, bringing things into values rather than properties, we change the dialogue a bit. And rather than aspiring to some property being exhibited, we can now, we can now make value a central part of the design of the system and force it to go into implementation. It could be just, uh, you know, word, wordplay. In my head, it rings true.

18:09Matt Coles Yeah, I don't disagree with you. I'm just looking up for, I had a running conversation with something about privacy and security basis. There's legal bases for things and ethics is part of that, has to be part of that, right? And so there's duty of care. I was just looking that up. Yes. So duty of care of system design, of a system, and this duty of care obviously predates computing systems, but it applies to them, right? So at some point, there was an expectation that a system designer would exhibit some sort of duty of care with respect to the system. And its capabilities and how it operates. And beyond that, I don't know much about that. If anyone's a legal scholar, they can feel free to chime in. But yeah, there's something that underpins all of this that may or may not include ethics.

19:17Chris Romeo So basically, duty of care means if you, that you have to care about what you're building?

19:25Izar Tarandach Isn't duty of care first do no harm?

19:30Matt Coles Well, I guess if you're building a weapon system, you would violate that, but you could still have duty of care.

19:37Izar Tarandach Okay, first do no harm to our guys? Yes.

19:42Matt Coles I think it's don't screw over your customer.

19:45Izar Tarandach Yeah, yeah, that's where I was going with that. Lord knows we don't see that a lot nowadays.

19:54Chris Romeo I'm asking a legal scholar that I know. Okay, a responsibility to act with reasonable care to avoid causing harm to others.

20:09Izar Tarandach Yeah.

20:10Chris Romeo So this is, this is what we should have been doing for the last 25 years building products.

20:16Izar Tarandach This is what I want to believe I was doing for that amount of time. Unfortunately, uh, it's not shared.

20:24Matt Coles Yeah. Okay, so here's an alternative, an impli— um, a way of saying that, Chris. Duty of care is a legal obligation to take reasonable measures to protect others from foreseeable harm related to their modeling. Well, yes. Because you want to prevent harm. Now, harm, if you look at harms, like, especially if you look at NIST's Risk Management Framework, which is a, which is outdated, I know they're working on that.

20:51Chris Romeo I'm gonna be honest, I try not to.

20:53Izar Tarandach Yeah, you shouldn't. Yeah, you don't.

20:56Matt Coles But the harms are important. They're not personal, they're not, they're not always personal harm, right? Yeah. It could be harm to the business, it could be harm to the mission, it could be harm to health. Sure, business continuity, disaster recovery. Yeah. All those other core things that we want to maintain when we build systems.

21:15Izar Tarandach Yeah. So, Chris, I would go and say that no, threat modeling is not duty of care. It's one way to carry it out. But how many times have we seen that we speak to people and say, you know what? This is fine from a security standpoint, but I don't think that we should be doing it. No, no, no, we need to do it. The business requires it. So, Where does that fall into the duty of care thing?

21:40Matt Coles Uh, your customer comes first, I think, right? Uh, so in other words, harm, harm to the business, you have a choice to make. And I guess that's your ethics discussion, right? Exactly. You know, if they, if they say, build me something that will kill everyone who uses this, you have a choice to make.

21:59Izar Tarandach My name is Kevorkian.

22:02Matt Coles Well, I mean, unless you're building an assisted suicide facility. Yeah. Right? I mean, that, this, it goes to the purpose of the system that you're building.

22:13Izar Tarandach But that's where the value thing comes to me. It's 'cause the property is something that's absolute. You have to, this thing has to be, has to keep confidentiality. But now if I translate that from, property into a value, this thing will keep confidentiality if confidentiality is something that's important for me, my customers, and my system, and my business, and all that stuff. So things become a bit more fluid, I think. But on the other hand, it brings them back to design and one step away from implementation.

22:57Matt Coles Was that it? I think so.

23:01Izar Tarandach I hope so. Been talking about it for a while.

23:04Matt Coles We'll just give you a little bit. So are you saying that if you say it often enough, it must be true?

23:09Izar Tarandach It works for everybody else.

23:10Matt Coles That's, that's, isn't that, isn't that the fundamental core of marketing?

23:16Izar Tarandach Oh my God. If you bring that to the fundamental core, then we are going to go into some dark, dark waters.

23:23Chris Romeo Wait, so Isar is now a marketer?

23:27Izar Tarandach Wow. As Mel Brooks would put in History of the World Part One, oh, a bullshitter. Have you bullshitted this way?

23:38Matt Coles So we suddenly went from PG to PG-13, is that what you're saying? Yeah. Oh, what? We just went from PG to PG-13.

23:45Izar Tarandach Oh God. That's, that's— We've been dropping those for a while.

23:50Chris Romeo It's about time. About time. Okay. So if we kind of wrap this—

23:56Izar Tarandach Oh, go ahead. We're wrapping? No, no. Just back to the article. I mean, the thing that killed me in the article after reading was that basically it says that security teams are overwhelmed by many frameworks. ISO, NIST, GDPR, AI Act, etc. Let's add one more model and call it a meta-framework.

24:20Matt Coles But how does that—

24:23Izar Tarandach Yeah.

24:23Matt Coles That's different than changing the fundamental properties of computer security. Yes. And how did—

24:34Izar Tarandach Oh, look, they're drowning. Let's drop some more water on their heads.

24:40Matt Coles Wait, wait. How? Do you wanna, do you wanna, do you wanna just drop them a buoy that happens to be a, a concrete brick? We can add gravity to the mix.

24:49Izar Tarandach I mean, no, if you're going out to solve, if you're going out to solve a problem, people, there's a very, a very simple method, way of solving problems. You define the problem, you propose a solution, you show that the solution works, ta-da, problem solved. But changing the— It'll create more problems. Changing the problem.

25:06Matt Coles Well, whether you're, whether you're not, The thing is, well, then it's a misleading article because a framework to maintain CI&A can be done without changing CI&A. Yes. True. We can add a meta-model, a meta-framework for all the frameworks that exist when all those frameworks are geared around maintaining confidentiality, integrity, and availability.

25:36Izar Tarandach Let me ask a rhetorical question here. Do you guys remember any—

25:42Chris Romeo We don't need it. Rhetorical questions don't need an answer.

25:45Izar Tarandach Any meta framework that ever helped you?

25:51Matt Coles I have no answer for you.

25:56Chris Romeo I have no idea what he's talking about. At least he said rhetorical question, so I don't have to provide an answer that doesn't even address what he was talking about.

26:06Izar Tarandach No, seriously, have you ever heard of a metaframework that helped you do better the other frameworks, the meta-sizes, whatever?

26:17Chris Romeo No, can't say that I have.

26:21Matt Coles Oh, um, alrighty.

26:24Chris Romeo So. What do we do with this?

26:26Izar Tarandach Now you get to work. Sigh.

26:31Chris Romeo I mean, is this just dribble? Is this just dribble that we're reading? Do we waste like 30 minutes talking about something that's just dribble?

26:40Matt Coles And so, uh, Brandon, English, look, I, you know, there's a, there's something in here, but they, but the message is wrong. I, in my opinion, at least, right? But. They, this is the throw the baby out with the bathwater problem, right? It's, we, something is broken, but we don't know what, so we'll take it down to the core and throw it all away and start fresh.

27:03Izar Tarandach But that's the thing, it's not throwing everything away. Putting a bandaid on the thing and building a pyramid on top of it.

27:10Chris Romeo Yeah, but under layer 1 core, he managed to work in zero trust and SBOM. Both. That's a lot of acronyms.

27:18Matt Coles Oh wait, wait, that's another time. How many times does we talk about it here? Wait on the core, on the core what WTF SBOM.

27:33Chris Romeo Well he said authenticity is the engine for the authenticity yeah and then under the accountability to ensure accountability security practices must extend into the software supply chain enforced by practices like SBOMs.

27:45Matt Coles SBOM doesn't do that.

27:47Izar Tarandach Oh, Matt, come on. Wait, wait, wait, wait. SBOMs don't do everything.

27:51Chris Romeo You can't even say it with a straight face. I tried. I was like, I'm going to be, I'm going to be serious and try to like argue the other side. I couldn't do it after 3 seconds. SBOMs don't do everything.

28:02Matt Coles Yeah. SBOM is not the solution to everything.

28:05Izar Tarandach And, and by the way, let's talk about hardware someday.

28:09Matt Coles Um, So anyway, well, it's a— S-Bomb is not the solution to everything. How? Hold on.

28:17Chris Romeo How's it caught another, another under layer 3? The number 2, a breach like Equifax in 2017. Dude, you can't use Equifax as an example anymore. It's the year 2025. It was 8 years ago. Like, there's, oh man, come on. Okay.

28:38Matt Coles A technical failure, but a contextual collapse. Yeah, that's great. The technical failure is what we care about. And it was a failure of confidentiality and/or availability. Like, come on. You go back to the core fundamentals. What? All right. Anyway, sorry. I vote we throw this thing out. Off my soapbox, sitting on my stool.

29:01Chris Romeo Yeah, this is, this, there's nothing here. There's no nugget here. There's, there's— No, this one is gone. There's just a, an attempt to take something and twist it to get people to read an article.

29:12Matt Coles And maybe for next time, we can talk about the companion article that goes with this that says, well, fundamental security properties are broken because AI agents exist. And we'll leave that as a teaser, maybe for next time.

29:23Chris Romeo And I'll just, I'll have to, I'll have to look up what that means. All right. Well, folks, thanks for joining the Security Table. Live long and prosper.

Transcript supplied by AssemblyAI. It may contain transcription errors.

Originally published as The Security Table. Part of the AI Security Table archive.