32 min

Innovations in Threat Modeling?

With Chris Romeo, Izar Tarandach, and Matt Coles

Threat modeling & security architecture

In this episode of The Security Table, hosts Chris Romeo, Izar Tarandach, and Matt Coles dive into the evolving concept of threat models, stepping beyond traditional boundaries. They explore 'Rethinking Threat Models for the Modern Age,' an article by author Evan Oslick.

Episode title image: Innovations in Threat Modeling?

Listen to the conversation

About this episode

In this episode of The Security Table, hosts Chris Romeo, Izar Tarandach, and Matt Coles dive into the evolving concept of threat models, stepping beyond traditional boundaries. They explore 'Rethinking Threat Models for the Modern Age,' an article by author Evan Oslick. Focusing on user behavior, alert fatigue, and the role of psychological acceptability, they debate whether broader human factors should integrate into threat modeling. 



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 · 32 min conversation

0:05Izar Tarandach Good morning. Oh, that's a—

0:12Chris Romeo Welcome folks to another episode of the Security Table where we focus on intellectual dialogue. We guarantee 5% of the time. My name is Chris Romeo and I'm joined by Izar Terrindosh and Matt Coles. Distinguished members of the Security Table. And today we're going to talk about a topic that's kind of foreign to us. We don't spend a lot of time thinking about it or, or considering it. It's this idea of threat models. So we should probably Google what one is real quick. It's—

0:45Izar Tarandach but wait, wait, this is different. We usually think about threat models. This one is rethinking threat models.

0:50Chris Romeo Okay. Rethinking. Okay. So rethinking, reworking. What's your threat model? Relapsing.

0:55Izar Tarandach Can you just follow me now?

0:57Chris Romeo All right, well, the article we're discussing is Rethinking Threat Models for the Modern Age. And, uh, yeah, let's just, let's just kind of start unpacking it. I know there was some, some things that caught my attention in this article that, uh, I thought were worth discussing. But, um, if we kind of make our way through this and consider what, uh, Evan, the author, talking about here. Um, you know, he's, he's talking about this shift in communication habits as a way to set up the conversation or the way he's going here. And just makes me think like we're all so inundated with so much noise. Like, do you guys even answer your phone anymore?

1:39Izar Tarandach What phone?

1:41Chris Romeo You don't even have a phone? Wow.

1:42Izar Tarandach You like next level, man.

1:45Chris Romeo Talk about baller status. No phone at all.

1:48Matt Coles I like my digital assistant answering the phone first. Wait, no, no, no.

1:50Izar Tarandach I, I, I do, but I have that, uh, that switch on my phone between silent, uh, uh, vibrate and full.

1:59Chris Romeo Yeah.

1:59Izar Tarandach And I don't think it ever went past silent.

2:02Chris Romeo Yeah, I do the same thing.

2:04Izar Tarandach If I'm not looking at it, I'm sorry, I'm not picking it up.

2:07Chris Romeo Yeah. I mean, and, and so in the article Evan's talking about alert fatigue, the, the challenges of all the noise that we are experiencing here. Um, and you know, I think it's true from both a voice call, text message. We live in this world where we're just bombarded with messages and some of us just turn them off completely and ignore them.

2:30Matt Coles So I think you're bringing— I think you're, sorry, I think, I think you're, you're a little late. You're in, our audience members are probably lost at this point about what we're talking about. So let's start with the definition. Let's start with the definition cuz we like to start with definitions and then, yeah, but I would, my only response would be, so what else is new? Big threat.

2:48Chris Romeo Sorry, sorry, sorry. That was a joke. It's comedy.

2:52Izar Tarandach So, and if it only was just the audio.

2:57Matt Coles So, so there's a, there's a fundamental part of this article that, that we need, we need to, we need to start with this definition. So what, what are, are threats? Uh, and, and so threats. Obviously, obviously threats are things that can cause bad things to happen in your, in your system, right? It's privacy or security or safety threats, uh, things that, you know, things that people will do or that, that will occur to your system that will cause some negative outcome. And, and, and understanding that and then looking at the article, right, where he talks about, well, uh, building an application without understanding How people communicate, does that represent threat? Right. Go ahead, Isar. Yeah.

3:46Izar Tarandach Yeah. No, I, I'm just thinking that it's, it's, let's even like use that definition and, and just reframe the thing here. The, the way that he got to the, uh, alert fatigue was by suggesting, uh, hypothetical application that in case of emergency calls a set of people that you predefined. And the link to roulette fatigue is people are not picking up their telephones if it comes from a number that they don't recognize or something like that. I know I don't, and many cases I don't pick it up if I recognize the number as well, but that's a different thing.

4:20Matt Coles But, um, I know.

4:22Izar Tarandach So he, he's saying that that fact that people don't pick up the phone if they don't recognize the number in the context of that hypothetical that can be seen as a threat and should be part of the threat model, right? And when I first, first read the, uh, the, the, the article, first of all, I, I really like the approach because it is looking at things that we may overlook on any given day. So I, I see a lot of value in, in, in just that call out. But in this specific example, the thing that, that stood out to me was, is this really a threat? Is this a failure of design? Is this something that I would ask as a threat modeler or that a VC would ask me when I propose the hypothetical, uh, uh, the hypothetical arc?

5:17Matt Coles I would think that this would be something, I don't know about the VC stuff, but wouldn't this be something that your product manager who's defining requirements or whoever's writing your business analyst who's writing use cases might try to put in Gene use case and have the organization go, are you sure about that?

5:35Izar Tarandach That, that, that's the thing. The, are you sure about that? For me, in this specific case, in this specific example, it's one of those of, uh, it's working as intended. The, the app does what the app is supposed to do. What the app can't do is get a hand out of the phone with a hammer and just bang somebody in the head until they answer.

5:57Matt Coles Right.

5:58Izar Tarandach That's the limit of the system, right?

6:00Matt Coles But would this be a fundamental fault in the way that use cases are defined? Meaning, so, uh, just to catch you, the, uh, the listener, anyone that's listening to this, uh, right, we're talking about an application that calls, that does a phone dial, basically a phone dialer for messages. And what Isn't this a fault when you do use case development? Hey, we're going to ring, we're going to call people on their phone and expect a behavior. And when that behavior doesn't get realized, now we don't know when this application was specified. Was this like 3, actually I don't remember the article, if it was a recent application or something that's retrofitted from before, like did this work in the '50s and now it doesn't work today. Um, but, uh, no, this sounds like a, it sounds like a use case problem, use case definition problem, requirement problem, not a threat, unless we're bringing it back to, um, even beyond before what we traditionally consider for threat modeling, which is more of a design review and looking at threats from a, from a security needs perspective.

7:09Chris Romeo I'm gonna, I'm gonna look at this from the other side. I'm gonna say this isn't a product problem. This is in fact a threat in the way that the author described it, because at the end of the day, the threat is a user does not receive a notification that they, that crucial notification about some emergency situation. And so what's the danger that could come from that? Well, you know, there could be any amount of danger.

7:35Izar Tarandach Wait, there, there, there's a difference here between not receiving the alert and not addressing the alert. Not picking up the phone when it's ringing, it means that you're not addressing the alert, but the alert is getting to you. The alert is not getting lost in some forest or something. You are getting it. So that there is a big difference in terms of tripped, because again, the app is working as specified. What's not working as specified is the human that's supposed to react. To the app.

8:04Chris Romeo Yeah, and I think his point is that can be, we, we, the threat landscape, I hate to use that term, but can be expanded as a result of application behaviors and user behaviors. Like we can go beyond, we, we've all said this and this is a good example of it. Threat modeling can be bigger than just looking at a web app and a feature within a web app. You can expand threat modeling. As far and wide as you want to. And this is just an example of him saying, hey, there are, you can go beyond the classic security and privacy Stride and Linden challenges to feature.

8:42Matt Coles Well, remember original threat modeling, original threat modeling before Stride did, right? That's, we're talking about things that eventually made its way into, say, the, the NIST risk management framework or the harms and hazards analysis that that, uh, medical devices go through, right? This is the, there's a, there's a portion of threat modeling that's supposed to happen before you even write any, before you write requirements and you're supposed to drive requirements from it and then go into design and then do a design risk analysis, right? So in, again, looking at traditional waterfall model, you start early and even I would argue that this probably does, can happen even in agile models because you still have to have a concept and make an idea and then have a plan. And so when you do that, if you do yet early threat modeling where you're looking at deriving, you know, security requirements and looking at harms and hazards and other things, um, then this may, I guess maybe this is a call to go back to that little bit more structured approach and not the winged approach that seems to have taken hold.

9:49Chris Romeo So you're advocating for bringing waterfall back?

9:51Matt Coles What I just heard? No, no, no, no, no, no, no.

9:55Izar Tarandach I, I like the idea of going back to things that are more formalized, but to address a different bit, which is to difference the complexity of the stuff that we are creating nowadays. I'm going to push back on the whole idea that this belongs in a track model simply because to me, this specific question in this specific scenario here is the question that happens when two bros are drinking their beers and thinking about their next startup and drawing on the napkins. And one goes, bro, but, but what, what, what, what if we called people every time that there was a text message? And the other one was like, but what if they don't pick up? Oh yeah, you're right.

10:37Matt Coles Yeah.

10:37Izar Tarandach Next law. Yeah.

10:39Chris Romeo But that, that, that, You're drawing this into like a startup, a product startup-focused thing. And I think, I don't think that's what the author's intending with this.

10:50Matt Coles Okay.

10:50Izar Tarandach And if you don't, what's the next step? The app is going to keep ringing until somebody picks it up.

10:57Matt Coles Yeah.

10:57Izar Tarandach That's going to introduce, that's going to introduce into the Olympics the next category, which is tossing cell phones as far as you can.

11:05Matt Coles Yeah. There's a technique to it, you know, you have to, you have to go sideways.

11:10Izar Tarandach Yeah, there's a whole Frisbee approach.

11:13Chris Romeo Don't actually throw your phone from the back. No phones were, were allowed.

11:17Matt Coles The hammer throwing, like the hammer throw, the hammer throw.

11:19Chris Romeo I just wish that Matt would have slipped and this phone would have gone flying across the room and hit the wall.

11:26Matt Coles Dang. Oh God.

11:27Izar Tarandach Oh, that's nothing. Imagine if you have a problem like the, the French bull, Walter, when it's throwing your phone.

11:31Chris Romeo Okay. All right. This is a family show.

11:33Matt Coles All right.

11:34Chris Romeo This is E for everyone here.

11:37Matt Coles Come on.

11:38Chris Romeo Yes. So I don't know. I disagree with where you are, where you're coming at this, Yezar. I think Evan's point is that there are threats beyond, like I said, there are threats beyond the things that we as security people put our hat on. Let me give you a hypothetical. Okay. You're doing a threat model with a new development team. You never worked with them before. And, uh, in the brainstorming session, a developer comes up with this as a threat. How are you going to handle it in that situation? Are you going to squash their future threat modeling dreams?

12:15Izar Tarandach With this specific thing?

12:17Chris Romeo Yeah, let's say they, let's say they came up with this. They said, hey, you know, in our app, what if, you know, we do these notifications for critical functions? What if, you know, somebody didn't answer and then the attacker was able to do something else or this or that? Like, how would you respond to that person?

12:31Izar Tarandach Oh, wait, wait, wait, wait, wait, you just added something very important there.

12:35Matt Coles An attacker.

12:36Izar Tarandach And the attacker would be able to do blah, blah, blah, blah, blah.

12:39Matt Coles Okay.

12:40Izar Tarandach Right now, the hypothetical on the table, the hippopotamus on the table, the only impact from there is the alert got missed. Right? So yeah, sure. Next time that I look at my phone, I'm going to look, there was a call missed or something. And things will happen. But if you have to, to threat model this from the point of view of, okay, yeah, sure, the threat, because now an attacker can do something with this, then you start looking at things like, okay, so whenever you register for the app, you have to add the numbers that it used to call you to your contact list so that when you look at it, it's not going to be an unknown number. It's something that you're going to see, oh, this is important. Even though we know that those things can be spoofed as well. So yay, one more thing from the threat model. But, uh, the moment you start putting a real impact in there and going beyond, oh, the app is not working as intended. No, the app is working as intended. Nothing changed in the ecosystem of the app. It's just that the person who was supposed to answer the phone was asleep or eating.

13:46Chris Romeo Well, yeah. And, and so you, so you're, you're concluding that that can, that's not a threat. That's a product deficiency.

13:54Izar Tarandach I'm thinking that as presented, it's not a threat. Now, the question, big question here is, even not being a threat, does that have a, a place in the process of threat modeling? And I think that Matt put it usually with his usual, uh, uh, great way of, hey, if this is a way of expanding and formalizing beyond what we do every day. And while I agree with that wholeheartedly, my biggest fear is that whenever you take a tool and you start using it for things that it wasn't intended or planned to use for, these things tend to start collapsing on themselves due to feature creep. So we already have—

14:45Chris Romeo By tool, make sure I understand, when you're talking, when you're saying use a tool, you're saying threat modeling tool that you're describing.

14:51Matt Coles Yeah.

14:52Izar Tarandach So if today we already have a huge challenge creating frameworks and ways for people to do the, what we seem to agree is the real hard part, which is the threat elicitation. Now, if I'm going to add other things that are more like business design process failures or whatever. I'm going to put that on top of it. Now, who am I going to ask this for? Or is the article actually saying, hey, we should have a different category of threat model and ask, as you said, Chris, product managers to do their higher-level threat model that's going to identify other things different from the threats that we are fighting every day for people to actually look at. And this should—

15:44Matt Coles no, I'm sorry, continue your thought, but I do have a response.

15:47Izar Tarandach No, no, I'm just thinking this should be like a similar but separate process.

15:53Matt Coles So I'm going to go back to the hazard analysis discussion just for a moment because I think— so if this— so right now there's no— the article doesn't— I don't think the article talks about the use case here so much, right? It's a, it's a thing that when a message gets generated, it generates a phone call. If that was a, if that was a message to a doctor in a medical setting, then your failure, that failure to reach the, the doctor because you're using a faulty communication medium. Is, is a hazard, right? It's a hazard that can result in a harm, right? You're relying on a functionality that, that is not effective and it results in, say, a patient, you know, dying. So there's a fine line, I think, here to, to worry about. It is a threat to the system, but it's a threat to the system at a broader level. And actually, maybe this is something that we should talk about. So the application is functioning as designed, ecosystem in which it operates and its fundamental reason for existing, right? So one system needs to send a message and make, or get, receive a message instead of fun— great. No threat there. The system operates in an environment that requires a delivery of a of that message because it's time-sensitive and critical nature. Either it's an incident that's happening or it's a— something's on fire or there's a medical issue, some other, other problem. Uh, and, and that app may function as designed, but it doesn't function within the context of that broader scope. So now you have a threat against the ecosystem, not just the system component, which is So maybe, maybe what really should have been highlighted in the article is the scope at which you do threat modeling needs to expand.

18:00Izar Tarandach But Matt, okay. So we get to almost a multi-party situation here where I could sit with a development team and have a 4-hour discussion about, because these messages must go through. Should we use a carrier that works only line of sight, or should we go satellite so that the message always gets through? Doesn't matter where the person is. And then at the end of the fourth message, uh, fourth hour discussion, somebody looks up, looks to the camera and goes, but what if they don't pick it up?

18:34Matt Coles All, all I'm saying is, all I'm saying is it's reasonable to have the conversation about, well, how to, how would this system stop? How would the system not meet its eventual goals? Broader—

18:46Izar Tarandach but, but that's the thing.

18:47Matt Coles It, it, it is, it is, you may not have a solution for it.

18:51Izar Tarandach Look, I, I, I, I think that there are two disciplines that actually, uh, focus on listing the ways that people fail and do things that it would be very reasonable to want them to do it. Psychiatry and psychology. I mean, we are fallible machines. The moment you, you think us as, us as, as a target of, of the function and put as part of the system and say, now try to model this. Today I, I wish I was being paid by the hour cuz millions are coming.

19:21Chris Romeo So let me, let me take something else from this article and try to bring this up to a 10,000-foot view and, and, and take some of what Matt's talking about here and some of what the article says and see if this makes sense as a potential strategy. So if we look at his third point under the expanding scope of threat models, collaboration with behavioral experts, um, involving experts who understand user behavior and communication trends can provide valuable insights to help shape more robust threat models. So we're talking about user experience, user interface designers. We're talking about bringing those folks. So what if we had another discipline of threat modeling that was, as I'm going to use Evan's words here, human factor threat modeling? And human factors threat modeling was focused on user experience, teaching user experience people how to sniff out security and privacy problems, as well as usability problems, because that's where they live is in the usability space. I think that's ultimately where I, I think that's the conclusion that I'm drawing from what Evan's saying here. I think we got really hung up on that one example. That's just one example of thousands though, of potential issues that could exist within an application system as a result of usability that would have a security and privacy tie-in.

20:38Izar Tarandach But again, that, that, that's, that one example is so representative of the, the last mile problem in computer systems. Like, we've had t-shirts saying you can't fix stupid forever now.

20:49Chris Romeo Which is a terrible—

20:51Matt Coles or, or, or PEMCAC, is it PEMCAC, uh—

20:55Izar Tarandach Problem Between Keyboard and Chair.

20:57Matt Coles Right.

20:57Izar Tarandach Or, uh, every time that you think that you created something foolproof, the universe comes and gives you a better class of fools. So the joke has been there for like hundreds of years, right? Now, whenever you come and you say, I don't want to say behavioral threat model or user experience threat model. I mean, isn't that the whole base of the whole A/B thing? We joked the other week about the shades of blue.

21:26Matt Coles Right.

21:27Izar Tarandach So you can actually apply, again, the knowledge of how humans think and, and work and, and do all the counterintuitive things that we do. And people do apply that to interfaces and to the ways that we relate to a number of different systems. I mean, look at any cockpit design and you're going to get a thousand different explanations of why that dial turns to the left and not to the right. Because people do take these things into consideration. But my problem here with this, what I think is a very representative hypothetical example, is that it's bringing into the maker of the system the responsibility of the user of the system. And at some point you just have to, you know, give up and say, this is how you use the system. If you don't use the, the way that it was designed, then I'm not responsible for it. It, it, again, it, it's like saying you can't complain to me that your airplane doesn't work as a submarine because it wasn't designed to.

22:31Matt Coles Psychological acceptability, psychological acceptability is a fundamental security principle.

22:36Izar Tarandach In my mind, they do not accept that.

22:40Matt Coles You're the, you're the threat. He's the threat.

22:44Izar Tarandach Yeah.

22:44Chris Romeo So we have a new threat.

22:45Izar Tarandach Oh, I have the tissue.

22:48Chris Romeo The new threat is a crotchety security person who takes part in your threat model. They keep shooting down all the good innovative ideas.

22:58Izar Tarandach So, Ludvig Damaj?

23:00Matt Coles Yeah. So, so psychological acceptability, right? This, this notion that if what you build from a security standpoint isn't usable by a user, then, then that itself is a potential area for threats to occur, right? So there's, there's active attackers who try to defeat your, your security, and then there's those who inadvertently, that's why we call it misuse cases, not abuse cases, right? So if you're build, if you're building a system, and actually maybe go back to, to an earlier part of this conversation, if the designers had put together misuse cases And that misuse case had documented what happens if somebody doesn't answer the phone, then maybe you wouldn't even gotten to the point of having a threat model in the first place.

23:47Izar Tarandach Yes.

23:48Chris Romeo Yeah. That's a good point.

23:52Matt Coles Now, something else I want to just throw out here, and because it's sort of inadvertent, there's inadvertent threats that occur that are like outside of the control of, of normal people. There's a whole class of, of weaknesses that are around environmental impacts. Right? So if your system is designed in a certain way, do you worry about electromagnetic interference generally? Do you worry about heat transfer and other, and other things, right? So there are, there are operational threats that we don't, that there are operational threats that often get overlooked when it comes to threat modeling, right? Because you're looking at security principles, you're looking at communications. You're looking at data and there's a limit there to what you can identify through that. But if you look at operational technologies, and I don't mean OT, I mean, if you look at operational technol— operations of systems that you're threat modeling, this happens a lot in OT systems, but not just there. Uh, you have a whole class of potential threats that come out from things like Is my data center going to be overheating? If my data center overheats, what happens to the system I'm building? Right? And so this notion of, well, what happens if my users misbehave or can't use the system properly? Why are those not threats too?

25:16Chris Romeo Now you're talking me out of my previous conclusions.

25:20Izar Tarandach I wish you'd complete.

25:22Chris Romeo But no, no, only for this reason. Right? When you started talking about all these operational threats and operational challenges, my, the complexity level started rising and rising to the point where threat modeling could become a monster, a villain almost, because it's so complicated. Oh, well, did you consider the heat transfer of the, you know, the, uh, single board A and board B? Did you consider, you know, famine in— Oh, seriously.

25:51Matt Coles And, and talking, and talking about the, and talking about how an attacker actually defeats a TLS connection or or even an HTTP connection over an adjacent subnet, and that's not complex. I mean, obviously we bring this up to a much higher level, right? So you don't talk about heat transfer, you talk about what happens if the data center's on fire. It's—

26:10Izar Tarandach So, big parenthesis here. Once I was listening to a professor explaining product development, and he got into the example of heat transfer. And the guy was saying that calculating heat transfer in consumer electronics is a really, really important thing. And somebody raised their hand and said, uh, oh, obviously you have never worked with a laptop on your lap, right? And the guy just stopped, looked at him, and he went like deadpan, it was transferring heat, right?

26:48Matt Coles What's, what's the heat resistance of your leg? With or without hair?

26:55Izar Tarandach If you get to decide, do you use spandex or, or, or cotton to go to the park?

27:03Matt Coles And what's the added impact when you spill coffee on yourself as a result?

27:06Izar Tarandach Yeah. And then if you have a, a, a matchbook of matches on your pocket. Oh, right. But no, again, you know, it's, it's, it's, I, I need to go back to my previous point of, again, a lot of times people, people ask, okay, you're doing threat modeling, how do you know you're done? And I think that by now we sort of have some good answers to that, sort of. But now tell me that I have to add the whole external world to the system to that in ways that are not immediately impactful. To the function and working of the system. And next time somebody asks me, when do I know that I'm done? I'll, I'll just give up.

27:53Chris Romeo And, and, but that's the classic separation of risk management versus threat modeling in my world, my mind. Like a fire in the data center should never appear in a threat model. That's a risk management problem that the, that the, there's a whole team inside of large companies that sits around and thinks about what if a tsunami hit this particular spot? What's the probability of it?

28:21Matt Coles Are those not threats?

28:23Chris Romeo But it, it's classically separated into a risk management versus threat modeling problem. Because once again, from a complexity perspective, are you going to ask a developer to consider what happens if a tsunami hits a data center somewhere in the world? No, because it's outside of their scope. It's a scope problem.

28:39Matt Coles It's outside their scope, right? But you might ask a product owner, you might ask a product owner who's working on a critical banking system, what happens if, what happens if something bad happens to that banking?

28:51Chris Romeo Yeah. But it's the—

28:53Izar Tarandach And their answer, let me ask the customer.

28:56Chris Romeo It's the risk management team that tends to own those types of scenarios though.

29:02Matt Coles But we're not talking about who owns the threat or the outcome of that threat. We're talking about, can you, should you be generating those threats?

29:12Izar Tarandach Should you be considering those threats?

29:14Matt Coles Threat modeling informs risk management, and whether or not we're doing it as security, you know, security folks, or developers are doing it as programmers, or GRC folks, and, you know, risk analysts are doing it as part of the broader, I agree with you, these things tend to sit at a much broader level, you know, a GRC organizational level. Than, than what the developers are focusing on, but it's still threat modeling.

29:41Izar Tarandach Yeah, but you—

29:42Chris Romeo whoa, whoa, whoa.

29:43Izar Tarandach Now you're going— what?

29:45Chris Romeo So you're saying threat modeling and risk management are the same thing?

29:49Matt Coles I'm saying that one informs the other. Threat modeling, threat modeling is creating a model and looking for what can go wrong with that model. Risk management is now that you know what could go wrong, how do you address it? Or how do you manage it?

30:02Izar Tarandach But now we're going like dangerously close even to compliance because a tsunami is something that can be a risk to everybody, more or less if you're close to the sea, but it's a threat to everybody. Uh, uh, uh, nuclear bombs, threat to everybody. Alien invasion, threat to everybody. So those threats exist for everybody. It doesn't mean that you have on every single threat modeling session to bring them up or in every risk management. Recession, like, I mean, hand of God, we, there is stuff that we can't depend against. Now, what's the probability that's going to happen?

30:35Chris Romeo I don't know.

30:37Izar Tarandach Not my betta weekend. I don't want it to be. But, uh, it's okay. The, the, the point here is that even in risk management, at some point you give up saying these things here are completely out of my control. They may happen, definitely. Are they going to impact me? Probably. What can I do about it?

30:55Chris Romeo Nothing.

30:58Matt Coles But you have heard of— I wanted to say, have you heard, but you've heard of me, but I can't. I'll just—

31:04Izar Tarandach well, somebody, somebody tried to call me about them, but I didn't pick up the phone.

31:10Matt Coles All right.

31:11Chris Romeo Well, I think we've taken this article as far as we can at this point. So we'll put a link in the show notes so people can read the original. And, uh, I get that.

31:21Izar Tarandach I think it's a great idea. We just have to approach it the right way.

31:26Matt Coles Yeah.

31:26Chris Romeo All right, folks. Thanks for listening to another episode of The Security Table.

Transcript supplied by AssemblyAI. It may contain transcription errors.

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