30 minSeason 3, episode 6

Threat Modeling or Threat Intelligence, Are they the Same?

With Chris Romeo, Izar Tarandach, and Matt Coles

Threat modeling & security architecture

Listen in as we debate the differences between threat intelligence and threat modeling. What distinguishes these two concepts in cybersecurity, and how do they inform each other?

Episode title image: Threat Modeling or Threat Intelligence, Are they the Same?

Listen to the conversation

About this episode

Listen in as we debate the differences between threat intelligence and threat modeling. What distinguishes these two concepts in cybersecurity, and how do they inform each other? The conversation explores definitions, real-world examples, and the interconnected relationship between proactive threat modeling and reactive threat intelligence.

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

0:08Chris Romeo Hey folks, welcome to another episode of The Security Table, where before we hit record, we were debating Asimov's Three Laws of Robotics and how they apply to AI. Bet you really wanted to be there for that conversation, but it was actually really quick. As always, happy to be joined by Matt Coles and Izar Terrindosh. Cheers.

0:33Matt Coles And we're all caffeinated.

0:35Chris Romeo Yeah, you're getting Izar at maximum caffeination right now, which is a good place for him to be because he just tends to fly off the handle even more, which is good. All right. So Matt, what are we talking about? Something about threat intelligence. I don't know that I totally understand what we're talking about. Oh my God. So set the table for us. All right.

0:54Matt Coles I'm going to ask a simple question and you guys are going to go, right? Discuss threat modeling or threat intelligence. Are they the same?

1:08Izar Tarandach No.

1:09Matt Coles Next.

1:11Chris Romeo All right.

1:11Matt Coles We're done.

1:12Chris Romeo That's it.

1:12Matt Coles Go home.

1:14Chris Romeo I'll argue yes.

1:14Matt Coles Mic drop. Gotta go.

1:16Chris Romeo I'll say yes, just to make this an interesting episode. Yes, they are the same.

1:20Izar Tarandach No, no, but there's the thing. There's the thing. I don't think that people are saying that They're the same thing, but for some strange reason that I simply can't fathom, they're saying that one is superseding the other.

1:33Matt Coles I, okay, we're talking about this article was written in DocReading about practices around threat modeling, but it heavily relies on using threat intelligence as a means of identifying and managing threats. And I think we want to argue here, or I want to argue that threat modeling has a different purpose than threat intelligence, and threat intelligence should be the consumer of threat modeling, not the other way around.

2:07Chris Romeo So when you say threat intelligence, let's work up a definition of that, because I feel like we've talked about threat modeling about 784 times on the Security Table.

2:18Matt Coles There's only a few books and a manifesto and a conference dedicated to this and a podcast that's dedicated to this.

2:25Chris Romeo And that manifesto does have a definition that we all agreed on, analyzing representations of a system to, I can't remember the last part, to, I always say to identify privacy and security threats. I always add to find because I can never remember what word we used to connect it, but to identify security and privacy problems.

2:43Matt Coles Concerns. Concerns.

2:43Chris Romeo Concerns.

2:45Matt Coles Wait, that just blew my mind.

2:46Izar Tarandach Okay.

2:46Chris Romeo Yeah. Well, we'll come back to that. But like, when you, when we say threat intelligence, Matt, like, what is that? What's your definition of threat intelligence?

2:53Matt Coles Well, I don't know if it's my definition, but my understanding of threat intelligence is if you are, if you run a service or you, or you're managing a system, or I mean, I guess product companies would have this as well. You know, if, but it's primarily I think it's traditionally cybersecurity where you're looking at all of the incoming attacks, 'cause of course every company is getting flooded with attacks, right? Whether they're scripted or not. And you're really looking at who from an attacker is doing what, those attacks, and what's their goal against the services that you run, right? So, of course, MITRE ATT&CK framework is a, a taxonomy and a dictionary, well, not dictionary, what's the word I want to look for? Some collection of attackers and their behaviors and their techniques that they use. And so threat intel goes from that understanding who's going to come attack me to what are the active, active, I'm going to use the word, active threats against my system and by who.

4:05Izar Tarandach So just to bring that to a one— of course, like many things in this field, comes from the military. The definition of intelligence in the military tense, not the other tense, is that it's a military discipline that uses information collection and analysis approaches to provide guidance and direction to assist in strategy and tactical definition, right? So it's people who are looking outside People who are looking at what exists, what they can predict based on what exists with good assumptions, and then come and say, this is what we believe is going to happen at a certain probability.

4:45Chris Romeo You know what that sounds like? Threat modeling.

4:49Izar Tarandach No. That's the thing. That's the mistake. That's the mistake. No, it does not.

4:56Chris Romeo It does awfully.

4:57Izar Tarandach It does not.

4:58Chris Romeo Thinking about what could basically go wrong is what— if I summarize what you just said. So there's a, there's a looking to the outside world to, to consider what could possibly go wrong. That's what I heard you say.

5:09Matt Coles Oh my God.

5:10Izar Tarandach I think, I think I just had a, I just had a, there's a basic difference here. And that's why Matt blew my mind with what he opened up with.

5:17Chris Romeo Okay.

5:18Izar Tarandach If we look at the pipeline, if we look at the pipeline of information, in my opinion, in my experience, Threat intelligence sets up the ducks. These are all the possible ducks, all the ducks that are coming our way with high probability that they are going to land. Threat modeling puts those ducks in a row by saying, this duck is more possible than this duck. And if this duck happens, this is how we're going to shoot it down.

5:52Matt Coles So threat modeling— So they go hand in hand is what you're saying.

5:54Izar Tarandach They go in a pipeline.

5:56Matt Coles Okay.

5:56Izar Tarandach From all the sources that we use to elicit threats in threat modeling, threat intel is a very important one. Now, threat intel also suffers, again, in my opinion, my experience, I could be wrong, probably am. Threat intel also suffers from the generic issue of all forms of intelligence. It all depends on how you collect your data. It all depends on who interprets your data. It's fueled by biases and opinions. So I would say that it's a subset of the incoming data that should serve your threat model. It's not the totality.

6:36Matt Coles So let me ask you a question there. So I hear what you're saying, and I think I understand this. So if you were doing threat modeling and you use, say, MITRE ATT&CK as the source of what could possibly go wrong, versus threat intelligence is what's going wrong today. Is that, is that a good way of like, there's the theoretical—

6:58Izar Tarandach Yes, threat intelligence is a subset. Threat intelligence is a subset of what might happen informed by what we see happening outside, right?

7:09Matt Coles Kind of like—

7:10Chris Romeo Yes, but that's across everybody though, right? So it's not, it's not what's specifically happening in our environment. It's let's look at the macro conditions and macro environment as all possible things that could happen.

7:22Izar Tarandach Yes and no, because once you go into the people who do this, I'm just talking from hearsay, I never did it myself. Talk about people who do this for a profession, they look into silos, they say, okay, these specific actors are targeting people in the finance sector, in the entertainment sector, so they're able to do some sort of subsetting, add filtering to say these are the threats that are more present in these verticals and sell that as targeted intelligence to those verticals. So you could go to one of these professional analysis people, threat intel people, and say, hey, I'm in this area, I'm interested in knowing what's prevalent in here and how and what.

8:10Matt Coles And that's the service, that's the service though, right? You're the, the general scope of threat intelligence is what's currently happening today, either against my system or my company or my industry or geopolitical, whatever, you know, um, nation state actions, you know, if somebody's cutting a telecommunicate, uh, undersea cable somewhere, my business is going to get impacted. So that's a threat I need to understand, right?

8:35Izar Tarandach True, but the way that I, the reason that I conflate both is that as you can imagine, you need so many resources, time, good people, content to be able to figure out that something is happening, have good intelligence. Having bad intelligence is worse than having no intelligence at all.

8:56Matt Coles Because you make bad choices.

8:59Chris Romeo I would like to understand. I'd like you to reverse engineer this then and explain why threat intelligence is not threat modeling according to the definition, because I'm starting to get a little bit fuzzy in the middle here. Like I thought, I started this conversation, we're like, we're going to talk about this. I'm like, oh, this is easy. We're all just going to be like, yeah, threat intelligence is not threat modeling. But then when you read that definition of intelligence and I started thinking, wait a minute, that sounds kind of like what threat modeling is. Maybe not exactly, but it's kind of like it. So reverse engineer it for me.

9:32Izar Tarandach Let me play with it for a sec. You're sitting on top of the castle, you're looking outside, you're seeing a cloud of dust coming in the horizon. The fact that there is a cloud of dust coming in the horizon does not inform what's the size of the moat that you need, how deep it is, how many sharks you're going to put in there. What's the size of the drawing bridge? Do you need a drawing bridge? Do you need two? Are they in the right direction? So threat intel is going to tell you there is a cloud coming, cloud of dust coming your way. That could mean that there's a crusade coming this, the way of your castle. But it's not going to help you dimension how you're going to build that castle or how you're going to defend it or—

10:20Chris Romeo So threat intelligence does not include mitigations.

10:25Izar Tarandach Yeah, no kidding.

10:25Chris Romeo Is that what you're saying?

10:27Matt Coles Oh, okay, hold on. Uh, I think we're dancing around a couple things here.

10:33Chris Romeo So, probably.

10:34Matt Coles I, but I, but I understand, I understand both your points now. And I, and I'm, and I wanna, I wanna modify what I said earlier about, about, uh, threat intelligence as a consumer of threat modeling. But threat intelligence informs a threat model. So, We are relating, and because we, we see ideally that threat modeling is done during architecture and design, right? You do it upfront. What can possibly go wrong based on what I know as I'm building this thing, or as I'm designing this thing? Threat intelligence is what could go wrong based on what I'm seeing in the real world now that I have a functioning system. Right. Or most likely, right. That's because that's the silos that, that cybersecurity space have, have put those things into. The threat model informs threat intelligence. What am I looking, what am I going to look for? Threat intelligence informs a threat model.

11:33Izar Tarandach No.

11:34Matt Coles Why? Why no?

11:37Izar Tarandach Because threat modeling, threat modeling, as you said, fixes the thing in time. In some way. Okay. And we are not talking now about a system. I think that we are talking more about an organization, about the way that it defends itself as an organism. Threat model exists. The threat intel changes with the threats. So the threat intel might possibly see threats that did not exist at the time that the threat model was executed, even though we keep talking about continuous threat modeling and all that good stuff. We know that not everybody does. So sometimes the threat model has been fixed in time for long enough that threat intel will be able to inform it with something new that wasn't, that didn't exist when the threat model was in place or that wasn't considered or something that was too complicated and all of a sudden because of the crawl of it became very simple.

12:32Matt Coles Right. So this is, this is where Chris's idea is coming in. So, and again, why I say threat intelligence informs threat modeling, and threat modeling basically becomes a consumer of threat intelligence. So it's a bidirectional situation here. Threat modeling is based on the information you know, the architecture you're about to build, and the threats that that architecture brings in from a— I'll be careful here— from an academic sense, right? You have a taxonomy of things that can possibly go wrong. And what, when you do threat intelligence, you're going to say, oh, well, I'm really finding these problems. So let's update the model because now I've maybe found a new method of attack or a new entry point or whatever that I didn't realize was there. And in fact, you know, if I design something and then it gets built, well, it may get built differently than how it was designed. And now I have new stuff to worry about, but I can, but I can also see the threat modeling would inform threat intelligence, and maybe this is limiting it too much. Well, I'm building a castle, I have a moat, I have portcullis, I have a drawbridge. I'm, if the model doesn't say I have an escape tunnel, you know, half mile down the road that opens up into a village, I'm not gonna look for threats against the village, although I should. And so the threat model could inform threat intelligence, as terms of what you want to look for. So there may be a feedback loop here, but they're not the same. I think they're different in that threat modeling is actively looking at the system from a, what, well, what can go wrong standpoint and identifying do mitigations need to occur. Threat intelligence is looking at what's coming at me right now or soon to come at me right now. And do I need to worry about it? Do I need to mount a defense? And Chris, you're on mute.

14:28Chris Romeo It's been a long time since I've done, I've done this, so I forgot how to, how to work the system. Um, now you, since you've both used the castle analogy, now it's my turn cuz it's only fair, you know, I should get to use the castle too. But here's what I've been sitting here thinking about in regards to the castle and the dust blowing at me. Whether the threat originates from a taxonomy or in my head from a threat modeling perspective, or whether it originates because it's actually happening to somebody in my network or someone there, seems to me like it's the same thing. Doesn't, why does it matter where it, if we're talking about eliminating threats in the system, Like the ultimate goal of threat modeling is to identify what are the things that could go wrong so that we can apply mitigations to them. Threat intelligence is, if it's not the same thing, it's a really close cousin. It's just a different source. And I did a little bit of research while we were talking here, and threat intelligence does often have recommended controls or configuration changes. So it does have a mitigations component to it. So it's not, it's not separate from mitigations. Although I then throw back on the table and say, this is very, this is more similar than it is different.

15:49Izar Tarandach Yeah, those mitigations, they are attached to the attack, not to the system.

15:57Matt Coles They're not, they're not foundational mitigations. When we do threat modeling, we would do foundational mitigations, add a moat, not go out and figure out, is that an approaching army or not?

16:09Izar Tarandach Those mitigations that come from definitely threat intelligence, they are more, this is the kind of threats, these are the things that we have on the drawer.

16:18Chris Romeo I mean, the example ChatGPT gave me is recommended controls or configuration changes, advice like disable Office macros by default, restrict PowerShell execution, or enforce MFA on remote access points.

16:33Izar Tarandach Ah, the thing is, ChatGPT is telling you something that you wanted to hear.

16:36Matt Coles The big difference here—

16:37Chris Romeo No, this is in a bigger context. I just asked it, what, how does threat intelligence include ways to fix the problems?

16:42Matt Coles Yeah. So I think the bigger, I think the, I think what you're highlighting is in threat modeling, and I was just, I think I was just talking about this, the, in, when we do threat modeling, traditionally threat modeling, we're going to look at architectural fundamental by design mitigations, right? If we have insecure communications, we're going to turn, we're going to enable security. With threat intelligence, when you identify what can go wrong, you can only mount a runtime defense, right? Either you're going to be proactive about reaching out to the attack and trying to figure out either its characteristics or to stop it there, or you're going to add a layer of, you're going to add, you know, spikes, spikes in front of the drawbridge to slow down the incoming army. You can only mount runtime defense. Whereas threat modeling traditionally, you can only mount an architectural defense, but you can inform runtime defenses.

17:35Chris Romeo I don't know that I agree with that because the example that my best friend ChatGPT gave me about enforce MFA on remote access portals, so—

17:42Matt Coles Will you choose a different model? Like, choose Claude or something.

17:45Chris Romeo No, this is my friend 'cause it helps me.

17:48Izar Tarandach We take no sides.

17:49Chris Romeo And I trust you.

17:50Matt Coles We are not biased.

17:51Chris Romeo We're no product of the platform here. If enforce MFA on remote access portals is a mitigation to some threat intelligence thing, That's an architectural play. That's not a runtime defense.

18:02Izar Tarandach No, no, that's a desperate play. If you haven't yet done that and you're waiting for threat intel to tell you, oh, by the way, perhaps you should.

18:12Chris Romeo True.

18:13Izar Tarandach Then you got a problem.

18:14Matt Coles So, oh yeah.

18:16Izar Tarandach It's what people say.

18:17Matt Coles That's really the important part. Threat modeling is proactive. Threat intelligence is reactive.

18:23Izar Tarandach There you go.

18:23Matt Coles Okay.

18:25Izar Tarandach And funny thing, it's reactive not to the attack happening, it's reactive to the attack going to happen.

18:32Matt Coles Yep.

18:33Izar Tarandach So in terms of time travel, it's, uh, we will have reacted, reacting.

18:40Matt Coles I'm gonna draw an analogy here because this is something else that's come up recently and I'm gonna take this, I don't wanna take us in a new direction, but I just wanna throw it out there. The debate between CVSS and EPSS?

18:56Izar Tarandach I don't know. I'm still working on OBSS.

18:58Matt Coles Well, CVSS is, you know, vulnerability scoring. What's the potential for damage? EPSS is what's the potential for attack of that, you know, exploitation of that vulnerability. Debate raging about how they inform risk. And, you know, and what action, you know, companies should take. Well, the same thing with threat modeling, right? So threat modeling is going to come up with a bunch of stuff that you need to address. Threat intelligence is more real-time. It's more reactive, but it's more from a defender standpoint. I can see that attack coming, whereas in threat modeling, it's theoretical.

19:42Chris Romeo Could be theoretical completely and never— Hot take.

19:44Matt Coles Hot take.

19:45Izar Tarandach Go for it. EPSS is threat intel on the cheap.

19:50Matt Coles I wouldn't say on the cheap. It uses complex algorithms to figure it out.

19:54Izar Tarandach But I'm not paying for it.

19:58Matt Coles Oh, sure. Sure. You know, there is, there are public sources of threat intelligence. You can gather your own information.

20:04Izar Tarandach Definitely.

20:05Matt Coles I mean, that's what, isn't that what like Shodan is and things like that where you can go and Yeah, you have to work a bit harder for that because—

20:12Chris Romeo I'm sure people don't try to put bad data in those sources either. Yeah. Mess with you.

20:17Izar Tarandach But I, I do, and, and that's, that's, that's a totally real thing. Like, if you want people to not look that way, you make them look that way.

20:25Matt Coles So I, I agree with you, Chris. It, they are, they are different, but they are not that different.

20:31Chris Romeo Yeah. I, I started this conversation with, with the, what, what had been my classic answer to this and that they're Ha! You're just wrong. They're completely different. And this is the more we've gotten into it, the more it's like, these are like really close cousins of each other.

20:46Izar Tarandach Yeah, I still think there's a pipeline.

20:49Matt Coles I, they definitely are distinct things, but their purpose is ultimately to secure a system. One does it a little bit more upfront, proactively. One does it a little bit more at the back end. Reactively, and they can inform each other. Data can go in both directions.

21:08Chris Romeo Yeah. It's like static and runtime, basically.

21:12Izar Tarandach Yeah, but I still think that the threat intel informing the threat model is a more dynamic and useful thing than the threat model informing the threat intel. Because the threat intel by definition has to have a broad scope of the things that it looks like. By using the threat model to inform it, hey, these are the things that we are not so certain that we did right, so perhaps you should look at it. You are scoping that thing. Can be too much.

21:43Matt Coles Only if you artificially limit it that way. I mean, if you don't know what information you don't need from a, like, we don't have PHI in our system. I don't need to worry about healthcare threats. Right, in my threat intel feed. Why? I mean, you can use it to filter information and maybe target the things that we made assumptions about. But that doesn't mean you don't look at anything else in your threat intelligence.

22:10Chris Romeo Your threat model informs threat intelligence because it's the best place you have a representation of what the system is. Like in threat intelligence by itself, how do you know what the system actually is? Configuration files or?

22:23Matt Coles You can't, and you can't possibly look at all the threat intel that all that possibly exists and understand how it impacts you without the threat model. That doesn't mean artificially limit things. Well, I mean, it may, but it shouldn't. Okay. So where'd we land?

22:46Chris Romeo So I think we got a definition. At the 22-minute mark, we've come up with a working definition, but we've ignored the article that got us to this conversation.

22:57Izar Tarandach I think that we didn't come to a definition. I think that we came to a working situation.

23:03Matt Coles Well, so the article is Threat Modeling is Useless and Threat Intelligence is the Way. And I think we've just counteracted, I think we just contradicted that, not contradicted, but refuted that.

23:16Izar Tarandach I find that, okay, and this is going to come from a loving place because I spoke a lot about threat intel without ever doing it. I find that a lot of people who do threat intel and say that have never done a threat model.

23:29Chris Romeo Yeah, that's true.

23:31Matt Coles That's for sure. You know, there definitely is this, you see across the cybersecurity landscape, you do see this a lot, right? There's people in the, there's people in the runtime defense camp who are unfamiliar with development practices, and there are people in the development camp who are very unfamiliar with runtime practices. At some point, you kind of need to know both in order to be effective.

24:00Izar Tarandach You are this close to talking about DevSecOps. I'm just alerting you.

24:03Matt Coles I'm not. I'm absolutely not. I'm absolutely not talking that. Throw a quarter in the jar. There's the understanding and then there's the bad implementation of that.

24:14Chris Romeo I'm reading this article again. I'm reading this article again. And I think it suffers from a common problem that we've seen across threat modeling for the last 5 or so years. And that is—

24:28Izar Tarandach Was it written by ChatGPT?

24:29Chris Romeo No, no, because that ChatGPT is my friend. I already said that I would have been complimentary if it was ChatGPT. Uh, it's people, people continue to try to define threat modeling for their own purposes and not, they don't use like the manifesto definition that we all agreed on and fought over and whatnot and got to a good place with. And then the more I'm reading this a second time and I'm like, all they're doing is they're trying to use threat modeling as a synonym to threat intelligence just because it works as a word, not because they're trying to make a stand or make an argument that threat modeling and threat intelligence are the same.

25:12Izar Tarandach I think that threat modeling suffers from the Humpty Dumpty phenomenon more than other things, where people just like impart the meaning to the word that they wanted to have rather than what agreed to be.

25:27Chris Romeo And I've seen that in the threat modeling kind of vendor startup space where some people are using the term threat modeling and it's like, not, that's not threat modeling according to the definition of what the manifesto calls threat modeling. Like you can just take, I mean, I guess I can take any two words and put them together and say that I do, like I do runtime security. What does that mean? Well, I'm redefining what runtime security is. Like it doesn't mean anything if you just take it and try to twist it to fit your strategy or your, and I've seen that a couple of different times where people And people are like, well, we're doing threat modeling. And I'm like, well, not according to the Threat Modeling Manifesto, you're not, because that's not threat modeling. And there's nobody else has really put out a definition. There's no competing definition.

26:12Matt Coles Where's your model? And where are your threats? I mean, it's really, it's kind of baked into the name.

26:17Izar Tarandach Yeah. You model the model, you elicit the threats.

26:22Chris Romeo This is like turning into like, it's like Seinfeld, like, describes threat modeling.

26:28Izar Tarandach You know how to get the threats, but you don't know how to model them.

26:31Matt Coles And if we, and if we If we use those words, if we use those words explicitly, you model a system, you identify threats, you have intelligence about a system to identify threats.

26:43Chris Romeo Yeah. I like that.

26:45Izar Tarandach But you got too many, you got so many words in between, it confuses me. Too many words. Too many words. You? Two of them are the same thing. That means that at least half of it is the same thing.

26:58Matt Coles Yes, I could, I could, I could.

27:00Chris Romeo Are you in management? Are you in management now? What just happened?

27:04Matt Coles Wait, wait, there's a, there's a, he, he's a—

27:07Chris Romeo I can get away with that cuz I don't, I don't ever have to have a job again in my life. So I, I could get away with saying that. That was me. Let the record show, that was my comment. That was the voice of Chris you heard there.

27:17Matt Coles That was Chris's fault.

27:19Izar Tarandach Ending page.

27:23Chris Romeo Uh, a lot of, a lot of pretty-haired people out there were like, wait, what?

27:28Izar Tarandach Draw Matrix for me, but make it thick.

27:32Chris Romeo Yeah. And again, they put it in a PowerPoint and in 200— in no more, in no more than 16 colors and explain it to me like I'm 5. That's actually a true management principle. I've used it before. I'm like, yeah, you know, I need to explain that to me like I'm in second grade because I don't know what you're talking about.

27:56Matt Coles I don't know. All right, we could talk at length about executive reporting and executive, uh, we're gonna slide to a new topic.

28:06Izar Tarandach Hey, it's, it's Monday morning. Let's not talk about management.

28:13Chris Romeo So did we come to some realization? Did we come to a conclusion? Like, where are we landing on this?

28:19Matt Coles I think, I think we've decided the article, I was gonna say full of something, but it's—

28:27Izar Tarandach Okay, let's see if we can agree on this. So we got two working definitions for each one, for Threat Intel and for Threat Model.

28:35Matt Coles Fine?

28:37Izar Tarandach We do agree that they do have a relationship. Yep. We do agree that that relationship is not one is dead and the other took over.

28:46Chris Romeo Yes.

28:47Izar Tarandach We do agree that there is value in them working together.

28:51Matt Coles Yes. Yes.

28:52Izar Tarandach Okay. So let's go embrace somebody who works at the SOC and everybody's happy.

28:59Chris Romeo We're going to sing Kumbaya or what's going to happen?

29:01Matt Coles So wait, you want, you want developers to go talk to the SOC people?

29:06Izar Tarandach No, no, no, no. I said we, we, the three of us.

29:09Chris Romeo Not even the security. We're not recommending security.

29:13Matt Coles You want, you want the SOC people to show up at the developer's desk? Can you imagine the pandemonium that would occur?

29:18Izar Tarandach No, no, you do not put the elephant and the hypo together.

29:26Chris Romeo No. Well, you know what? We're at the 30-minute mark. I think this is a good time to wrap up this conversation. And we have made the world a slightly better place today. So once again, thank you for tuning in to the Security Table. For Matt, for Ezar, I'm Chris, and we'll see you next time around.

Transcript supplied by AssemblyAI. It may contain transcription errors.

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