# We'll Be Here Until We Become Obsolete

URL: https://securitytable.ai/we-ll-be-here-until-we-become-obsolete/
Published: 2024-10-23
Duration: 1668 seconds
Season: 2
Episode: 30
Hosts: Chris Romeo, Izar Tarandach, Matt Coles

## Show notes and chapters

This week we explore the multifaceted concept of obsolescence in technology, detailing its planned, unplanned, and forced forms. We delve into the security implications of outdated or unsupported devices and software, with a spotlight on cloud-connected vehicles and their vulnerabilities. We discuss architectural decisions, regulatory requirements, and real-world incidents like the OnStar hack, reflecting on the need for robust security protocols. 

  
  

Follow AI Security Table:

➜ Home: [https://securitytable.ai/](https://securitytable.ai/)

➜ X: [https://x.com/AISecurityTable](https://x.com/AISecurityTable)

➜ LinkedIn: [https://www.linkedin.com/company/ai-security-table/](https://www.linkedin.com/company/ai-security-table/)

➜ YouTube: [https://www.youtube.com/@AISecurityTable](https://www.youtube.com/@AISecurityTable)

## Transcript

**Chris Romeo (0:08)**: Hey folks, welcome to another episode of The Security Table. This is Chris Romeo joined by Izar Taran Das and Matt Coles. And we talk about security, privacy, the meaning of life, and Very few sports references. I try to sneak them in, but then Matt and Isar look at me.

**Izar Tarandach (0:34)**: What are you talking about?

**Matt Coles (0:35)**: Sports? What's that?

**Izar Tarandach (0:38)**: Never played baseball in my life.

**Matt Coles (0:39)**: You'd have to go outside for that, right?

**Chris Romeo (0:42)**: Yes. Matt, go outside.

**Izar Tarandach (0:43)**: What is this outside you keep talking about?

**Chris Romeo (0:47)**: Come on, you did a live— or you did a security table remote from the woods somewhere.

**Izar Tarandach (0:53)**: Did you see the hat? Did you see the hat I had on me?

**Chris Romeo (0:56)**: Yeah.

**Izar Tarandach (0:57)**: That, that's a lot of fear. Like half an hour ago, I was sitting here turning on my computer and then I felt something in my leg. It was a freaking spider. Wow.

**Chris Romeo (1:08)**: Now, if it was a radioactive spider and it bit you.

**Izar Tarandach (1:11)**: Oh, then everything would be okay.

**Matt Coles (1:13)**: Yeah.

**Chris Romeo (1:13)**: I mean, origin story.

**Izar Tarandach (1:16)**: It was a bug.

**Matt Coles (1:17)**: It's okay. Was there a flaw? Was it a weakness?

**Chris Romeo (1:21)**: Yeah.

**Izar Tarandach (1:22)**: I put it in a jar and I'm calling it future.

**Chris Romeo (1:28)**: So Matt, you're struggling for coffee.

**Matt Coles (1:31)**: I am. I, you know, my Nespresso machine died last week. And which is interesting because it brings up our topic for today, I think, around obsolescence, planned or unplanned. Coffee machines die at random for whatever reason, whether that's intentional or not. It's interesting. But I guess we have bigger problems, it seems, with internet-connected devices and obsolescence. Yeah.

**Chris Romeo (2:05)**: So planned obsolescence is when a product is end-of-life. Like the vendor is saying, hey, you've got a couple of years left on this. And I'm not, I'm not making any more software updates for it. Yeah.

**Matt Coles (2:18)**: Or, or it's, uh, reached its, it has reached whatever lifespan the, the vendor or manufacturer believes it should have, whether that's from a software standpoint or from a hardware standpoint, right? Planned obsolescence, it goes end of life. Unplanned, on the other hand, is where it just stops working. Yeah.

**Chris Romeo (2:37)**: Like the coffee machine.

**Izar Tarandach (2:38)**: So we have planned We have unplanned and we have forced.

**Matt Coles (2:42)**: Yes.

**Chris Romeo (2:43)**: Okay. Planned, unplanned, and forced.

**Izar Tarandach (2:44)**: Those are the 3 categories, or am I forgetting something?

**Matt Coles (2:47)**: No, I think that covers it, right? So planned is when you expect it to end. Unplanned is when you don't expect it to end. And forced— now forced is interesting because that's the vendor saying, well, it's time to end, even though it may still be a functional device.

**Chris Romeo (3:02)**: Yeah.

**Matt Coles (3:02)**: Or functional software.

**Izar Tarandach (3:04)**: The next question is easy.

**Matt Coles (3:07)**: Planned?

**Izar Tarandach (3:07)**: Is forced part of planned or is it a separate thing? I keep feeling the spider in my leg.

**Matt Coles (3:17)**: So, so I think, I think you'll, I think you'll find, at least in my experience, and I, and I, you know, from coming from working with a number of device manufacturers and, and, or companies that, you know, that build physical things that ship to customers. Planned obsolescence is, there's certainly a plan. You don't want to expect a lifetime of something beyond, you know, technology, right? If you do, if you have certain technologies, you want to be able to, and of course you want to sell more, right? So you want to be able to get these things into the market over time. Unplanned, unplanned happens, but forced obsolescence, I think, is a plan. But it's not necessarily a good plan. Yeah, I know. I can't do anything about it at this point. Um, so, uh, yes, yes. Forced, forced abstinence is a plan. Is it a good plan? Not, not sure it's a good plan.

**Chris Romeo (4:22)**: So we have the issue that people claim, and that is, let's take a device like the iPhone.

**Izar Tarandach (4:31)**: Cool. Straight into the belly of the beast. Yeah.

**Chris Romeo (4:33)**: Yeah. After, so after, after 2 years or so, it seems like the device is operating less optimized. And then it always, it begs the question that a lot of people on the internet have made this accusation that big companies like Apple or Google Android or whatever are purposefully architecting their software in such a way to push the upgrade cycle. Because I don't know about you guys, but I've gotten off the, like, every— I need to upgrade every year. Like, I just upgraded my phone after like 3 or 4 generations of iPhone because I'm like, nothing really changes. Like, it's the same thing.

**Matt Coles (5:18)**: Yeah, I mean, first Go ahead, sorry. No, no, you go. Go, I've been talking. Go.

**Izar Tarandach (5:25)**: I'm like, okay, first of all, why are you using an iPhone? But beyond that, anybody who says something like that, my first reaction is, okay, so you never touched a Windows computer for some period of time. Because to be surprised by that is like, excuse me, it's a fact of life. The thing with the alleged iPhone thing is that while people have made a sort of peace that Windows at some point you're going to have to reform it, that thing just to keep working, right? The iPhone, that there is an assumption of reliability on those things that we put in our pockets. And we spoke about handsets a number of times around here. We do put trust and we do expect reliability from those things in a way that we don't from other from other things in our lives. And I think that if this is forced obsolescence, now comes to the question, okay, what does obsolescence has to do with security? Because at the end of the day, this is the security table. And I think that last week we looked into, in between, after the last episode, we looked into this article that was talking about this car company car maker, right? That they have electric vehicles and they sold themselves as software-defined cars. And apparently one day they went bankrupt and their servers went down. And all of a sudden what you have is a great long moan that looks like a car. Those things just don't move without the servers. Yeah, so in this apocalyptic scenario, what are the security implications of obsolescence when all of a sudden the services that you rely on, the servers themselves, go away for one reason or another? They could be obsolescence, could be a denial of service. Somebody could have maliciously decided that no, you're not going to have access to the service. So where does obsolescence meet security?

**Chris Romeo (7:39)**: I think it's an architecture problem. In the Fisker's case, we can speak about them there in the article as the company that did this. Their reliance on the cloud/server-side software to even make the car move is just a terrible architecture. Like Tesla doesn't work like that. I don't have to have a connection to the mothership for Tesla can operate on its own. It does rely on a lot of services that it consumes over the internet, but it doesn't rely on the internet to start the car.

**Matt Coles (8:18)**: And that would, and I mean, let's just talk about car auto safety for a moment. That would be a huge safety issue if you relied on a reliable network connection and you're driving a vehicle. To even get the vehicle going. I mean, that's just insane.

**Chris Romeo (8:33)**: Who— in the middle of nowhere, like you're in the cornfields in Iowa and there's no cell service, your car stopped, you stop on the side of the road to examine a corn stalk because you're a city dweller, you've never seen one in live, and then you go to start your car and it's like, sorry.

**Izar Tarandach (8:49)**: But that's, that's the thing. Who thought that that was a good idea? From the article, we learned that the And the rationale behind this was that they wanted to be able to not only sell you services to your car, sell software upgrades to your car, but they wanted to, in case you sold the car to somebody else, resell those same services that the first owner already bought to the new owner.

**Chris Romeo (9:12)**: Sure.

**Matt Coles (9:13)**: Great idea. That sounds like a great value add.

**Chris Romeo (9:16)**: For them.

**Matt Coles (9:17)**: But don't make it the core way that the car functions.

**Izar Tarandach (9:21)**: So from a business standpoint, okay, that makes sense. Now, how do we make that work? The things that from the technical point of view, and that's where I thought that you were going with the, you need an API to turn on the car. Think about the attack surface that you are exposing here just by making those things dependent on a cloud service. All of a sudden you have an API that says apply brakes. Right? Or turn off engine or pull eject lever.

**Matt Coles (9:53)**: Okay. Yeah.

**Izar Tarandach (9:54)**: But nobody's—

**Matt Coles (9:55)**: Let me blow your mind a little bit more. Think about the architecture there. Does the car reach out to the cloud and get its instructions? Or does the cloud— does the car have an open port and an internet access? Is it on the DMZ or is it not on the DMZ?

**Izar Tarandach (10:12)**: Logic says that the car should be reaching back. Reality says the car has many open ports.

**Matt Coles (10:20)**: So here's a question. Could you run a Shodan scan and find these cars on the internet?

**Izar Tarandach (10:29)**: Most probably, yes.

**Chris Romeo (10:31)**: Interesting question.

**Izar Tarandach (10:32)**: Unless somebody went and said, okay, let's do our own cloud in here and keep all the— well, we know how those things work, right? They work wonderfully most of the times. So no.

**Matt Coles (10:44)**: So that's me. So no. So yeah. So I mean, so there's architectural decisions that can be made and it will be interesting. I mean, it would be interesting to know who really was in charge of making those decisions. 'Cause the software developer may be like, well, you know, I need these cloud services. We don't have compute power on board. Although granted cars have like a almost supercomputers these days. And so, who made that choice? Who said—

**Izar Tarandach (11:15)**: Product owner.

**Matt Coles (11:19)**: So what you're saying is they didn't have a security champion who is guiding the experience though.

**Izar Tarandach (11:24)**: Not only that, it sounds like something that somebody in the marketing department came up with and said, hmm, that's a great idea. Let me go sell it to the product owner. And the product owner said, hey, you know what? That's a great idea. Let me go and talk to the architects. And the architects say, hey, you know what? That's a terrible idea. But you know, Since you pay my salary, sure, why not? Whatever. I'm just going to buy Tesla.

**Matt Coles (11:44)**: So weak-willed architects put us in this position. Is that what you're saying?

**Izar Tarandach (11:51)**: I think that more than ever, I'm trying to push the view that at a certain level in the lifecycle of a story, somebody, somebody's got to act as a firewall and a champion for Good sense. Somebody has to put a stop to the madness and say, up to here.

**Chris Romeo (12:13)**: Well, I think that's represented in this equation if we're talking vehicle safety by, is it the transportation? I don't know what the government entity is.

**Matt Coles (12:23)**: NTSB or—

**Chris Romeo (12:25)**: NTSB. There's a host of others. Yeah. And something transportation.

**Matt Coles (12:28)**: Yeah, NHTSA or TSA or whatever it is.

**Chris Romeo (12:30)**: And so they have— If, for example, with self-driving, you can't just, in the United States, you can't just start a car company, build a self-driving feature, and let it loose on the road. Like, you have to go through various protocol safety testing and everything to prove that the things aren't just going to crash into each other or things at random.

**Matt Coles (12:53)**: So that's the— Just to have a car on the road, you have to go through that. I mean, whether you're doing self-driving or not.

**Chris Romeo (12:57)**: But in this case, what we're discussing is more of an attack surface that is opened, you know, apply brakes is not likely gonna be a control, a command that is sent from a cloud service back to a car unless it's in a self-driving mode or it, or it has a safety detection of some, it thinks a collision is about to occur.

**Matt Coles (13:18)**: Uh, recall, recall 10 years ago, not even actually, uh, more like 6 years ago. Uh, do you remember the OnStar hack. Yes, they were able to apply brake and kill engine on a vehicle driving on the road through open APIs. Well, not open, but APIs that were made available.

**Chris Romeo (13:38)**: Yeah.

**Matt Coles (13:39)**: Right. So it's, it's not just, I mean, it happens. We know it happens because it's been tested. Um, yeah.

**Izar Tarandach (13:48)**: I'm sorry, but I can already see it. Somebody thinking, We know where the outdoors are. We know where the outdoors, like the sides. We know where the car is. If I want to maximize exposure of the driver to the outdoor, I'm going to slow down the car when it drives by that specific one. So yeah, apply brakes.

**Matt Coles (14:16)**: So what you're saying is using, using using the environment, the ambient environment to enhance driver experience and giving a better value to the vehicle occupants.

**Izar Tarandach (14:29)**: And surely putting another LLM in the loop somewhere.

**Matt Coles (14:31)**: Sure.

**Izar Tarandach (14:32)**: There must be at least one.

**Chris Romeo (14:33)**: So let's come back around to obsolescence because that's where we started and we teased that we were going here. And so in the case of The Car Company with this article, that was a forced obsolescence because they went out of business. The car was reliant on the internet, the cloud services for operation.

**Matt Coles (14:52)**: It was unplanned.

**Izar Tarandach (14:53)**: I think it was forced and unplanned. It was forced by, by outside factors, and it was unplanned from the side of the company.

**Matt Coles (15:02)**: And it was a bad idea to begin with.

**Chris Romeo (15:04)**: Well, it was planned. It was planned.

**Izar Tarandach (15:07)**: You don't—

**Chris Romeo (15:07)**: bankruptcy doesn't just hit you.

**Izar Tarandach (15:09)**: Oh yeah, right.

**Chris Romeo (15:10)**: You have to declare it. You have— it takes you— you spend in a corporate environment, you're spending 3 months preparing for the event.

**Matt Coles (15:19)**: But at some point they forgot about the servers that control the cars. I don't think they forgot about it.

**Chris Romeo (15:24)**: Here's the interesting stuff. You don't— you mean you think— let's, let's play with that for a second though. Like, you think executives at a car company that's relying on cloud services didn't know that the cars were going to shut off?

**Izar Tarandach (15:36)**: You know what, I'm— abstain, thanks.

**Matt Coles (15:39)**: Yeah, that's not out of the— that's not out of the realm of reason.

**Chris Romeo (15:44)**: No.

**Izar Tarandach (15:44)**: That is part of my threat model, yeah. But the thing is—

**Matt Coles (15:50)**: You did what?

**Izar Tarandach (15:52)**: We had examples of hardware in the past where they had proprietary servers, but still the protocol was known. And when those servers shut down, the open source community did step forward and create its own open versions.

**Matt Coles (16:11)**: Yeah.

**Izar Tarandach (16:11)**: Now, I would imagine that something that's under so much scrutiny as a car on the road, somebody somewhere would expect those protocols to be, if not public, at least not proprietary. Somebody would have a copy of that thing. And upon failure of the servers, they could be published or made known to other people who are interested in keeping that thing running open source.

**Matt Coles (16:37)**: Although there's the danger of that, a good issue.

**Izar Tarandach (16:41)**: But, but nobody needs to do that. It's not a requirement. It's not part of the, if you want to do this, then do that. If you want to create this new thing, then put on escrow your protocol so that in case of emergency, somebody can step in and, and fill those shoes.

**Matt Coles (16:56)**: Right now, the danger of having your, your protocol in escrow is, well, what's, what stops a threat actor from standing up a rogue server?

**Izar Tarandach (17:06)**: That's the thing in escrow, right? Of course, it would fall. I want to really, really, really believe, really, really believe that the authentication between these cars and the original servers was well made and nobody would be able to stand that malicious server very easily. I really want to believe, like I'm thinking, I'm knocking my heels here many, many times.

**Chris Romeo (17:30)**: And you guys are being way too cynical here. The world is not the world. The world has been, there is some security goodness that exists in people's designs.

**Matt Coles (17:40)**: I mean, yes, there's a lot of crap out there, but if I'm designing a car from scratch, like, be quiet before I get in a lot of trouble. You know, don't rely on military-grade encryption here. If you get triple DES, you're lucky.

**Chris Romeo (17:55)**: I'm not suggesting, I mean, AES is not, in my mind, military-grade encryption. right?

**Matt Coles (18:00)**: Like, well, well, it can, it can be if you use a strong key and you have good key management practices and you're baking it into a hard root of trust and you're doing a host of other things that have not really been done well in the, in the automotive industry for a long time. Go take a look at the research papers by, you know, NCC Group and others over, over the years.

**Chris Romeo (18:20)**: This is, those are, this is not a new problem now though. They are, but they're, NCC Group was 10 years ago.

**Matt Coles (18:27)**: If I remember correctly.

**Izar Tarandach (18:27)**: And yet in some things, the state of the art has not moved too far forward.

**Matt Coles (18:32)**: Yeah. I mean, there's now an automotive and transportation hacking village at DEF CON.

**Izar Tarandach (18:38)**: True.

**Matt Coles (18:39)**: And they're still finding tons of issues.

**Chris Romeo (18:41)**: I mean, we're finding tons of issues in everything.

**Matt Coles (18:43)**: Yeah.

**Chris Romeo (18:44)**: Right. Because there are tons of issues. There's no mature category. You can't point at a category and say, well, this one's doesn't have a village at a security conference in—

**Izar Tarandach (18:55)**: Let me keep believing that aviation is a different level. Please, please let me keep believing that.

**Chris Romeo (19:02)**: They have an aviation village, don't they?

**Matt Coles (19:03)**: They do have an aviation village, yes. Because you remember the hack on the infotainment system in the—

**Izar Tarandach (19:09)**: Yeah, but it's very high level.

**Chris Romeo (19:12)**: He knows that. I really need that sound effect of the drums. Thursday, folks. 2-drink minimum. Remember to tip your—

**Matt Coles (19:24)**: we're here all week. Yes.

**Chris Romeo (19:27)**: Uh, security comedy. All right, so what do we— what do we do with this?

**Izar Tarandach (19:31)**: So wait, wait, wait, let's feel more confused.

**Chris Romeo (19:33)**: I feel more confused than when we started about this.

**Izar Tarandach (19:35)**: No, no, let's close the loop there.

**Matt Coles (19:37)**: Sure.

**Izar Tarandach (19:38)**: We agree that we have 3 kinds of obsolescence. 2 and a half, actually. Planned, unplanned, and forced.

**Matt Coles (19:45)**: Okay. And force could be planned or unplanned. Exactly.

**Izar Tarandach (19:49)**: And we agree that there are ways around, if you want to do something as smart as these companies have done. We agree that there are ways to get around it. We agree that there are, if you build a community of users, somebody in there knows how to write enough code that they might be able to stand an open source version of the thing. And that you have ways of protecting yourself, your IP, until that event, planned, unplanned, or forced happens, that would allow the thing to continue working. So you would be respecting, not protecting, but respecting the rights of your consumers. The problem is that nobody is actually building with this in mind, planning these things in mind, unplanning this in mind. And the government doesn't seem to be stepping into the thing.

**Matt Coles (20:40)**: You know, you, you're opening up another can of worms here though, because respecting the rights of their consumers, I wonder, has anyone looked at the EULA and said, oh, oh, the, the, the privacy value of this thing, I won't even start talking about it. No, no, no, I'm not even talking privacy. I'm talking about when you, when you buy software these days, you're not buying software. You're buying the license to use the software. and a EULA governs how you use that software. If you buy this car and you get a EULA and that EULA says, for the duration of the, of the time that you own the vehicle or the duration of the time we believe the vehicle is practically functional, you can have access and use the software. When we say, when we say otherwise, you lose access to the software.

**Chris Romeo (21:28)**: Matt is the only person I know who's read the EULA for anything.

**Matt Coles (21:31)**: Yeah. I have read EULAs. Yeah.

**Izar Tarandach (21:32)**: So just, just, just from, just from listening to that, I am already half asleep.

**Matt Coles (21:38)**: That's perfect. That's exactly what the product owner wants.

**Izar Tarandach (21:41)**: That's what I think. I think that there should be studies of the effect of reading the EULA and getting a good night's sleep.

**Matt Coles (21:49)**: You know, there is a system that does judge privacy policies and EULAs on reading comprehension levels.

**Chris Romeo (21:58)**: Yeah, me.

**Izar Tarandach (22:01)**: I can tell you I understand this or I don't understand this.

**Matt Coles (22:05)**: Yeah. So the Mozilla site Privacy Not Included has a, uh, has a great rating, rating comprehension system for, uh, for agreements. And, uh, I think they do it primarily for privacy, but I'm sure they could apply it to other things. You know, they don't expect anybody to read it, but also it's dense legalese. You're not gonna be able to read it in most cases.

**Chris Romeo (22:26)**: If we categorize the threats that exist for these 3 categories, Or maybe they, maybe some of them cross all the boundaries. When I start to think through this, I think, okay, uh, I buy a device, it's made obsolete by the vendor. They say they're not going to provide any more security updates or patches or anything. And so threat, one threat is that now I'm vulnerable to new vulnerabilities that are found.

**Matt Coles (22:57)**: So first off, you're, You're talking about the consumer doing the threat model, not the, not the vehicle manufacturer.

**Izar Tarandach (23:01)**: Yeah, I think that there are two threat models in here.

**Chris Romeo (23:03)**: Right.

**Matt Coles (23:04)**: But also not having software updates is different from having the device in question stop functioning properly.

**Chris Romeo (23:12)**: True. Yeah. True. Well, so those are different. There's, so that's a nuance in obsolescence. Does the device get bricked in the Fisker's case or does it continue to work? In a diminished, well, not even a diminished state.

**Matt Coles (23:26)**: Well, not even diminished, just whatever it was at the last update, right? So Chromebooks is an example, right? Google has planned obsolescence, not obsolescence per se, but the device doesn't stop working when it doesn't get its final software update.

**Izar Tarandach (23:37)**: It just doesn't. No, they just fired the team that forced it something.

**Matt Coles (23:44)**: So bad. So bad. But so real. You get less function. You get less functionality 'cause you get whatever functionality there was and then you get nothing beyond that point, right? And, but it doesn't stop working, although you in theory need to stop using it 'cause you may not get security updates. So now you're vulnerable to whatever's new coming out, right?

**Izar Tarandach (24:05)**: Oh, oh, oh, nice. So that's another category. That's the customer planned obsolescence claim. I'm going to stop using something that I bought because there is no security support to it. Therefore, I should consider it, uh, But whatever. And, uh, I shouldn't be using it anymore.

**Matt Coles (24:23)**: And that's why you would go to, you know, you would jump, you would jump, uh, iterations on your iPhone or your Android device. 'Cause you know, you're using a Pixel 3 XL that hasn't been getting software updates for 5 years and like, oh, uh, yeah, I can't, there's now this disastrous malware attack. I'm gonna upgrade so I get protection against it.

**Izar Tarandach (24:40)**: Yep.

**Chris Romeo (24:41)**: And there's always the, the, the nuance of the OpenSSL Heartbleed vulnerability that I always found hilarious. People were running versions that were so old that it didn't have the Heartbleed vulnerability in it. Like, there's a— there, that's a security policy right there, is just run software that's ancient. And there's no— none of the, you know, most of the attackers that are, that are around today weren't born yet when that software came out.

**Izar Tarandach (25:09)**: Can't break 3DS if you can only do one round.

**Chris Romeo (25:12)**: That's true.

**Izar Tarandach (25:19)**: But I like where Matt went with the fact that actually we are talking about two different threat models here. There's a threat model of the maker who should take into consideration obsolescence, forced and enforced, and forced, planned, unplanned, and forced, and of the consumer who before they make a decision to actually buy it should put into their rough model, what if this thing becomes, well, what's the, the obsolescence, obsolescence, obsolescence. Thanks.

**Matt Coles (25:49)**: So, um, well, or, or if the main, if the manufacturer or the service provider goes out of business, I mean, if you're looking at AWS, why aren't you doing the exact same thing, right? AWS goes down on occasion. What do you do when your business relies on it?

**Chris Romeo (26:03)**: Right?

**Matt Coles (26:04)**: I mean, it's extending that outside of obsolescence because obviously AWS isn't going obsolete.

**Izar Tarandach (26:08)**: If AWS goes obsolete, I'm going to say that I have bigger problems than my threat model.

**Chris Romeo (26:15)**: But that's, that's why people go multi-cloud, right? People, people do, you know, big companies run multi-cloud sites that have the ability to fail over to a different cloud provider.

**Matt Coles (26:25)**: Right.

**Chris Romeo (26:26)**: It's not uncommon.

**Matt Coles (26:27)**: It's, it's, uh, it's on, it's on. It's not unexpected is what you're saying.

**Chris Romeo (26:30)**: Yeah. I mean, it's a business continuity play, right? And so to match—

**Matt Coles (26:34)**: But when you buy a car, but when you buy a car, do you expect to need to do business continuity?

**Chris Romeo (26:39)**: You guys don't buy two cars at the same time from different providers?

**Izar Tarandach (26:44)**: I do.

**Matt Coles (26:45)**: Yeah.

**Chris Romeo (26:46)**: I mean, that's what I do. Two or three.

**Izar Tarandach (26:49)**: And it's even worse. I don't buy, I lease.

**Matt Coles (26:52)**: I buy one from Matchbox and one from Hot Wheels.

**Chris Romeo (26:55)**: And, you know, it's even worse than a lease. I just take test drives and I don't come back.

**Matt Coles (27:03)**: Well, that's the sign and drive.

**Chris Romeo (27:06)**: I can sign and drive. No one said anything about pay. Just sign and drive.

**Izar Tarandach (27:11)**: Well, there is the Lincoln Lawyer, and now there's the Tesla security practitioner.

**Matt Coles (27:17)**: That's right.

**Chris Romeo (27:18)**: All right, with that note, we're out of time for today. Thanks, folks, for listening to the Security Table. I guess I don't want to say stay obsolete. Stay non-obsolete.

**Izar Tarandach (27:28)**: No, it will be here until we become obsolete.

**Chris Romeo (27:32)**: That's right. That's our new motto at the Security Table. We'll be here until we become obsolete. You can judge.
