 Hello, and welcome to this episode of the Security Angle. I'm Shelly Kramer, managing director and principal analyst here at theCUBE Research, and today for this episode of the Security Angle, I'm joined by Joe Peterson, fellow analyst and member of our CUBE community family of analysts, and I'm joined by David Lenticum, a new member to our analyst team here at theCUBE Research. Hello, friends. It's great to see you. Hello, Shelly. Hi, Dave. Hi. It's great to be here. Absolutely. Absolutely. And we are so excited to have you. David, I know a lot of your focus is on cloud and AI and all things in between, so we are thrilled to have you on this week's episode of the Security Angle, and we are talking today about the transformative nature of cloud and AI together. And this is a topic that I am sure you have many thoughts on, David. Absolutely. Although, I kind of think you have many thoughts on a variety of topics. So I want to start with, you know, I've read about this phenomenon called the Fujiwara effect, and it's what happens when two tropical cyclones get close enough together to create sort of a shared center, and this forces the two storms to kind of whip around that shared center, and then if one storm is stronger than the other, it usually leads to the weaker one sort of being gobbled up by the more powerful one. I think there's many situations in life actually where we have the Fujiwara effect, but anyway, I think this is kind of happening with AI and cloud, and more specifically, I think that we're seeing AI actually accelerate the adoption of cloud for a variety of reasons, and of course it enhances the platforms that, you know, the hyperscalers platforms. So David, what's your take on that? It's a boom for cloud. Everybody's in a bit of a confused state right now. They don't know, you know, is it going to be on premise? Is it going to happen in the cloud? Things like that. The reality is that the cloud is the most convenient place to build AI systems because the ecosystem is already there. You can choose the engines you need, the tool sets you need, you can choose the databases that you need, you can choose any number of systems. We have about 40% of our enterprise data that's already on the cloud, so we can leverage that as training data, and it's just a matter of sitting at your terminal in a godlike way and connecting the dots between these various things, and it automatically happens behind the scenes. And so this is versus walking to a data center and unbolting stuff from a rack and, you know, bolting new stuff in. So as long as we can get by the money that's going to be thrown at this, then I think we're going to be okay. But right now, if you look at cloud computing, we saw the numbers come back last few weeks in terms of their quarter earnings and its way up and AI is all over that, and that's going to continue on for years. Yeah, it's really interesting. I will say some of the conversations that I had this last week in Barcelona at Mobile World Congress around AI wasn't that cloud wasn't important and wasn't playing a key role, but there are many clients today who actually are taking sort of those initial steps in terms of embracing AI and gen AI and that sort of thing, and they're doing it on-prem and kind of resisting moving to the cloud at these early days. And I think maybe my estimation on that is that, you know, there's still so many unknowns. People are a little bit nervous about using the cloud. What do you think about that? I think that on-prem is going to rise as well. You know, in sizing these systems up, what it's going to cost on the cloud versus sometimes these on-premise systems, even some of the micro-clouds that are starting to emerge as well, and that's kind of Plan C and some of the managed service providers as well, many enterprises are choosing to put their initial AI built systems in the data centers. And I think that's perfectly fine. As an architect, I don't care where you put it. We have to put it in the place that's going to bring the most optimized value back to the business. And sometimes that can be on the cloud, certainly if we need the ecosystem around the AI systems. Sometimes it's going to be an isolated system. It's going to do a very discrete thing. It's going to do the same thing over and over again. On-premise is preferred because it's going to be cheaper. And if you notice, the last 10 years, hardware prices are limited. Yeah. Yeah, I know. Well, and I think there are lots of conversations around AI is wonderful and it's not cheap to make it happen. But I was interested to hear so many vendors saying that they are taking some initial steps. They are very excited about how they can use AI throughout their organizations, but many of them are taking those initial steps on-prem. So I think that makes perfect sense. So the cloud, as you mentioned, we've just come off a big earnings week. The global cloud computing market continues to boom. Let's see. It's expected to reach about a $680 million market in US dollars. That's a 30% jump in the space of about two years, again, to your point, that we're seeing lots and lots of cloud adoption. So do you think that one of the things that I think that we're seeing here is there's a build versus buy mentality at play. And to a certain extent, some of the functionality, I think that people want to be able to use some of the AI-powered functionality already exists in some of the hyperscaler environments. And so do you think that some of this is why we're seeing an increase in cloud? Do you think that's attractive to people that, oh, if we want to use virtual agents and chatbots and things like that, we can use the AWS ecosystem, we can use Azure, we have some things that are built in. So do you think that's part of this as well? That's the most compelling reason to use cloud for AI. If you think about it, because all the tools are there, it's you measure the number of services and cloud providers in the thousands, not the hundreds anymore. And in terms of AI services, go to the console and type in AI systems and you'll have hundreds of different options for different tool sets. They provide training data, data hygiene tools, things like that that are needed for AI. And so we don't have to learn and integrate all these various disparate tools. They're there already for us to leverage in any way we want. And so that's why the cloud is chosen. It's the most convenient way to do it. You can certainly do it on premise and sometimes cheaper. But if people are trying to get something built fast and quick, and they're trying to get all the tool sets there, they're trying to get the ecosystems along with it, they're going to try to build it on a single public cloud provider and it can't blame them because you're getting into another set of complexity. Yeah, no, I agree. It makes perfect sense. Oh, go ahead. No, no, no, I was good. Go right ahead. Oh, okay. So they whiz in their state of AI and the cloud report came out and for this year and it showed that more than 70% of the organizations that they surveyed were looking at managed AI and the adoption rate, why both that of managed Kubernetes? And I guess that's a surprise, but not a surprise. Why do you think that the managed AI rate was so high? Yeah, it's making it somebody else's problem, which I would make the same sort of argument. I mean, the great idea behind managed databases and managed Kubernetes and managed cloud native systems and managed AI is you put it in the hands of people who do it many times for different clients. And so therefore, I don't have to hire the very expensive data scientists for expensive AI engineers and people like myself to knit it all together. I just say, hey, make this so and they go ahead and build the AI system that they need and they also maintain it, which is the harder thing to do. Because you got to remember, you're moving a lot of data, you're doing a lot of data hygiene exercises, all kinds of things that are very problematic in operations. And I think that's why we're seeing managed services and including AI as a popular option. As a business decision maker, I have to say it makes perfect sense to me. I mean, I don't want to get necessarily in the weeds. And when I'm making a buying decision like this, I know what kind of resources I have internally. I know how hard it is to hire highly skilled talent. I know we're at the beginning stages, we're dipping our toe here, we're experimenting in some use cases, that sort of thing. I mean, I want somebody to take me by the hand and do that heavy lifting all day long. And I think it makes perfect sense that we're seeing a rise in this because this is kind of uncharted territory. And if I can work alongside a vendor partner who's going to help make this process easy, this is a no brainer for me. Absolutely. You're going to have to hire some people you've never had around before, like an ethics specialist and a data scientist and an AI engineer. And they're very scarce. The better ones out there are all hired. And if you're not willing to go into it 100% of the way, in other words, get the talent you need to make you successful, then I wouldn't attempt, I wouldn't attempt, recommend that you attempt to do it. I recommend that you turn it over to somebody else who has to know how. Well, and the reality though, is when you work with, you know, I don't really care, I'm a huge fan of partnering and the benefits that that brings. And I think that's, you know, we're seeing this across the whole digital transformation landscape these days, right? Smart organizations learn to partner with one another and collaborate. And but I think that, you know, when it comes to this, this whole, you know, you talked about ethical AI specialists and risk mitigation as it relates to AI. And there's so much expertise that AI brings into this equation that many of us don't have on our staffs right now. And so to me, being able to again, be beating the work with a trusted vendor partner, you know, when you can work with somebody that you know is bringing their vast expertise and working with other clients who are resolving the very same issues, there's no learning curve, there's no nothing, you know, and again, a lot of us are, you know, how many clients have you ever worked with that said, Oh, take as much time as you want to get this rolled out, we're patient, you know, everybody wants to be able to have this rolled out to be able to get adoption to be able to start experimenting to be able to start seeing some use cases and how can we get to ROI here and everything else. So it just makes perfect sense to me that this is a very popular thing to outsource. Yeah, right now we're at the introductory stages of it. I think people are just kind of feeling their way into what it is. When I look at what people are doing, they're learning, they're not implementing as much yet. And so we're doing a lot of prototypes and a lot of smaller systems and, you know, leveraging, leveraging LLMs as a service and all those sorts of things are in play when they finally get into it. And they realize they're going to need about 25% more talent than they would in just deploying a regular cloud system. And it's going to cost them twice as much, let's set the metrics what I see so far. So if you're unwilling to go all the way and to get the resources you need, my recommendation is to outsource what you can. Just make sure you're outsourcing, you just hit the nail on the head, a trusted partner, someone who's going to lead you down the right, the right direction. And whether it's a consulting firm or it's a, you know, vendor led organization, whatever, it's just you got to find people who know how to do it. The biggest concern to me is I don't think there's a lot of people out there that know how to build these things. And so I suspect we're going to hear a lot about a lot of field projects where they did outsource to what they thought was a trusted partner. They were led in the wrong direction and it didn't work out. And we saw that in the early stages of cloud. Suddenly everybody was a cloud expert. And I never saw them at the meetings. And they led them astray and so we had to fix a lot of things back in, you know, 2000, you know, 2013, as where people were in the cloud systems. And we're going to see the same thing here. Probably it's going to be more, it's going to be more damaging because a lot of organizations can benefit from this technology and providing the technology, allowing the technology to be an innovative durator for them and in their marketplace and build something that's really going to change their presence in the market. And they're going to miss that, that could damage the business completely. I was going to say, Dave, to your point, you know, I started architecting cloud in 2009 and people stuck their toe in and they did things like object storage. Remember that, you know, they were just doing really, they weren't even doing full systems or stacks. They were just taking a piece of the puzzle and doing that work. And the other thing, just a point you made earlier, I think it was Shelley about ROI. I don't know if you guys have checked the numbers, but the hardware and the software associated with AI is exponential in comparison to just a normal server build or a normal, right? It's so much more. So to make the numbers work, you have to get it right the first time. Yeah. Yeah. Maybe GPUs are cheap. Oh, wait, no, not. And all those, all the equipment is very specialized. And if you're not willing to put together the system that's going to provide you with the basic performance that you need, then it's going to be very difficult for you to do. And you have to have a special engineering talent to pick the processors that are aligned directly to the AI systems that you're building. You know, the reality of it is I was listening to you talk about, you know, back in the cloud days and people made the mistake of working with the wrong vendors, you know, the reality of it is that that's not a new phenomenon, right? I mean, you know, I can't tell you how many, you know, in my career as a marketing brand strategist, you know, how many times I came in after somebody else who had sold an unsuspecting client things that they absolutely didn't need. They hadn't, you know what I'm saying, like cleaning up after other people's messes is something that we've, all of us have had to do in some way or another, right? So I think this is just another mess that somebody could, you know, potentially find themselves in. But I think to both of your points, you know, one is that I would hope that it would be fairly easy to figure out quickly if you're going down the wrong path. Not 100% sure about that. But the reality of it is their time is to your point, Joe, very much of the essence in there are things that you can miss out on if you try to go it your own way or if you choose the wrong vendor partner or whatever. So do your homework on that front, right friends? Absolutely. I think we're going to see a lot of failures and successes that are directly related to their ability to weaponize AI for the purpose of their businesses. Yeah. Yeah. You know, one thing I thought was interesting, Joe, I think you mentioned the whiz research, you know, while we're having these conversations and these are timely and important conversations, the reality of it is, is that, you know, many organizations are experimenting with AI. But the reality is that, you know, only a handful are really what we would call, you know, power users. And so some of this research showed that, you know, 32% of organizations appear to be in the experimentation phase with AI tools, totally makes sense. Deploying fewer than 10 instances of AI services in their cloud environments. And only about 10% of those have moved beyond experimentation and they're deploying 50 or more instances in their environment. So, you know, we're at those early sort of dipping your toe in the water stage for people. So, you know, it'll really be interesting to watch this play out. That's for sure. Yeah. Dave, I, you know, it's, it's a heated battle right now. When it comes to the hyperscalers in AI. And, yeah, everybody's throwing up news as fast as they can. It seems like, it's like, ooh, something else came out today. Okay. So, it was interesting in that whiz report back to that. Azure AI, which includes Azure Open AI is leading the way with 39% of the organizations using the cert service. And, right? So, if you looked at a four month slice of utilization in 23, the number of deployments rose 228%. Simple question. Why so many Azure bands? That's an easy one. Sorry. Go ahead. That's all right. I think that Microsoft saw the generative AI bubble inflating before everybody else. And so, they were the first to the party. They bought open AI, spent, you know, put a billion dollars where their mouth was, or many billions of dollars where their mouth was. And I think that's paying dividends to them right now. So, eventually, the industry will normalize, they're all moving at a pretty good pace. And, you know, what's true now in 2024 won't be true in 2025 and 2026. It's just going to be a big battle royale. It's going to be very much like the cloud in 2010. Yeah. Well, and the reality of it is, you know, AI is not new for hyperscalers, right? I mean, many of them have been using AI, we've been using AI, all of that sort of thing. But what Microsoft did is they capitalized on an important moment in time. And I think, and I think what it has resulted in is that when people think about who's doing amazing things with AI, Microsoft comes to mind because of the open AI alliance. And yeah, they were just really smart there. You got to hand it to them. In other words, they saw the value of generative AI early on. I think you got to remember, it hasn't been around for a while as far as a hype driven concept. We've known about it for a while. Certainly dealing with AI and machine learning for many years now. And we understand the ability how to operationalize and how to add it to the business. But this new dimension brought a whole new group of interested parties into this world and trying to get their generative AI systems up and running. And those are the people who are spending the money right now. And many people I talked to, they thought AI was invented two years ago with chat GPT. And so I was like, no, it's been around since the 50s. Oh, yeah. And so with that kind of mentality, people are just grabbing a market that really isn't educated yet. Yeah, absolutely. Well, you know, it isn't educated and everybody, you know, feels the pressure. I have to be doing that. I don't even know what it is, but I have to be doing it because if I'm not, you know, I mean, I think they're very much is a low factor at play here. You know, David, you mentioned money, let's talk about money for a minute. Now, you know, without question, the iber scalars have already spent hundreds of billions on AI innovation. But you know, the financial pressure that comes with the demand for AI is putting massive pressure on cloud infrastructure itself, you know, so the capital expenditures needed to expand and retrofit data centers and deploy GPUs and TPUs and all of that. You know, so have we already begun seeing the impact of some of that investment in earnings reports? Where do you think? Where do you think we're seeing that? Or are we seeing that yet? Yeah, I don't think we've seen it yet. I don't think they've spent the money, the big tranche of money that they need to build GPU based and HPC based systems to support AI. And so they're just doing things on regular CPUs and some of the GPU instances that they have. So we're running this stuff on traditional infrastructure at the end of the day, they're not buying GPU based systems to do that. And I think in many instances, that's going to be perfectly fine for the use cases that we're seeing now. But moving forward, they're going to have to spend a tremendous amount of money in putting very expensive, very power hungry processors into the various data centers to support the loads that are looking support. And that's going to be interesting because number one, they have a traditional business that they need to support the people who are leveraging cloud based storage and cloud based compute and things like that. And that's what people are running their businesses on these days. Are they going to divert resources from that to focus on the AI space? And is there going to be a downside that's going to happen on the traditional business side as they push more investment into the AI side? Those are the questions I've been getting a lot of recently. They see the investments being made and they say, you know, Dave, we put all we're all in on AWS are all in and zero are all in and Google are all in and all three. And we're concerned about them not building or maintaining the core systems that we need to run our business. Because guess what? AI is five years away from us. We don't really care about it. We just hear about running traditional infrastructure. That's a bit of a balancing act. I think that the cloud providers are going to have to figure out how to do. Yeah, it is definitely an interesting challenge for them. So what about security? Joe and I talk a lot about AI and security. And I know you think about it and talk about it a lot as well. So, you know, in December of 2023, Amazon, Google, Microsoft open AI joined forces to with the cloud security alliance on an AI safety initiative. And the whole goal was to create an AI framework that could be released within about a year. I've got two questions for you both here. So are AI systems subjective to normal security vulnerabilities that need to be considered alongside standard cybersecurity threats? Personally, I believe the answer to that is yes. Joe, I'm going to do, you know, you've been quiet. So talk to us a little bit about that in your thoughts. Yeah, they are. I mean, if you look at the traditional tools in the market, first of all, they're text-based, many of them. They're not object-based. And AI introduces the element of object into the scenario. So what we're starting to see is some of the tools that exist on the market being retrofitted to facilitate AI data. For example, DLP, you and I've talked about that. DLP is a tool that there are five or six vendors that have gone and retrofitted their tooling, right? And you're going to see more of that. The second thing to consider is, and I think it's a bigger problem, we as companies, you know, organizations can put out initiatives and got guardrails to their employees about acceptable behavior. But how do we police that behavior? And that becomes a problem, especially when we think about how people use and do their work. So we've got private AI and we've got public AI. And how do we put in those guardrails and make them effective when those systems start to combine? And that's what I think about. Yeah. There's a whole new realm of security vulnerabilities that come up with AI. And the big thing would be poisoning on the models, that people have a tendency not to think about. So in other words, we have the ability to deal with the authentication of the systems, but people who have access to the models can do a lot of very creative things, you know, like make them pay out, you know, millions of dollars to their SWIFT bank account on Tuesdays. And those things are very difficult to detect. So we're going to have to put parameters around that. Joe hit the nail on the head. We got to put education in place, so people are using these things correctly. And also remember, we got lots of data flying around. So excuse me, how are we going to deal with the data? And securing it, make sure it's secured in flight and make sure it's at rest. And I don't think people really get that now. They assume that the different security data standards are going to transfer to the AI systems. Sometimes that's not going to be the case. And the way you're protecting these models has to be the same in terms of the way you're protecting data. Also remember that the data can change state within AI systems. In other words, we can take anonymized data and load it into an LLM. It can figure out who that data belongs to. And suddenly you have BII information and suddenly you're violating law. And so lots of different dimensions that I have these conversations all the time. I don't think people really kind of understand the different dimensions, even people who are security experts haven't really understood the changes that are needed that are going to come up and around. Also weaponizing AI, so they're attacking systems. We've already seen that already. And that's the bad news. The good news is we can weaponize security systems using AI-based systems to become more proactive and the ability to spot attacks. And so the fun begins. Here we go. Spy versus fly. You know what? It was always a game of whack-a-mole, right? This whole cybersecurity business. But now it's like whack-a-mole on steroids. They've got their AI. We've got our AI. We're duking it out. Very interesting times, for sure. It's going to be, it's going to be, the next few years is going to be incredibly interesting because we're going to see some amazing breaches and attacks that are going to be caused by weaponizing AI-based systems. And we're going to see a lot of industries and a lot of enterprises that are caught unaware because they don't have these security parameters in place. And when they do the course mortem on why the attack occurred and they stole millions of dollars or did a ransomware attack or things like that, they're going to say, well, listen, look, we didn't understand the new vulnerabilities of these systems. We built them because we felt we needed to build them to be competitive in the industry that we're in. But we really didn't put the security parameters in place to really protect these things. And that should be something that's systemic to every step of the way. I always tell people who are doing architecture, including AI architecture, there should be security at the design phase. The minute you're dealing with the use cases, all the way to operations and all points in between, this is not something where we start bolting stuff on at the end state. And I think that's what everybody thinks is going to happen. Well, yeah, security is absolutely not an afterthought in this in this arena. And if you do, you're rolling some really dangerous dice, I think. Yes. So as we wrap the show, the last the last thing I want to talk about is just, so how far are we away from AI regulation in the United States that has even a tiny bit of teeth somewhat like what we see with the EU's AI Act? And I will say this, the EU has long been certainly ahead of the United States in terms of any kind of regulation around technology. So it is interesting to see sort of what they've put in place. And then we have our own AI Act that is relatively mild, I think a first step. And how far do you think we are here in the United States from having some AI regulation that has some teeth? I think we're pretty close because as soon as you realize the privacy vulnerabilities that we have that are exposed to us, that's going to be a call to action for the government to put some regulations in place to make sure there's some laws and some teeth and making sure we're not weaponizing this stuff against us in the marketplace. We're already concerned about privacy. We give away data, data is mine for about us all the time. But the ability to take that data and derive it in different domains and use it in different ways that are now possible with AI gets to a whole level of anti-privacy concerns that privacy concerns that we haven't understood yet. And I think once people are testifying in front of Congress in terms of what these capabilities are, we'll start seeing some bipartisan regulations starting to emerge, putting some limits and some restrictions on how we use it. It's very much like HIPAA. HIPAA grew up out of the fact that we were concerned about patient data and privacy and the data and dealing with that in HL7, things like that. This is going to basically run the same path, but it's going to be, has to move faster. They're going to have to get legislation through in the next few years or else the cats are going to be, the cats will be out of the back. Yeah. I think, Dave, that you're going to have, I'm halfway with you, Dave. I think you're going to have vertical specificity as it relates to the regulation. So I think regulations in healthcare, for example, like you pointed out with HIPAA, I think PCI retail will be another, but many of the disclosures that are required are largely state regulated. It is not driven by federal. So there are states that have lots of security regulations like California and then those states that don't have hardly any. So I don't know if I'm on the same page with you with this. You know, I think that we're, I think that something that we haven't talked about as part of this conversation in the whole idea of risk mitigation is that I think the onus in many ways needs to be on the brand to respect the power of AI and to respect what it is consumers are saying. You know, I think that in any, any bit of research that I've read on the topic of AI and Gen AI and, you know, questions posed to consumers, they want to know how their data is being used. They want to know when something, they're reading something that is Gen AI generated and that sort of thing. But you know, I'll give you one example. We have gotten very lazy in terms of terms and conditions, right? How many times do you sign up for something and you know, you're just your, you know, yeah, yeah, except, right? Because you want to get to whatever it is you want to get to. And so we're lazy as a body of human beings, most of us. But I was just came across something in the past couple of days, people being a little bit up in arms about docusigns, terms and conditions. So docusign, who's not familiar with docusign, right? You know, electronic signatures, we use them all the time. But in docusign terms and conditions, it talks about the ability for them to use anonymized data that they feed into their LLM for any of these documents that are signed using the docusign platform. And by the way, that's automatic, you have to opt out, I believe. So you don't opt in, you opt out. But what my point here is that I've seen over the course of the last couple of days, I don't know yet how up in arms people need to be about this particular situation with docusign. Is it a big deal? I've talked to a fellow analyst here and you know, he really didn't think it was that big of a deal. But I've seen lots of consumers being really pissed off about this. And they feel like they're kind of being taken advantage of and it leads to not trusting docusign and do we need to use something else? And so my point here is that I believe that the onus is on brands to really communicate to customers what kind of data we have, how we're using it, how we're using AI in there. Because protecting that customer loyalty and brand trust is an important part of the equation here. So I think that brands really need to step up on this front. You make it great. Go ahead. Go ahead, Joe. Well, I'm not a marketer, Shelley, and you are. And so let me ask you, is there enough juice for the squeeze here from the CMO's standpoint? Is there enough differentiation to put a stake in the ground and say, hey, here's what we're doing? Because they believe that it will matter to their customers. I think it is important. All of the confusion and the excitement and the trepidation, all of these words make sense when it comes to AI and Gen AI. It's exciting times, but it's scary times. And I think that we see some amazing possibilities, but we also see big concerns. And if you're a consumer, and we think about this and we're immersed in this space all day, every day, right? And we're talking with and we're advising clients and that sort of thing. So we sort of have an advantage. But for ordinary average people, there's still so much unknown here. And there's very much, there are very much things to be nervous about. So I feel like that's something that consumers are looking for brands to be honest about, to be transparent about. And here's how we keep you safe. You both make some amazing points. I think it's going to be, people are going to have to be burned, because it's amazing to me how much privacy we're willing to give up. And then it's, you know, as a security person, it's just beyond, it's beyond belief. And I think that ultimately, people are going to have to see some downsides of this. You got to remember, there is no anonymized data anymore with AI. I can take data and I can backfill it. And suddenly I can put attributes to it and make it PI information. I can assign my name to a contract that I signed. And it can go into some big database, which is going to be useful to marketers or people who are, you know, are, are even the FBI subpoenaing that data for whatever reason, all kinds of things are going to happen where we're going to find we're giving away a lot more than we think. And I think once people realize that, that's when the pitch works and those torches come out. I agree. I agree. It is interesting times for sure. You know, I was telling, I was having this conversation when I was in Barcelona this last week and, you know, there's some really cool things about being older. And, and what I mean is that, you know, all three of us are old enough that we lived through the dawn of the internet, right? And so we know what it's like to live a completely analog life. And we also remember advancements like, oh, my God, a fax machine. You know, I don't have to like go across terrible and erphetic something to you overnight, David or whatever. You know, and, and then, you know, the internet and I'll never forget, you know, when all of a sudden now we all have computers on our desks and we're communicating by email. And, you know, I remember running into somebody proudly saying, I don't use email. That's stupid. I'm never going to use email. I've got a secretary for that, you know, or whatever, you know, people being proud about it, not embracing the internet and all that. But, you know, we've seen mobile devices, right? I mean, do you remember the first mobile devices and they were car phones? Because the only place you used them was in your car. And we call them a car phone for that reason. But it's so interesting to have lived through all of this evolution of technology. It's now be here. You know, we're not really at the cusp. We're kind of well in, you know, beginning the journey, but seeing another, another faction of technology that's really going to change everything in some pretty amazing ways. It is. And I think we have to prepare ourselves for that. In other words, we know what the change can bring. It can bring lots of good things and lots of value to our lives, because lots of bad things as well. You know, too much screen time on the phone and things like that that we do that we do now as a bad habit. So we're going to have to put everything in perspective and start putting logical, logical, uh, uh, gateways around this stuff. So how we're using the technology and make sure it's good ethical use. It's used for a good purpose. We're doing good with it. And we're enhancing your business. We're trying to create innovative different shaders in the marketplace. All the sorts of things are possible now, but also all the bad things are possible. And I think both things are going to occur. We're going to see some amazing successes in the next few years, but some amazing, tragic things. Yeah. I agree as well. And you know what? It is a very interesting time to be alive. I'll say that. Well, Joe and David, I think that, uh, we have reached the end of our program today and this, I knew this was going to be a fascinating conversation. I know it's the first of many that we will be having on this topic, but thanks so much for joining me on the security angle today and we will see you again next time.