 Welcome back, everyone, to the show here. On location, theCUBE is in the press area this week for AWS annual user conference reinvent, of course, part of our special edition SuperCloud 5, where you can run out of our Palo Alto studio, Lisa Martin and Savannah Peterson, hosting, great to have you guys in the studio. But we're on location, we're getting all the data, talking to the top people, talking to the experts, CISOs, CIOs, all the executives on Amazon and more. We have CUBE alumni, Merit Bear here who's the field, CISO, Chief Information Security Officer for Lacework, great to have her back on, former Amazonian, CUBE alumni, back, great to see you. That's true. We've got Margaritas here and everything. This is Lemonade, I swear. Yeah, we promised to call it Lemonade, you know. Just like we do identity in the cloud, we always know what's in it. Yeah, no, really happy to be back. Thanks for having me. It's exciting. There's always so much energy in this room. And yeah, as you mentioned, this is my first reinvent as a Lacework employee, the field CISO and used to be in a small team in the office of the CISO at AWS, doing security of AWS. So it's always really interesting to see how customers are interacting. And what's great is that Lacework is a real big partner of AWS. So you have that kind of also connection. So you kind of know how the mothership works here at AWS. And also as Lacework. The mothership. Lacework. Yes, hopefully it doesn't call me back. No, it is interesting. And it's also interesting from a security perspective because so much of my work when I was at AWS was on securing the cloud itself, but really those customer conversations revolve around how they can secure themselves. And for that, they need something like a Lacework. So let's get into what the show is this week. I did a preview. Adam Squist gave me an exclusive and talked a lot about what they're going to do. Obviously chips and models together. So you have this new layer, infrastructure layer powering, this new model, abstraction layer, foundation models and more. So kind of a new glue layer is coming in this kind of middleware part. Hardware is going to be a big performing factor. All that goodness. So the question is, what is the security threats in the new gen AI? I mean, our large length, our foundation models themselves going to have like a firewall around them. Is it like an endpoint? It's not an endpoint, but like you got to look at the security posture. And of course, you know, Amazon's comments, oh, it's security intrinsic and everything that we do. Okay, I get that. But like, let's call it reality. Security is a huge concern as this gender of AI wave comes down the pike. Yeah, there's so much in what you just said that we could spend a PhD unpacking, literally, you know. So one, there's the hardware layers which, you know, have always been important and are today really critical. We've talked about AWS nature before and some of the confidential computing benefits that folks get from the like fact that AWS built it to not be human accessible. And so you don't have to like pay extra for that factor. And you know, this is part of a longer tail around the chip industry and other things. So it's important in that I think right now, of course, AI is a buzzword, but what we're really seeing is the culmination of a lot of technologies coming to bear. So yeah, it starts at those hardware layers. Of course, that's not something that customers need to toggle. You're just getting that if you're using any, you know, EC2 or Lambda or anything that's running on those compute layers in AWS. But what I think is really interesting as we're talking about the uses of AI, we're talking about, as you alluded to, you know, folks expecting that they might generate code, for example, well, to do that, they're using their own substantive content. They might be putting code in to get other inputs out or outputs out. There's all these ways in which you need to be using real data to get ML outputs. Meanwhile, you need to be protecting those algorithms themselves. And then, of course, the uses of AI for security are also really interesting. Lacework has been doing what I would consider like old school AI for a long time, which is to say that like, you know, we're looking at how folks are accessing what and in what patterns and manners. We're now actually doing cloud aware analysis, or sorry, code aware analysis earlier. And so folks who are using agents are able to get data around what they're actually running instead of just kind of a laundry list of CVEs that they have to look at. So, you know, we see this relevant, we see AI relevant in both how we're securing the processes themselves, how we're securing the outputs, and then how we're using AI for security. So one of the things that comes up is how do you regulate this? Or do you not regulate, are you an accelerator, an enthusiast, no, is it's, effective accelerationism is a big kind of persona now. And then there's the decels, slow it down. And so it's interesting, we have a cultural kind of war going on, only war, but like tension between open AI just showed us that you have a board that says, whoa, AGI, which I don't think is in the future anytime soon, but they're basically, we're making these decisions around AGI, the fear and slow it down, and then you get the 90 billion dollar valuation. Let's go make money. So you got the capitalism meeting, kind of like the utopia that could be generated. So the security are people that break pragmatics. So I want to get your thoughts on this because security is a huge part, you can't break it, it's got to work, it's got to be secure. It's got to run, yeah. So you can't have this like gray area of philosophy. What's the security industry's general principles around go fast, slow down? Is there a theme? I think that a lot of what we're seeing, I personally am like a security optimist, so I see security as part of how we're going to move forward in a productive way. I think that security around AI and also the security of your AI will be areas that we care about for the foreseeable future, but we're going to be doing stuff at an accelerated pace, with that high power compute, with the ability to do more and know when you're hitting a wire. You know what I mean? Perverbial wire, like a threshold. So being able to get real time alerting, being able to get, or near real time, sorry, the engineer and me, but being able to get really low latency alerting around things that look anomalous. It is still true that the most likely source of an attack is a valid credential that's being misused. So what we really want to be looking at as a security community is, what's lighting up around the heuristics, and there's so much here that like, the power of accelerated compute and so much data is, you know, everything in the cloud is an API call, so it's really beautiful. The way that we're able to triangulate around to get the right kinds of alerting, you still have to take action on that as a security team. And as you alluded, you know, a lot of this also revolves around, how companies work, how the industries work, and I think, although we're not perfect at it, we're having this reckoning. And I think that's important. We should be talking about the values we want to be living by. I think we should be defining those. Tech is something that is human constructed and like, who but we, to start to define some of these and get better at it. And I really, I also see possibilities here for underserved and vulnerable communities, including folks who, you know, may not be ones for whom security was traditionally a source of power to get more accessibility. I mean, I'm a fan of Andy Grove, the legendary Intel innovator who, you know, speed of Moore's Law, and that was back then fast, it's going to get faster now, the AI is. His philosophy was, let chaos reign and then reign in the chaos. Let it run a little bit. Well, of course, Moore's Law isn't an actual law, it's a presumption. But yeah, I want to see more, better, faster. And I also want to see us do it in a deliberate or a conscious way. And I think that's, we have all those pieces and I want to see us construct them in a thoughtful way. So let's get it, last time you were on the queue, we talked about security not being a cost center, being more part of the business proposition. Okay, now that business proposition is clear, the user experience is gen AI, this is going to be all layers of the stack. What are you seeing as you're on the field, talking to customers at Lacework around what they're thinking, what's the psychology, what's the mentality, what's the mindset, and what things are being put into practice that you see happening in the field right now? Honestly, what I consider to be my job is making sure that customers have those muscle groups and mechanisms to actually do the work of security. So I think that Lacework is delivering really impressive across the board capabilities, but you have to go have those muscle groups and that connective tissue to be able to take action, to go do stuff, to refine that flywheel and get better over time. And there's no approximating that, you have to just go, it looks like suspenders and hard work because it is. So as you get better as a security team, you may be asking, why didn't we notice this for two hours instead of just like, you know, there's an outage. But it is still one of those things that you get better at over time. And so what I talk to customer CISAs about mostly is what are you doing around incident response? What are you doing around threat detection today? What are you doing around identity and how do you notice when something is going off the rails? How quickly can you recover? You know, these are just trueisms and also, you know, as I've kind of talked about before, being a CISO is a fairly lonely job. They don't have a lot of room to talk to other folks about what are my peers doing or, you know, is this the kind of way that I should be approaching this? And I think one of the ways that I love to talk to fellow CISOs is to understand, yeah, everyone works hard at this versus, know that you could be automating this away. Like that, you know, that email, that should probably be a trouble ticket. And by the way, those trouble tickets should increasingly be resolved by robots. And by the way, those could also be scripting automations that notice when it's been resolved because your exceptions process that is scripting around your templatized environments for when you're allowing an exception. You know, like these muscle groups, they just get stronger. It's like working out. So I love the analogy, because all muscle groups work out in their spine, work bigger muscles. You got, is it leg day today? Is it upper body? Exactly. So, and also when you play different, let's say sports, for example, different muscles are used. So I have to ask the question, do you bring up some of these things about automation, which is clear, great opportunity to use AI. Is the game changing a bit from a muscle standpoint? What is needed to be augmented or developed that you see coming that is addition to the bigger muscles? Obviously threat detection, data protection. You know, data protection, a ton of people doing that. You got threat detection is always hard. Not many people are doing threat detection. Is there new muscle that needs to be built? Other we are, yeah. No, it is. Yeah, exactly. I think- Well, it's more crowded in data protection than it's threat management, I think. I think that's right. I mean, I think that, look, you want to reduce, you as a CISO, right? You want to reduce the likelihood of a bad day. You want to notice when your bad day starts and have it have little to no impact. And then you also want to get that for the ROI that you are aware of. Like for something that you have already bargained for. So is it a known, known, is it an unknown? You know, like these are business decisions and security executives are increasingly realizing that they need to present this in a sense that makes investments relevant. And so I think this is where, you know, like the, so when folks go off the rails on like, oh, well, AGI is going to like take over humanity. Have you talked to Siri lately? Like, it's not, you know. It's not close to being taking over anything. Right, your people are going to be working on different stuff, hopefully, right? You know, you'll, I think that we do see your, the security work you do will take different forms and you'll get better at those too. You know, this hopefully becomes this flywheel where you have folks who are increasingly automating. You know, we used to joke that we like to hire lazy security engineers because we want them to never want to solve the same problem three times. And also that there is kind of this beauty in being able to know stuff on a logical level. So, you know, old school security was like checking to make sure your VMs are plugged in. Now you're able to do everything at an API level. You know, you're beaconing, you're this, you're that, you know, like there are ways in which your logical controls are what you're constantly refining. But of course that means that you're dealing, again with like data lakes and analytics on analytics. And so it's really important that you're getting high fidelity alerting and you're not getting alert fatigue. That it's stuff that matters to you. So like refining those alerts into what you can really care about. So the game's changing. Great to have you on theCUBE. I wanted to ask you while you're here, what's your plans for re-invent? And what do you expect to see? I saw Steven Schmidt walking around this we're in the big meeting area here, the press area. All the execs saw Adam earlier. Obviously they're going to talk about their JNAI. We're going to see some chip improvements, chips and models together, pairing like wine and food. Changes a lot more decoupling of storage. For instance, you're seeing a lot of these, a lot of speeds and feeds kind of getting, trying to nail the performance and cost equations. As that happens, that infrastructure is going to get faster, better, maybe more secure. What is Amazon doing here? What's your takeaway from the show? And then what's your agenda? What are you going to be doing here? Doing meetings? Are you going to do some media as well? Let's take those two questions. What's Amazon doing? And then what's your plans? My sense is that while AI is kind of like the theme, there are not a ton of new service announcements. It's mostly refinements on existing services, which is reasonable given the explosion of services in the last few years. I am excited to see folks be building out just easier ways to make it prescriptive for how folks should be using technologies. And we are part of that, so doing more built-ins, having it be more managed, really embracing that customer experience of not just saying like, oh, you can do anything, but here's how we see folks succeeding at getting the right kinds of responses. And then I'm really looking forward to, obviously learning from my peers, but also finding ways that I can help folks understand what to focus on. The biggest question I get from CISOs is, am I focused on the right stuff? And I think that as we look at this, some of this AI stuff is new. A lot of it is just the kinds of muscle groups that you need to have in your organization to both embrace new technologies, to fall in love with them and make it so that security is part of the heartbeat of what you do. And then also, part of what you take in as, I don't want to say a threat, but like take in as something to deal with as a security team is like, yeah, let's embrace this new technology, but put guardrails on it. And you have plans, you're doing some media here, and you're camping out. Yeah, I am, yeah, I'm hosting a dinner and a lunch, and I'm running into folks, and I am looking forward to interactions with folks. And also, of course, probably going to be falling asleep at 7 p.m. since I'm on Eastern Times. I'm sure that's not gonna happen. Well, we appreciate you here in theCUBE. We appreciate who you are and what you do. Love how you're on social sharing, a lot of data with folks. I know you talked to a lot of customers as well on the business side, but great show, great content you're putting out. We really appreciate you doing that. We learn a lot and we like to work with you. So thanks for doing that. Well, John, the feeling is mutual. Okay, we're here in theCUBE in Las Vegas for AWS's annual user conference. It's called Reinvent, part of our special SuperCloud 5 Palo Alto live stream out of the Silicon Valley Bay Area Studio. We're bringing you on location Dave Vellante. I'm here. All of our top people, our analysts are here. And keep watching. Move right back after this short break.