AWS Graviton Advances, Cloud Security Risks, and AI in DevOps | TSG Ep. 980

Train Graviton Blackwell. Oh my. You're watching Textron Gang.

Hey everyone, it's Alan Hummel and we are still here in Las Vegas. You know, I do have a three day rule at Vegas, and I'm running up against the edge of it here. Um, I hope I've got a real hot shower waiting for me at home.

But, uh, it's been a great AWS reinvent in many ways. It's, uh, I think a watershed moment in AWS, uh, history with, with a lot of the announcements here. We've got a lot to bring you today, including a, a, a new version of the gra or a new generation of the Graviton chip that was announced, uh, yesterday.

Um, you know, hardware sexy, there's no doubt about it. We spend a lot of time talking hardware. Let me introduce you to our gang, though.

We've got a lot more to cover too. But if I wanna introduce you to these two guys, they've been sitting here to my far right. You know, our friend Mitch Ashley, future Analyst, and the man in the middle, the Bard, Mike Ard, who, uh, happy anniversary to Mike and Charlene by way.

Yeah, There you go. Um, so Mike, congratulations on the anniversary. Thank you.

It's a lot of bookies in Vegas that are losing money. Yeah. They took the under, they took the under, but, um, no, seriously, congratulations and lucky you.

Um, but let's talk, let's talk AWS reinvent. All right. So, uh, they previewed the new Graviton five, which is 25% better than the GRAVITON four, which is both of which are based on an arm design that AWS custom bills itself.

And the other thing that was interesting is they revealed for the third year in a row, roughly half of the new instances being spun up on the AWS clouds or GRAVITON chips, which is, you know, a significant number of workloads that probably used to be on an X 86 process or somewhere. Yeah. Mm-hmm.

And it also turns out that all the managed services that AWS provides by default run on AWS custom silicon. They do not run any of those things on any processors that they offer other customers elsewhere. And basically they're saying, look, um, there are certain databases that require an X 86, so you know, we will provide that capability.

And if you really want Nvidia, we'll provide that capability as well. But, you know, they themselves are going with their own custom silicon just because it's less expensive for them to run. And they say, Hey, the TCO is better.

And of course, it's more profitable for them as well. But it's pretty clear that AWS is saying, you know, we're leading heavily with our own custom silicon here first, and everything else is kind of, you know, if you want it, you can have it kind of mentality, but you gotta actually like raise your hand and say, make sure I'm running on that. That's why it's such a different world for Intel these days, right.

I mean, and it's, you know, apple has its chips, et cetera. All right. On down the line, it's, it's about price performance.

Mm-hmm. You mentioned cost, right? It it's price performance.

It's also control Control Supply chain. That's True. Yeah.

In this day and age, very important. Mm-hmm. Yeah.

Um, you know, it, it's an interesting thing because look, we all grew up in, in basically a Wintel monopoly. Yeah. It was Intel inside baby.

Yeah. Even Mac was running Intel back then, and I, I, I sometimes think that ARM doesn't get enough credit for cracking this case wide Open's Huge. Yeah.

Right. It was ARM that led the way to these custom chips and, and compatibility and, and you know, it was, it, it wasn't easy getting Microsoft to port windows to arm. Mm-hmm.

Or, you know, it was easier with Linux and some of the open source stuff because you had the army of, you know, crowdsourced open source engineers who, who could make that happen. But, but this is the future we see now we live in now where we're not locked into Intel inside. We're not locked in, you know, a MD poor A MD spent all those years making a grade 86 chip.

Mm-hmm. And you know, a little late for that though. They're here with the GPUs maybe.

Mm-hmm. Um, one thing I've gotta mention that I want to set the record straight and, and I might be wrong if I'm wrong. Tell me.

So AWS doesn't own foundries. As a matter of fact, AWS doesn't actually design the chips. They, they hire out, This is O em outsource.

Yeah. Right. They they provide the specs and the design.

Yeah. Do some of the design themselves. Yeah.

They do some of the design, but they're not, they're not building a foundry. No, they're not. O and that's the whole, that's the beauty of the armed design, right?

It's licensing the technology to be able to design your own chips and mm-hmm. It's So, are are, are the new graviton and trains, who, where, what foundry are they using? I don't know.

They don't say, and I think they just shop around because there's a ton of these arm foundries out there now that you can use to generate, it's a three nanometer chip is the latest one. So they're down to that particular, um, size generation. But, you know, the thing about it is that, um, it gives them options, right?

They can move from one founder to the other. They're not as dependent upon a particular, Um, well, that, that's the point, right? Is there, they're in the chip business, but they are, they're not a foundry, they're not a direct manufacturer.

They're an OEM client. They have Chips designed for themselves, or they design 'em themselves. It gives them the ultimate flexibility to have chips made to their spec.

It gives them the ability not to be subject to price wars supply chain mm-hmm. Geopolitical kind of, uh, instances. And you know what, to me it it, it validates Pat Gelsinger vision of Intel, which is be the foundry that makes all of these different chips, not just your own.

Mm-hmm. We, we are living in a world where this isn't just an AWS thing. Google has their own, Microsoft wants to go their own way.

Oracle, IBM, everybody wants to kind of own their own custom silicon, if you will. It's back to, you know, Mitch, when we, when we first started selling security appliances mm-hmm. Right?

There were the ones that were just X 86 inside that we OEM from Dell. Then you were like, And, and then there were the ASIC ones. Yeah.

You know, purpose built that really couldn't kind of handle the kind of traffic that we were seeing and stuff like that. It's the same thing playing itself out right now. Yeah.

They are saying that a, a Java app runs pretty smoothly without any tweaking, but some of the C apps and their rust apps require a little bit of tweaking to get onto the, um, arm processor. But they're saying that's gotten a lot easier. And they're also pointing out that the classes of workloads now are starting to move higher up.

And you're gonna see more with this latest processor of things like databases, analytics, and, uh, electronic design automation was another use case that they were looking at. And so the percentage and the types of workloads that can run on these things are starting to, you know, creep up into the higher end. And you can imagine they're saying they're on a two year cadence for the next generation.

So if we're in 20, you know, 2028, you're probably looking at a new class of processors that can run just about anything that you have on those arm processors. And, um, it's gonna be an interesting world because you see Microsoft doing pretty much the same thing. And if all the cloud providers are doing all the same thing, it's like, well, what is the market opportunity for the Intels and the a m as A foundry, not, not as, uh, a next 86 producer next.

Well, and this didn't all happen, you know, overnight arm has been on a multi-year quest of getting organizations open source projects, getting drivers, getting everything running on, on, on arm. Mm-hmm. And they've had them added a lot of support to the open source projects, for example.

Absolutely. I never seen three years at one of the conferences talking with our folks there. Why were they there?

And that was, that was the whole reason. If you kind of build the stack up so you've got support and it's closing the gaps where you've got 'em when, when that occurs. And as opposed to, we can't do it 'cause there's so such a big hill to climb to get there to run on.

We're not, we're not there anymore. We're actually in a good place. I don't know if they're thinking about it or has said anything in the kind, but I could imagine a world where maybe there's a arm processor optimized for AI inference service.

I would be surprised there. Think I, I think we'll see it sooner than later. Um, you know, though, it makes me wax nostalgic a little bit, right?

Think back before X 86 was the data center machine of choice. Right? I was an ultra Spark guy, right?

Yeah. You had deck and you had, uh, silicon graphics and you had, you know, think back to some of those machines and the difference, and you know what, and it wasn't just the chip, but each chip had its own flavor of eunuch. Mm-hmm.

Right? There wasn't just Linux. Think about before there was this, it was before Len Linux.

Yeah, yeah, yeah. But we had all the unixes and, and even had Motorola. I mean, what, Well, Motorola was a big apples on, you Know, the max were, Um, and you know what, you, let's talk about Apple in this context.

They kinda led the way, not in the data center certainly, but on the desktop they showed it's and device possible Device. Yeah. Right.

And mobile. Right. They showed it's possible.

And, and, and we should talk about that too. You have, you know, what we consider mobile, the sort, the lightweight, more energy efficient, the Qualcomm snap js and stuff. They're running desk laptops and desktops all over.

Might we see a Qualcomm enter into the data center space because hey, we're building $3 trillion of data centers. That's a pretty big pie. It's possible If there's a market, right?

If they see, if they see themselves in that market and they see a competitive edge that they can bring something to market, they will, Lemme throw one other thing out at you guys and tell me what you think. Is this a lock in play by AWS If you optimize your app for Graviton five, does it make it a hassle to move over to Azure or Google? Uh, it's one more little thing that makes you kinda wanna stay with a AWS.

'cause you might have to refactor your app a little bit to take advantage of whatever optimization those guys have put together. Right? So eventually you might decide that it's not worth the effort.

So, because That is what we lose by having a, a PAX 86 world, right? Where you, you, you do have that portability. I I think you're more likely gonna have stickiness from the AWS services that you're using, the Dynamo dbs and all the things that you're tied into the ecosystem.

Yeah. Less so at the chip level. 'cause there's not gonna be that much differentiation.

Well, I don't think, I think they're doing it for cost performance, that they don't wanna lock customers in that way. That's not their lock in on software services. So, but, but AWS does have this kind of interesting little lock-in that they do use, right?

They have this Nitro system. They do. And Nitro is essentially offloading the running of the VM and the storage and the security onto a separate nic cart.

Well, I can't replicate that over at Google per se. So if I'm gonna take advantage of that, then I gotta go figure out what is the version of that over at Google. And I'm not sure what Google has as the equivalent of their nitro, but, you know, uh, it's a significant amount of effort to kind of rework all that It is.

And it's an advantage from a security perspective and security within ai. That's kind of how they really pitched it two years ago when they went into more depth about how they built Nitro or how they pulled things off the chip on onto, uh, into the separate Yeah. It's kind of a chip within a chip or a chip on a board that Yeah.

Offloads those processes. Yeah. If You go talk to the AWS engineers, they're like, that's the best thing we ever did was Nitro.

Because that's what gives them a, a performance edge in their mind that this, that differentiates them from other cloud service providers. Absolutely. Hey, we gotta move on to our next segment here.

Okay. So we're gonna take a break. We're coming at you from Vegas.

AWS reinvent. You're watching Textron game. You've Earned it.

The spotlight, the responsibility, the weight of teams, companies, and entire industries fall on your shoulders. Lives depend on your decisions, your home life included, that work your are protected physically and digitally. Nothing gets through your team without a fight.

But in a globally connected world, everyone sees you, including those who mean to cause you and your organization harm. And now home your sanctuary attackers see an opportunity. Your digital front door is wide open.

And what compromises your home can breach your boardroom. Because the devil's greatest trick isn't targeting your workplace firewall. It's convincing you that your personal life isn't at risk.

Black clerk, digital executive protection, defending the new attack surface your personal life. Hey folks, we're back and we're gonna shift gears a little bit and talk about, well, security in AWS they've, um, made generally available this security hub that they announced earlier this year. And it's basically a, uh, a huge data lake for aggregating and analyzing all the telemetry data that you collect from all these different sources.

I think AWS has been doing this internally for themselves for years, and this is the productization of that capability for their end customer. The concept is that, you know, you're gonna collect all these signals and the metadata, and when you see an anomaly, there'll be an alert, and then eventually an agent will look at that and kind of pass on a message to, um, some DevOps tool that says, Hey, there's an issue here. Maybe with this configuration, do you want us to re-fix that for you?

Automatically, it might close that loop, but eventually we seem to be moving down this path where a lot of this is gonna get automated. And then the other thing they were talking about is, you know, how they're gonna move everybody to short-lived tokens? Because part of the issue is that, um, too many developers are making use of, um, longer time tokens that when they get stolen, can be used for longer periods of time.

So they're trying to say to everybody, Hey, use a short-lived token so that if that laptop goes missing, um, you know, it won't be as big an issue. And they're also pointing out the fact that they tied their MFA to the browser security. So they look at the browser and then they say, Hey, if you can authenticate it, then at least we can pass that authentication in a federated way back to the CLI tool.

So they're, you know, making some efforts here to like, make the whole thing more secure and address a lot of the, the issues that folks have with cloud security. But I kind of always scratch my head about this. Are we, when we talk about concerns about cloud security, are we really talking about we think that the cloud is less secure than on premise?

Or are we really saying we don't know how it works and we don't have the processes and the tools? So we kind of still prefer the on-premise mindset because we don't wanna learn something new. I mean, how much of this is cultural versus technical?

Yeah. Well, I think these announcements are more about similar to what Cisco did with their ai, with their Cisco data fabric. Mm-hmm.

Right? Which is kind of the same idea of instead of just aggregate aggregating things together, let's put more contextualization around it, uh, rather than just a bunch of siloed data kind of co-located together, let's put it into some kind of a data lake or some, uh, data method, data structure that can be used by their tools and help reduce some of the complexity. I think that's a, a big part of it is since you don't run all these tools yourself, you don't necessarily know how to integrate them, or maybe you can't, right?

Mm-hmm. Especially if they're, they're a third party like the cloud providers, so you have to rely on them to do this. So I assume this is a big push from customers to say the, we need easier ways, we need less complexity about managing security in the cloud, at least for the, for the data fabric part of it.

I think, I think the definition of shared responsibility keeps shifting a little bit. There's only one responsibility and that's yours doesn't matter. I mean, talking about Sarah, You want, you could, you could share, Ultimately it falls in Administration, but you don't share responsibility.

Look, here, here's my, here's shimmy's stake on this. Look right? All the noise about cloud security, there's only one thing in cloud security that, well, you're responsible for everything, but there's only one area where you could really kind of jig, you know, rig the deck.

And that's identity and excess management. That is the killer app of cloud security. And that's what this is all about.

IAM um, in terms of token a, this ist AWS's idea, NIST has been pushing this, Google's been pushing this, this is in the browser. They weren't right. They weren't browser certificates to expire.

Yeah. We went from yearly to quarterly. They want to go to 30 days and get to a week or a day even, right?

Because good security practices is, if we could make you re-authenticate as often as possible, the the least chance we have that you are give it stolen, Say somebody can, it's only good, Right? And, and there's, and there's a short window, right? Of mm-hmm.

Of its viability. So this is, this is kind of security 1 0 1. Now that being said, Bravo to AWS for doing this because this, this is what cloud providers need to be doing, right?

Google's on board with it. I'm not sure about Microsoft. AWS appears to be, but I think this is the future we're gonna look at.

We're gonna live in, right? We're gonna MFAs and, and biometric and stuff like that. Um, you know, we could just get rid of these damn passwords.

I just wish that the MFA stuff was a little smoother. And it's not like, you know, every time I'm kind of like logging into something, I'm like, damnit, where's my phone? I gotta get the authentication.

Yeah. You gotta have your phone next to you. If in the world of MFA, my problem is, is I'm a, I'm a multi, um, I'm a multi lotter things, but I'm a multi-device user, right?

So it's like, yeah, I've, I've authenticated the browser on this device, but I haven't authenticated the browser on that device. I use about five different devices, and I'm constantly playing the authenticate this particular version of this particular browser on this particular device. And the permutations drive me crazy.

And it's still clunky. Yeah. I mean, it's clunky even with PAs keys.

I mean, yes, it's simple when you can use it, but oftentimes it's like, well, passkey didn't work on that one. Or choose what you want to use, which PAs key or which C or passwords. I, we got what PAs keys.

I, So we, we kind of kick ourselves when we do good things for security by making it complex to use by presenting options. Yeah. It Defeats the burden.

Flexibility brings complexity. And that's often where we fall down on security. I would like, like a clear ID for all this and be done.

Mm-hmm. Charge $45 if you don't have a clear ID when you show up at the Airport. No, that's real id.

Okay. All right. Real id.

Yeah. Well, look, someone's gotta pay In theory though, right? We are getting to a world where there will be sensors everywhere.

So at some point, the sensor should be able to know that you is you, and send that information over to somebody and say, yep, that is Alan at the edge of that particular device. And maybe we will get to something that's not nearly as clunky From your mouth. Let's take a break.

We'll, we we're gonna ruminate on that a little bit. We're gonna come back and talk about our third session today here for our special Techron gang from AWS Reinvent, Discover Techron Group, the epicenter of tech innovation. We are your go-to for reaching IT leaders and practitioners worldwide.

Our secret impactful content that sparks awareness, engagement, and top quality leads with us. You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more. Join our satisfied clients.

Let's revolutionize your tech journey. Contact us today and tell your story to the world in the most powerful way with Textron Group. Hey folks, we're gonna continue a conversation we've been having all week about AI and DevOps and DevSecOps.

And the folks at Harness are here, and they announced that they're adding support for the AI coding tool from AWS called kero. And we've been trying to figure out like what is gonna be the relationship between how far our DevOps process is gonna be shifted left. Because Kero has a lot of AI agents that will in parallel do things that historically a DevOps engineer did, and including observe and maybe fixing code and patching.

But, um, harness is saying that they too will have AI agents and their AI agents will communicate with the Kiro AI agent. And then I did talk to the AWS DevOps people and they were saying, we wanna manage that process more, but we do not want to provide the code repository or the, uh, or, or the CICD platform either. They're like, we just wanna layer the agents across to take care of the friction.

And if the agents are from third parties, that's fine, but they seem to be trying to find a happy medium here. I mean, what's your take? I see a lot of harness electric billboards in my window up here on the 34th floor.

Um, harness has a big presence here. Mm-hmm. Um, kudos to them for not just laying down, rolling over and saying, we'll use the AWS agents.

I, I think in a world of billions and billions of agents, there's room for CICD DevOps specialists, just such as Harness to have a very effective DevOps group of a, uh, group of agents that help you with DevOps, where, you know, in no disrespect for AWS but their agents will probably be the 80 20 agents, right? 80% of the functionality of what that dedicated harness agent is. But at, uh, you know, 20% of the price.

And I think that may be enough for some people, and there may be other people, especially if they're already using Harness for CICD who, who say, no, I'll, I'll use the harness agents. And the fact that it does integrate with the, uh, kiro and, and the AWS infrastructure is fantastic. Mm-hmm.

It was interesting 'cause when I talked to the AWS folks, they were saying the part that they really want to get after is between, is what, what occurs after code gets checked in and when it gets deployed in the production environment. 'cause they were saying it's just, there's just a lot of friction. There's a lot of inefficiencies, and they want to build, you know, the AI frame framework to manage that.

And if that involves, you know, agents from the providers of the repositories and the CICD so much the better. But you've been looking at that workflow forever. What is it in that between when I check in the code and it runs in production, that is like creating this infinite amount of friction?

Well, well, first kind of like, one thing that I see happening with all the vendors is folks have realized we can't be the agent of everything. We can't be the AI of everything. What do we do?

Well, what's our part in this ecosystem and how do we take an AI strategy that builds on our strengths as opposed to going after, in this case, it doesn't make sense for, uh, AWS to go after all the development agent process part of it or, or become the code repository or whatever. Um, that's got huge inertia behind it already with other options, right? Mm-hmm.

So rather than forcing people into an AWS solution for that, they are tackling it with the IDE, though the, the part of it, it's deployment. We've always run into this with CI ICD Allen, it's like ci, big little cd, right? Because that trust around automation of deploying things into production, we've always had steps to review and checklists and production ready, whatever kinds of things.

Now, not everybody does that, you know, you can get up, sit all the way into production, but there, there's a lot more caution about releasing it. I think it's goes back to the trust issue. Mm-hmm.

How much we will automate, but we still won't kind of turn it over to even a, you know, whatever a Pearl script or a a Python script, whatever might be underneath running that. Will we do that with ai? I think that's the goal.

How do we move that? And, and by the way, I think Harness is super smart for not saying, well, we're just playing our little sandbox here. No, everybody, whether it's security, whether it's deployment, et cetera, compliance, they're moving up the, and they're shifting left by, by getting into the agent ecosystem.

AWS did say something that's gonna warm your heart. They were saying, Hey, we're using AI and coding tools and agents to write more code than ever. And if you don't have a CI platform to manage all that, you most certainly will.

I need one soon. Yeah. So I think that's for sure.

I, I think there's something to be, to your point on the CICD stuff, harness is a little bit of a different animal than some of the other CICD players in that. And, and I've spoken about this with Jodi, their, their founder, Jodi Bonsai, they were designed from the get-go to BCD, most of the CI ICD solutions out there started life as a ci, right? Mm-hmm.

And then, you know, added the, you know, big ci, little CD harnesses, big cd little ci. So it is different. And it did embrace GI UPS pretty early on and say, Hey, we, we work really well in that GI ops type of mm-hmm.

Uh, environment and that GI ops model. So I, I think they are, they are well positioned to do this, perhaps more than some of the other Yeah. The ops players.

The only thing I hear out there that people are looking at for CD as an alternative, as you hear a lot about Argo, which is, you know, as far as I can tell looking at it, is a much more graphical approach to CD rather than program. That depends, that depends if you are using the pure open source Argo cd right? Or you are using like the Octopus deploy version of it, which is more graphical and has a better ux mm-hmm.

Supposedly, you know, look, UX is right. What you like may not be, I had to behold it on that. Yeah.

Yeah. But those people tend to argue that the CD side of the equation is really, you know, an ops function and it's not something that the developer drives as much. Uh, So that's now you're getting into the nitty gritty of what's Dev, what's ops and what's dev ops, what's welcome.

We're all in there. Yeah. Welcome to the religion.

Yeah. I know, I, you know, I, I don't talk about politics and religion, you know, me. Um, but excuse me, be that as it may, you got a little choked up there.

Excuse me. Yeah. But you know what?

It, it is gonna be interesting, but hey, I, as I said before, harness has a huge presence out here. Uh, they obviously feel this is where their people are. Mm-hmm.

And, and they're making a play for them. And Mitchell, they are, they're shifting left, moving up, but they wanna be a player in this, this new ecosystem. Yeah, absolutely do.

And I've heard their story and they've got a great story. Yeah. I mean, it really Is very impressive.

I was on the show floor yesterday walking around, and I was just taking note of all the DevOps companies that were here. I mean, they're all here with big boots and Yeah. Including some outfits I never heard of.

And I was like, looking at some of their boots and I was like, wow, this is huge, and I have no idea who you are. Mm-hmm. And Well, you know what's interesting, I overheard of Venda talking yesterday, Hey, we gotta go reserve our booth for next year because we need a bigger booth and we're too far out.

We gotta be more in the center. Let's take money from other events and double down here because this is where our audience is. Interesting.

And, and you don't hear that very often. It shows these days, You're just complaining about how expensive it's Right. It is.

But, and this company was already in for a hundred K. Mm-hmm. So they're, they're moving Now.

I did hear, and I, I cannot divulge sources here, even if you want to, you know, Methods and sources. Sources, can you, Yeah. Deep throat, but, uh, easy.

But, um, I heard reinforce might be going away, well, not going away, but they're gonna merge it into reinvent. It's not gonna be a standalone, The security conference. Yeah.

Enforce. I haven't heard that. I could, I could see that.

I mean, well, reinvent started as a developer conference. I mean, I went to the first reinvent, it was very, because that's who was really, but I, I'm gonna tell you something. This reinvent was a return to its route.

It was, It was developer first, the reemergence of the developer conference. Right. A developer led conference.

But it'll be interesting to see if Reinforce gets merged into this or stay, stay standalone. Mm-hmm. You know, look, we have, we have RSA, we have Black Hat as big security conferences reinforce was never that size.

Mm-hmm. But it, it had a nice following and what I liked it, it was, they're good on the East Coast and mm-hmm. I think there's something to be said for bringing everybody together to have the same conversation at the same time about the same thing.

I, I think the problem is it's already really hard to get put 60,000 people into the same conversation. I don't know if I want to add 20 more, I would argue spin off and do a developer, a real developer conference. So it says the developer Guy, backpacks and hoodies.

I'm, I'm serious that that's the crowd you're going after. You want the GitHub style, you want the build or style kind of event or the Ignite Style event, um, tech exchange, you know, those kinds of conferences that are for, lemme give you an example. You, you go to IBM Think versus IBM, tech Exchange.

Two very big, big, big differences, right? It's all about sandboxes and trialing things and working with tools and, you know, experimenting and showing demos and code c***s and hackathons and things like that. Versus you go to the corporate conference.

It's all about kind of the bigger things that are happening at the, the mothership. I think developers, if you wanna attract developers, give them what they want. Right?

Fair enough. Hey guys, we've gotta wrap up. You know what, it's been an amazing week out here at Reinvent.

It's lot, so much news. I think we'll still be digesting all of this for the weeks ahead on, on the gang and throughout our tech strong, uh, properties. But Mitch, you know, I appreciate you coming down every day, Mike, as always doing yeoman's work, walking that floor, ferreting out the latest, the underground, the, the, the seldom scene putting the t on the ball for the three of us to hit it on a Textron gang.

Absolutely. I appreciate it. I think all of you who've been catching any of our, uh, videos this week, I think we probably have done 20 something videos in the three days, maybe more.

So it's, it's been, uh, uh, great. We had the FU analyst on, we had Daniel Newman on, we our friends, I can't help but thank Susa. Oh, this isn't their logo, but they, they had, we've had their logo behind here.

Susa has, uh, partnered with us this year in, in, uh, bringing this to you. We love coming out to reinvent. It was a little warmer in Vegas in the winter.

See what we could do about that. But, um, we hope you've enjoyed it. We'll be back tomorrow though, with our regular gang show, you know, from Boca.

And, uh, I think, anyway, and, uh, God willing, and we'll, we'll go from there. But for now, on behalf of Mike and Mitch and everyone from Fu intech Strong and suse, thank you for joining us. It's a wrap.

Bye-bye.