Techstrong TV October 8, 2025

Data centers on fire. No, literally data centers on fire. You're watching Textron Gang.

Hey everyone. Happy Wednesday. Welcome to our Strong Gang Show for a hump day.

We've got a lot to talk about. It's been a busy week. I feel like we put a week in already.

It's been a busy week, but we've got some interesting stuff to talk about with some interesting people. Let me introduce you to them. We've got our friend Chris Blak, the one and only Kate Scarcella, Dan o Dan O'Brien, and joining us from Barcelona, Mike Ard.

Hey, Mike. Hey. You able to get the Yankees out there?

I'm gonna find out in a couple hours, but, you know, wonders of the internet being that was yesterday. Yeah, This is true. All right, let's jump into it.

So, you know, we spent yesterday talking about the, uh, exciting news on a MD and, and, uh, open AI the day before that. More exciting news, it seems every day there's exciting news around AI and data centers, right? I was talking to Dan o yesterday or the day before.

You know, something like half of the growth in the U-S-G-D-P this year can be attributed to AI and data centers. JP Morgan put out a report now for the first time in a very, very long time, consumer spending isn't driving the GDP growth. It's actually AI and data centers are bigger than consumer spending.

Consumer spending is usually 70% of GDP. You know, are we in danger of, of kind getting in over our skis here? Can, can AI momentum outpace our ability to, to, to build out the infrastructure?

Mike, what, what's the deal? All right. Well, we have two things to talk about here.

One is there's a report from our friends over at Volter who confirms that well, demand for AI is far ahead of the available AI infrastructure, and we're not able to keep pace. And I don't think that's gonna surprise many people. However, at the same time, Fujitsu and Nvidia have been talking about a new design for AI applications and AI agents in general, and a lot more integration at the silicon level.

So you would see GPUs more tightly integrated with CPUs, in their case, their arm processors, but also there'd be asics and networking and infrastructure all built into the system at a much lower level. And Dan, I'd love to get your impressions on this. 'cause I wonder if a lot of the stuff that we're using today might seem fairly antiquated soon as we look at these new platforms that people are building.

And are we finally building platforms that are made for agentic ai? Yeah. Well, I think we're getting closer, Mike.

Um, you know, we've thrown, you know, really general purpose. GPUs a lot of work on the software side as well as the hardware side over the past several years on really building efficiency and right, the move to rack scale systems, you know, getting to the point where we can get much higher utilization rates out of existing GPUs. Uh, that was a lot of what was behind the news yesterday with, uh, open AI and a MD was, that was really the first AMD's first rack scale system, which integrates this all much more tightly.

Um, and I think, you know, there's, there's certainly a mindset out there. I mean, you look at the power, yet the power efficien, the power efficiency of the human brain, we're a really long ways away from that. Um, and so, you know, I do think we're gonna continue to see more and more efficiency gains.

You know, there's a lot of people out there that believe, and I probably count myself as one of 'em, that we need a fundamentally different architecture to kind of scale a high to the point we're eventually gonna get to. Um, and I think, you know, a lot of the announcements here of like immediate jiujitsu, you know, they're all incrementally moving the ball forward on that map. Um, I mean, just, you know, recently as we moved from, you know, LLMs and the more the agent world, you know, you're seeing 10 X tokens and under x growth in compute.

So, you know, the, the type of AI we're doing is we're getting more reasoning. Um, and some of these, you know, and more advanced styles of gen ai, uh, the compute demands are just off the charts, right? You know, we're, we're years away from, you know, catching up on the supply side as we build out this infrastructure.

And, you know, there's a couple gating factors there that we're working with too, right? We're hearing more and more that it's really about the power supply, uh, but also, you know, there's not enough chip making capacity to build all the chips we're gonna need for this stuff. And I think, you know, TSMC being somewhat of a monopoly in the market, you know, is pretty good at kind of policing that overbuild on that side of things feels like we're really gated by the power at this point.

Uh, but an enormous amount of compute will still need to drive over the next several years to get supply to catch up with demand. But lots of incremental innovations that will happen over time. And I think, you know, ultimately a fundamental architecture shift.

I think we're, you know, right now in kind of the system level design thinking, but, um, still using a lot of the, the same architectures we have that that will certainly have to evolve over the coming years if we ever want to get power efficiency down to the point where we can actually afford all of the supply that we, you know, see to need on the compute side of things from a power perspective, You, you know, Dan, what you've just described is exactly the issue. Why we always want software versus hardware. Because software as we improve it, you just update your software, right?

It's a download. It's a di the old days we put a disc in, and you got, you got the next version. It's more efficient, it's better, it's faster, it's everything else.

With hardware, you can't just update your hardware, especially when, you know, you're, you're literally putting hundreds of billions of dollars, trillions of dollars into building out these hardware factories. And even though it's incremental, the drag time between how fast the human brain can design AI assisted, of course can design new hardware, you know, uh, con configurations versus the time it takes us to actually build those configurations. It always seems like we're going to be driving last year's model next year, right?

We're building like the world's fastest depreciating asset. Yeah. But we're gonna need to shrink, you know, the IRS is gonna have to come in here and let us, you know, write it off after 24 months 'cause it's obsolete.

Um, Just feel like there's some sort of financial engineering bomb around the depreciation schedules of this stuff that will eventually hit the industry. But yeah, it seems like knock down the road for now, You know? And what about, you know, how tech people are?

We always, we want the fastest shiniest, you know, trinkets. Now, we don't wanna wait. What, what's, what's a tech bro to do?

Chris, I saw you had your hand up. I, I'm just thinking, you know, I, I love this. You know, for me, we do this show once a week, right?

And we've been doing it long enough and we all the terabytes of video, much less the exabytes of stuff we're talking about as I'm listening to this, I'm thinking of these tiny little attestations between us. Like literally Dan, you remember a, a power conversation a month or so ago talking about the grid, you know, and we're all, you know, adamantly agreeing, we need more power and all these little spots, you know, week to week through this year stitched together. And the the answer is to these questions are, are in there, right?

You know, so whether it's, you know, from the power issue, yes. The way we're scaling power, you know, is queued up for, for this segment. The grid actually doesn't match the expectations.

But we're going to do those things. And you know, we, we've had a couple conversations about the architecture and workflows and workloads. You know, we're, we're putting everything in the mass of ai, you know, to do math or, or to, you know, change the capitalization in, in a paragraph.

And on this, as I've mentioned before on this tiny little laptop back here, there is LLM running. This is for Clifford Brewery. Shout out to Darlene at Clifford, who keeps the management there, keeps all running, and of the benefits of all this stuff.

So we've been working up and down the, the supply chain down to the micro business, the underserved who has a million different applications. And all the stuff we talk about here hits them coming and going. They can't see it coming.

But these systems we have now allow us to serve that need. And when we talk about this ai, we're talking about taking little bits of human text and turn into data and vice versa. So people like Darlene and operations like Clifford can do what they they do.

And if it works at the foundation like that, then maybe these e esoteric things we get to talking ahead about here actually matter, because the, the physical grid is not moving faster than, like you said, Alan designing things physically implement. These are huge, long pro uh, projects. A lot of the tra trajectories that we're predicting, living, experiencing right now don't match those.

What are we gonna do? Maybe some of the things we've talked about on the show are, are right, we'll spread it out. We'll get more intelligent and we'll do all these things.

'cause Oh yeah. We're, we're, we're embedding AI in everything. The benefits are too huge.

So something that we've talked about though on this show has been the idea that we are building for tomorrow, um, on yesterday's thought, on yesterday's architecture, right? And Dan, you've talked about that a lot, about the architecture. Um, AI is sprinting on a runway.

We haven't finished paving. And I think with that being stated, I I do think the architecture is, is changing. I think it will actually become more distributed, uh, for a lot of different reasons from a sovereign perspective, right?

Um, having, you know, to have agency over, you know, geopolitical, our, our, our data, right? And I think the whole architecture of a cloud model, it, it, it's not gonna work for, for the chips that we're building. And, and it doesn't even really make sense.

0, um, distributed model, I think. And yeah, I mean, and it, it will make us, um, more resilient and everything. Again, I think we're building on, on, it's almost like we really have to start thinking differently.

And we're not there yet. I think we're, we're there to do so. So, Kate, follow up that thought for a second.

If you would, do I, as a developer, let's say I'm building some sort of a gen AI app and I'm starting it now, do I assume a level of compute capability that might be available a year and a half from now? Or do I kind of not trust that and I just go with what I know I personally would, would go with the chips that we're not sure what are what? 'cause we're not utilizing them that, that we know for sure we're not utilizing those chips to their fullest.

And I would go for it. I think the one who goes for it is the one who's gonna win. Oh, I was just gonna say to me, the, the long poll in the tent here feels like power.

You know, the, the total capacity and the location of it. And to me, that's ultimately gonna drive the constraint based thinking across the industry. You know, there will be somewhat of a power ceiling we need to design for.

And, you know, I think we can count on innovations down at the hardware level and then spilling down with the software level that allow us to essentially accomplish, you know, kind of what we need within that envelope. Uh, that to me is, you know, kind of the right way to think about this. Um, I think the chip market will, you know, basically builds, you know, to the point where, you know, you can deploy it and you've got the power to run it.

Um, that's kind of the constraint on the upside of the chip side. Um, but, you know, there's, there's a lot of levels at which, you know, we can optimize here. You know, we're, we're very much still in the build and move as fast as we can phase, eventually we're gonna hit a little bit more of a plateau.

And that's when you tend to see those real efficiency gains kind of come through, right? Because the data centers we're building today, they're where the power is and the road of power is gonna be, but that power is gonna sit there and that power will be relatively consistent for a very long period of time. You know, ultimately, like tens of generations of chips will probably come through those data centers and, you know, think about the amount of compute you're gonna be able to get outta that data center 10 years from now with that same power envelope versus the, you know, amount of compute you're able to get it out to out of it today.

Um, that, that's gonna be a, a tremendous increase. And I think, you know, right now, early in the cycle, little bit harder to see the light at the end of that tunnel. But I think once we get a little bit more mature and that efficiency starts to really compound, that's kind of what we'll get out of this, you know, spiral that we're in right now And, and all, all of this.

Because, because now we finally can, you know, I like looking for patterns that, that are, they've always been there. They're sort of inevitable to fill in. And then whether it's cybersecurity, the internet writ large, I mean, Kate, you've touched on one of the key things, and when you said that the, the resonance, like literally Darlene, again, I told you this one, right?

Um, data sovereignty, because to date, I mean, if you're, most companies, much less small companies have 'em help you, the millions, tens of millions of micro businesses, your data is out there with somebody else. But it doesn't have to be because Moore's Law, blah, blah, blah. You know, the, the amount of data, you know, you actually care about as an organization, as a person, whatnot, tend not to be that much.

And no, it doesn't have to be in the cloud, and it doesn't have to be somewhere else. You could actually have it all right here because storage is cheap and compute is cheap. So you follow these curves up and you find us doing things because we finally can't.

Not because it's radical, it's because the way we've been doing things aren't really the way we wanted to. We had no choice. You know, 1999, 2000, 2001, I'm, I'm out in Colorado, near Boulder.

I'll, I'll give you some names. Interlochen. Anybody ever been to Interlochen, Colorado, right outside of Boulder, right?

Uh, I forgot the hotel there where they have some good, the Omni the Omni Hotel. They had some good conferences there. com company.

And when you got to Interlochen, it's in the foothills. And as far as the eye can see, you saw two and three story buildings with companies with names like level three. And, and, uh, was it Quest, be Pack Bell West became Quest or something?

They, they had a few names, eventually Century Telecom or something. And there were, there were all of these fiber carriers, and then there were, there was the storage companies, mixed storage, storage, tech storage, this store that, as far as the eye can see, because we were building out data centers all over the place. This is when Northern Virginia became the data center capital of the world.

We had a beautiful one over in Tysons, you know, and it was the same thing. We've gotta build storage centers. And back then it wasn't, power wasn't the limit, it was bandwidth.

We couldn't, we couldn't, you know, you could build the biggest g*****n machines you could with big racks, but if you only had a straw to send the data through, it wasn't really doing you much good. And, and we had these same conversations, same conversations we're having now, and then the bubble burst. So it Took 10 years.

I would go for so far to say though, like, if I look at the last decade, hardware was kind of boring commodity. And then GPUs came along The last decade, the last 20 years, 25 years, when was the last time hardware wasn't boring, But, but I'm surmising or at least postulating that hardware might be cool. Again, as I look forward to see what people are starting working on and all these engineering things that people are talking about, for example, what Fujitsu was saying, that stuff's all gonna be crucial.

So I don't know damn hardware cool again or what Hardware's very much cool again. I, I think it became cool again, probably during the COVID pandemic when everybody realized how critical chips were to pretty much everything you make, you had a hundred thousand dollars cars sitting on the lot, the lot 'cause of a 10 cent by per controller. Um, and I think it's, you know, it's become more mainstream, you know, ever since.

But I think you're right with, there's a ton of innovation happening, right? You know, whether it's, um, you know, some of the, you know, other instruction sets like R five, um, you know, new IP out there on the compute side, you've got in-memory compute, you've got analog AI techniques which use extremely low power relative to the type of stuff we're seeing today. There's a lot of innovation that's in been incubating during this, you know, unsexy period for hardware.

And I think it's gonna have its stay sometime in the next, you know, next decade or so. All right. I was, I was, I was just going, I was just going for cool there.

I didn't go all the way to sexy, but, okay. I think even back to a hardware. Cool.

Cool. Now because you know, it, you were talking about chips and, and you know, to be clear, almost all chips are sitting there doing nothing almost all the time because our systems, we just, you know, we've overloaded the hardware because we had no choice. But now we we're smart enough and we can make complex enough systems that go back and use the old hardware to do things.

You, you know, you could never imagine. We're doing it now. It's fun.

Good stuff. Hey, we we're a little over on this segment. I'm gonna need to pull the plug on this one.

Let's take a quick break here on the gang, and we're going to come back and as I promised you at the outset, data center's on fire. Wow. You're watching Text on Gang.

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. You 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 cloak, digital executive protection, defending the new attack surface your personal life. Hey folks, we're back in.

Yes, is Alan alluded to. We're having an issue with fires and data centers. It turns out that there are lithium batteries that tend to blow up and systems get overheated.

And well, maybe the most important thing that any AI company can do is make sure there's a firehouse nearby. But Chris, does any of this kind surprise you? And are we gonna see any more of this stuff before maybe it gets better?

I don't know. You know, the metaphor kind of writes itself. You know, AI just doesn't burn cash and carbon anymore.

It'll burn your actual wrecks. Um, it, it, it comes, you know, as we said in the last segment, you know, it, it comes back to these issues we keep talking about. You know, the demand is real, you know, and the demand is real.

The market and cash will, will land and market and cash lands. Heck we'll build things. And is the, are the actual, is the infrastructure altogether.

We're sort of, well, well, al al you know, my, you know, my boating experience. Like let's build boats while you're sailing 'em and, and so on and so forth. See what happens.

That's what we're doing. So yeah, sometimes things are gonna burn down, right? You know, when your ambition is melting your racks, right?

That's just, you know, hubris with a warranty, but, uh, hubris with a warranty. Yeah. Nothing particularly surprising here.

You know, look at it as indicators, right? You know, as a, a individual, you know, fire event is always a bad, bad thing. But looking nationally, globally, you know, we're building 8 billion of these.

Some of them pop because heck, there's no infrastructure there. And we didn't think this through very well. Maybe there's steering lessons.

How many of you have actually managed a data center or I worked for a company managing data center. So I, I have too, one of some of my funnest times was checking out the fire systems, right? The fire suppression systems, what's the stuff they use?

It's not, you know, it takes the oxygen out of the rooms. It is a chemical halo, right? I thought it was, it's halo.

You know, I thought we've licked this problem already. Apparently not. Because I think part of the issue is the lithium batteries that are being put into the backup systems are at the core of the issue.

I think, Dan, Yeah, no, I think you're right, Mike. It's the batteries. We've produced these, you know, really, really difficult Andre devices in as part of the backup storage here.

And the big takeaway for me is, you know, fire safety comes in. It's an architectural decision. When you build a data center, you don't, you don't plan for a fire by having, you know, brave souls show up on a truck like this is truly built into the design, right?

You know, compartmentalizing, you know, keeping servers further away from batteries, having fireproof barriers between them. I mean, it's really built into the industrial controls of the data center itself. Um, but, you know, beyond just the, you know, the prevention of fires for, you know, all of the, the risk it poses the, you know, health and human safety and all that, but the economic impact that these things going down is tremendous.

I mean, it reminds me of similarly when a, a semiconductor fat goes, you know, as the firing goes offline, I mean, it's hundreds of millions, billions, maybe tens of billions of dollars in, in, you know, lost opportunity for these things to come off, right? I mean, you know, what Jenssen's quantified, uh, you know, a gigawatt, uh, data center is, you know, roughly equivalent to like, you know, $50 billion in value. Um, in terms of, you know, the amount of compute capacity you can sell out of it.

Um, these things coming offline is not only gonna be really economically damaging, but you know, a lot of mission critical, you know, critical services for, you know, our economy and our society. You're gonna be running in these things. And if you add a massive data center come offline, that's gonna ripple through and it, you know, it impact everyday people, um, at a certain level.

So I, I think it's definitely a topic that we gotta take seriously. And, you know, the right time to think about this stuff is when reporting concrete and doing the building, uh, because it truly does have to come down to an architectural decision to make these things as safe as possible. And, and you bring up a good point, Dan, and that is, um, you know, it is actually the potential of costing lives becomes a part of this equation as well.

And something that we, we do need to take seriously. I mean, is it time to update the fire coats? You know, this is, again, is not a new problem that we've dealt with at in societal levels.

It's the same way after Hurricane Andrew. They changed the building codes down here in South Florida, or in Florida in general. Maybe we need to re-look at the, uh, fire regs.

And, and Dan, to your point, you know, what, what do you need to get a c of o certificate of occupancy to get, you know, an approval here? What, what fire suppression systems do you have? What architectural decisions do you have?

You know, this, this seems, hey, maybe AI can help us with this, right? Reasonable to revisit it. I'm certainly no fire code expert, but, you know, my concern is it does feel like, you know, actually at a a federal level, and even at a state level, there's a lot of deregulation and kind of squirting around some of the, you know, processes and procedures we have around this stuff.

Um, you know, really, 'cause I think, you know, the knock on America is we can't build fast enough, you know, whether it's fast, it's data centers, Well, it'll take what, what, That's the way that we're actually deregulating and making it easier to do this stuff versus, you know, versus potentially harder by, you know, really kind of clamping down on, you know, rules, policies. What was the fire? Were the women burned in the Garin district?

Uh, in Chicago? I, I, the Linen it was, think it was the linen factory in New York, right? No, but there, there's a name, I don't remember the name, but it, it took one, the Triangle shirt factory.

The triangle shirt factory. Correct, Mike. It takes one of those and then people will, you know what Dan, they'll say, okay, maybe we shouldn't, we shouldn't cut those corners.

You know, maybe we should take that stuff seriously. Unfortunately, it may, it may take a tragedy like that. So a lot of the guys I grew up with are firemen, and they are frequently remind me, they say, you know, Mike, when your house is on fire, we're not coming to save your house.

We're coming to save your neighbor's house. 'cause your house is already toast. The data center's pretty much the same thing.

I mean, they're just gonna come there and try to contain it, but I don't think they're there to save it. I, I think if you're relying on the firemen, it's too late. I think you, you've gotta, you've gotta design for fire, if you will.

Right? So last, last week, last Wednesday, I was on the, uh, at, at sector, the, the, uh, black Hat Canada event. Uh, shout out to Ja, Jamie Arland was moderating the failed panel, and I got to be the old person on it, right?

You know, with two younger folks. And Jamie had brought props for us. I got to, to hold the 19 89, 9 and a half inch floppy that had the industrial control system, Dan from the Orfield water, uh, water system, uh, back in 1989.

And the folks next to me had a CD ro and a five and a quarter, three and a half. And we talked about, uh, the, are are we getting better? The classic con think of been 19 years they've been running this fail panel, right?

So it's literally long enough itself to, to be an example of, of what we're talking about. And, and I, I guess the point that I was trying to make there is the same one I guess generally make here, you know? And because we're moving along in a path, we're at this point, this is not new.

You know, everything from, you know, I think everybody here, you know, said parts of this in, in this segment. Um, we've been down this path before, you know, we're having a lot of things catch on fire. What do we know about fires?

Back in 89 with that disc was around, you know, I worked a, in, in, uh, South Carolina at a, a data general var, and we actually had the raised floor and the hay lawn system, and yeah, we kind of forgot about that and, you know, but, but it's not that we need to reinvent the wheel, like it's so many other things. We need to now, we can both do what we've already done before, put that back in place and connect things, uh, appropriately and fast enough to, to solve the, the current issue. And, and maybe as we cycle forward, you know, raise our qualities and have different levels of failure in the future, which is the whole goal.

Yeah. To fail more fun, right? But just not the same way over and over again.

Yeah. I always wanna be introduced to small modular nuclear reactors to power these. Like, we've also talked a lot about that.

Uh, but it's something Why, what could go wrong there? Things Always go Wrong. But I, but, but here's, here's the, the, I I don't wanna use the word irony, but it's ironic.

It's not a question of could we be designed them safer? Of course we could design them safer. I think the question is what Dan said, which is, do we have the appetite to design them safer if it means delaying putting them online?

Well, and that's the whole, we don't weigh everything together, right? You know, and fail failure is, is a wonderful thing, right? You know, like if you're not failing enough, then maybe it's really stable, that's fine.

Uh, but it's certainly not a growth area. And this is, you know, in data integrity of, of IAM there's so many different areas. I think we've been driving towards this.

How do we get to perfect? And we need to understand, stop trying that don't, there's no such thing, you know, there's, to your que question, Alan, it's a balance. If the benefits are X and the cost is y of data centers burn down.

Maybe let the data centers burn, but know you've made the choice and put it together. Sounds like a song burn, maybe bird. Yeah, that's what I was thinking.

I was like, okay, somebody who paid for that data center's gone. Yeah. Yeah.

You know, that data center. Tens of billions of dollars. Well, anyway, hey, you know what?

This is gonna be, I, I think an issue we're gonna see, unfortunately, uh, you know, until something bad happens and, and you know, it's what, what my mom used to say, funny till someone loses an eye or something, right? Mm-hmm. And, and that, that's what you got.

With that, let's take a break though on this one. We're gonna come back and let, let's turn over to managed insecurity. Discover Textron 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 back. And we're talking about, well, the number of instances involving security breaches that get traced back to some sort of third party provider of an IT service or security itself seems to keep climbing and folks are getting a little bit worried about that. But there's a converse to this argument.

Some folks would say that the providers of those services know more about security than the folks who run the internal IT TE systems. So you're better off relying more on those folks. But Alan, I know you put a post together over on digital CXO talking about this in terms of the risk and how we should be thinking about this.

But where do you fall in this debate? 'cause it kind of sounds like you're saying don't rely on the third party. So look, Kate, Chris, I'm sure you guys are gonna have a lot of thoughts on this one as well.

And, and remember, this wasn't on Security Boulevard. This is a digital CXO. 'cause it's, it's a leadership discussion.

Of course, you're gonna rely on a third party security's too damn hard, right? For, for, except for a handful of companies. Maybe the Fortune 50, maybe not even all of the Fortune 50.

Security is too hard to do by yourself. Almost by definition, you're gonna have third party providers helping you. We live in a cloud world where we have, by definition, seated some of our, uh, capability for security to the cloud providers 'cause they control the stack.

But the real point of my article is whether you are outsourcing, Taylor, I'm hearing a echo. The real point in my article is whether you are outsourcing your security or giving, sharing it with a cloud provider or some third party MSSP or, or an AC provider who you give access to your network, as in the case, I think it was a target, right? That led to the target breach.

You can outsource the work. You can't outsource the responsibility when the stuff hits the fan and your customers are impacted. They don't want to hear, it was one of our partners, they don't want to hear it was a third party.

It's not our fault. The the buck stops on your desk. And so you gotta remember that whenever you bring in a third party who is in any way remotely responsible for your security posture, for your compliance, I'm not saying don't do it, but you've gotta be thoughtful and, and kind of follow best practices and make you making sure those third parties have, are doing what you think they should be doing and what you feel is reasonable and managing that risk.

Because when the, when the music stops and you don't have a chair, you are the one in trouble. You can outsource operations, but not accountability. And if you think about, um, you know, hey, we lived in Florida for 20 years.

I'm sure you know Chris and, and Alan, you remember, um, the, the deep, um, the water oil, the deep water oil issue, and it was, you know, bp, but, you know, who were they blaming? They were blaming Halliburton and, uh, Transocean and it, but at the end of the day, it came back to bp. So we see that over and over again, right?

That, you know, you gotta, it's gonna go to the brand at the end of the day. And we see that. Absolutely.

So, I, I'm gonna, I am going to, uh, bring the, uh, late grade Admiral gr amazing Grace Hopper into this, you know, because she was real admiral. And and for those of you who don't know the name, read a book, but, uh, she, uh, in the Smithsonian, there's a log book with a beetle taped into it, you know, that was caught in a, in a relay in a computer. And she said, found bug.

So if you talk about bugs, you know, amazing grace. And she's, you know, famous for having all sorts of great, uh, sayings. And what I've always loved as a sailor myself is sail.

You know, a a ship port is safe, but that's not what ships are for. Sail out to sea and do new things. And to be clear, she's not just, that's, that's an analogy.

She's an admiral. And as I was billing the, the, the Twains, my two silly solar power boats and sailing 'em up, uh, from the keys to St. Augustine, remember explaining this to some of the, uh, uh, uh, the Coast Guard folks in the water.

You know, to be clear, my lifeboat is another boat, and both lifeboats have lifeboats, right? So, you know, you know, you, you guys may have to come get me someday, but to be clear, I'll just be standing here. I'm not gonna be swept out to sea.

And the nice thing about that sort of thing is what, what, what, uh, Grace Hopper was meaning is that take responsibility and, you know, there's nothing else after that. There's not take responsibility, just take responsibility. And then whatever happens you're responsible for.

So yeah. Whether you're literally sailing out to sea or whether you're, you know, building an oil rate deep horizon, right? Uh, Kate and or running a data center or anything else is your responsibility.

So you can outsource your cybersecurity. You know, Alan, your intro is exactly right. It's all very complicated stuff.

You shouldn't be doing it unless you are a security company. But if that is your one link, if you decide to have one fragile link, you know, again, have a lifeboat, have a second boat, have a second. It depends.

But you've decided to have one brittle link, and that's your, your choice. So Chris, I I hear redundancy in five nines. Yeah.

Redundancy to your redundancy. Well, when it's life critical, right? You know, I literally Go back up.

I love it. Yeah. Right.

You Know, at worst case, on those boats, I'll be, I'll be in there in the ocean with an anchor, you know, floating, but the storm passes and I'm still there, right? So yeah, you gotta think this all the way through where it's your fault, period. I wanna think a little different approach here.

I, I think what we're ultimately saying is you can't outsource the business impact, right? Ultimately, something goes wrong, it's your business that will impact everything in between is just the blame game, right? And, you know, whether you have security totally inside, how many CEOs have survived an incident by throwing their CSO under the bus, right?

That's a short shelf fly job if you ever seen, right? So, you know, you're running the business as a CEO. Ultimately, something goes wrong.

It is your business gonna be impacted? The rest is just the blame gate. Do you wanna, do you wanna have a third party outside?

You can blame, all right. You know, BP probably got off, uh, a little bit easier because Halliburton took some of that blame for them, right? Kate, you know, to go back to your analogy, um, but, you know, end of the day it's your business at the end of the day that's gonna be impacted.

And, you know, whether you, you, you decide to do it yourself or you outsource it, you know, that's really much more a matter of cost, risk mitigation and skillset. The term outsourcing is getting more nuanced though, because what a lot of the folks are doing these days is they're calling them co-managed services, and it's a, a better blend of an external service provider and an internal team. So everybody's kind of got skin in the game, but it's not like I'm dumping all of it onto one provider and hoping for the best.

I think that's gonna be the better approach, and frankly, it's taken a long time to get there. Much to my amazement. No, that's a good point, Mike.

And, you know, co-managed, um, models with shared telemetry and clear SLAs, um, are the future. So it, it has taken though way too long. Yeah, go ahead.

That's part of a good resiliency strategy. Yeah. You know what, when, when CrowdStrike had the incident, which is just about a year ago now, I believe, right?

Um, the blue screen of death and Delta went down. I didn't see all those Delta passengers in the airport saying, God, Don CrowdStrike. No, they didn't.

They blamed Delta. Delta Tried as much as they could to get Everything they did. They sued him.

I, I don't know what happened with that lawsuit, but they did Sue CrowdStrike, but people were p****d at Delta. That's that, it's your business. And when you outsource you, Kate, you said it, I think in the beginning, you could outsource what not res uh, capability, but not outsource Operations, but not accountability.

Not accountability. That's, that's the bottom line here. And, and if you're a digital leader out there, I'm not saying do all your security in-house, you probably be more danger to yourself doing that.

But you've gotta recognize that you still are accountable to your customer. It's still your business. And, And the ironic, wonderful thing about it is once you do that, it turns out you get more business.

Right? You know, and, and this, I'm sure I, uh, I'm old. I repeat my stories all the time, so forgive me, but, right.

You know, in a vvc early VC life, for me, giving a pitch, um, after the, the, the Cisco firewall thing, um, one of the inve investors present, leaning forward and saying that, that, you know, billion dollar, uh, uh, firewall thing, how'd you do that? And I just honestly said, well, we built, built a team that died for each other. We took responsibility with the market and built trust.

And this, this individual leaned forward and said, no, you can tell me. And it key. So the key to what we did, literally, I think, uh, Danny said it, uh, somebody said it a minute ago, we had this horrible thing happen that was totally our fault.

So instead of saying, oh my God, you know, we immediately went out and said, we're gonna go, I'm Chris Blas. I'm responsible. It's our fault.

It's as bad as it looks. It's in fact worse, and we're going to do everything we can to fix it. And here's our plan, and we did it.

And that actually works. You people trust you. They give you more money, they give you more trust, as opposed to saying, you know, the whole blame game thing.

We all look at the, the, the looks on our face when we bring that up. It's like every human being goes, oh my God, it's blame game. But if you could actually develop trust, that's business.

That's money, that's efficiency. But you have to actually mean it. Oddly enough, you know, your, your grandmother told you these things, right?

Remember back to when you were a kid? It's true. You know, I think George Kurtz, so I know George a very long time, probably 25 years.

Uh, George Kurtz is the CEO founder of CrowdStrike. George did, that's exactly what George did. He went before Congress.

He went before the cameras. He went before everyone, black hat. He got on the stage of black hat and said, yeah, we, we screwed up.

We messed up. And here's what we're doing to make sure that don't happen again. And, and I think that is better than playing the blame game.

You're, you're right, Chris. That's how you build trust. That's how you build a business.

To, to be fair, Microsoft had a hand in this, but they blamed CrowdStrike too, so, well, I'm not gonna throw dirt on them, but yes. But Microsoft's been blaming everyone for this, for their security issues for a long, long time, unfortunately. But, but they had their moment too, right?

There was the time where Bill Gates stood up and said, security's gonna be important. We're doing trustworthy computing. They hired the best security people in the world.

And there was that period of time from maybe 2005 to 2010 or 12, where Microsoft was stellar, stellar security. And now, And, and now the president of Microsoft shows up in front of Congress once a year and does a Meia coppa, and that's that. Send Hail Mary's make and like, make a donation at the box on the way out.

You know, that's, that's security today. But anyway, it is, it's important for leaders out there. It's something as security people I know we, we deal with all the time because we know, we, we don't, you know, look, if you are using a cloud, you're, you're, you, you're outsourcing security to a certain extent.

Absolutely. There's no way around it. So anyway, gang, what a terrific, terrific discussion today.

Of course, we had our double dose of ai, a little security thrown in. Doesn't get better than that. Uh, as usual, we have a full Textron TV show following our Textron gang here today.

If you're watching it during our stream, if not, whether you're watching it on O TT or Textron tv, or text on tv, YouTube channel, thank you for watching. We have a lot more content out there. Go check some of it out.

Mike, good luck to you in, uh, Barcelona. And then of course you'll be at Amsterdam. You'll be, I, I'm sure, recording videos and filing stories like mad, uh, Chris, Kate, great seeing you.

We'll see you on a gang soon. Dando, I'll be talking to you, I guess, in a little bit, hopefully, but keep doing what you're doing. Enjoy your day, everyone.

Stay tuned for Text Drunk tv. This is Alan Shemel, we're out. Hey, everyone.

Welcome back here to Techstrong tv. I'm really happy to have our next guest on. Let me introduce you to Christian Rig Rodriguez.

Christian is a CTO over at CrowdStrike. When he is not surfing, it looks like, Hey, Christian, how are you, man? I'm going, I'm doing great, Alan.

You know, thanks for having me on. So I always like to find out a little bit about and share with our audience a little bit about, talk to us about the boards. Oh, so, and I just got into surfing, uh, last year.

My fiance's a big surfer. She's kind of surfed all over the globe. And, um, I basically am at this point now where I'm learning to not drown, uh, and also look fancy during that process.

And so I've had the, the, the pleasure of going, uh, through like Central America and a few parts in on the East coast, do some surfing, getting used to, you know, standing up and, um, you know, not embarrassing myself as I get better at it. But, uh, yeah, that's, that's, Yeah, no style Style absolutely counts in surfing, so It Sure does. It absolutely does.

Yeah. No, good for you, man. It enjoyed, it's a great hobby.

Yeah. Um, so Christian, when you're not surfing, and before you were CTO at CrowdStrike, talk to us a little bit about kinda your, you know, give us the Christian Rodriguez story. Oh, Wow.

I, we may need another hour. Um, and so, um, you know, not we got time not including all the, the, the, the inner child work and meditation, um mm-hmm. But, but, uh, haven't we All, Don't we all Exactly.

Uh, I've been with CrowdStrike for almost 11 years. I started here as a, um, sales engineer, probably part of the first 120 employees when we, we basically had like three offerings of the company from EDR to intelligence and, and hunting. And, you know, now we're, we're, we're, there's 30 plus modules, but, um, but my background even prior to coming to CrowdStrike is big in the solution architect slash systems engineer, sales engineer roles.

I've worked for the likes of, of Websense and Fishnet and Zimperium. Uh, so roughly 20, 21 or 22 years in the industry, you know, working for sure for cyber companies in some capacity. Yeah.

Gary's an old friend of mine as well. Oh yeah, I know Gary. Yeah, he's good people.

Yeah. Yeah, absolutely. One time we've given a chance off camera, my first time to Kansas City to fishnet way earlier.

This is early, early two thousands. And the first time I met Gary Fish in Jody Zel was the CTO that is now CEO and Fire Mark. But yeah, all, all good people.

Yeah, great people. Um, so you, you've got some real, you know, cyber, cyber creds as they say. How long you been CTO over at CrowdStrike?

So, Uh, field CTO of the Americas for roughly three years. Um, where, um, I feel very fortunate this, this role is, I feel, um, I just love doing what I'm doing, right? I get to, to understand the way we think of problem solving, but then I also get to, uh, work with all of these enterprises on understanding their vision and how they plan on maturing different disciplines in their security practices and, you know, how they think of problem solving as well.

And then figuring out how we align our strategy with their strategy and their vision with our vision. And, um, it's just a really great opportunity to just speak with a very wide array of different customer types. Uh, and then also just evangelize what we're doing here at CrowdStrike and what we're seeing in terms of new threats and, you know, what, what, what thought leadership is based upon customer feedback and how we're building new things.

So Very cool. Very cool indeed. Um, you mentioned when you first came on to CrowdStrike, you know, they had three modules.

Yeah. Yeah. And they've got 30.

Yeah. Yeah. There's a little something for everyone, it seems.

Yeah, that's right. When it comes to security and CrowdStrike. But you know, if we, if we, we don't have to go through all 30 modules 'cause we really will be here an hour, but, you know, Christian, gimme like big garbage cans, if you will.

Yeah. You know, break it into big cans Yeah. The areas.

Yeah, I think, I think there, they're really, um, man, it's, there's some, there I would say, uh, four to five kind of major pillars, if you will. Um, you know, naturally we, we, the flagship capability of the, of the company or the nucleus, if you will, of the company's success has been around the endpoint, right? So we, we do endpoint security, uh, extremely well, right?

And that's, you know, the EDR, the next gen av, anything that we can see on the endpoint has been kind of the big, you know, thing for CrowdStrike for since our inception. And then we got into identity as well, doing identity and authentication analysis and behavioral analysis on who's logging into what, why are they doing that, you know, what do we do to prevent something that could be bad? And then there's cloud kind of the third pillar, if you will, uh, where we can understand, um, ephemeral workloads or the configurations around your systems and the likes of any of the major CSPs, like AWS or, or Azure, uh, or GCP or OCI.

Um, and then there's also even the, the, the SaaS side of the house, right? Where we, we, we plug into doing SaaS analysis and authorization of analysis across SaaS applications. Uh, and then even if you were to think about, you know, those kind of four pillars, then there's this AI model that sits on top of all that very high fidelity to telemetry that we're capturing where we can start to help our customers augment their stock efforts by having AI help them with decision making or hunting, or, you know, just doing analysis and triage.

Excellent. I think that was a nice way of kinda compartmentalizing into some people wrap their heads around, man. Yeah, absolutely.

Yeah. All right. com if anybody wants to go to the website and explore all 30 modules and Yeah.

And jump into all that. There's lot, There's lots of brows. The website.

Yeah, there is. It'll keep you busy while, yeah. Hey, look, it's job, it's job security for someone who does the website, right?

Totally. Um, well, let's AI replaces them soon. Who the heck knows.

Yeah, that's whole nother, that's another podcast episode. Get into that if you want. Yeah, I was gonna say, Alan, that i'd, that's a podcast Episode where We, We'll, we'll, you know, we'll, we'll commiserate Exactly.

I mean, like keeps me busy just talking and writing about that particular subject. But anyway, you know, well, I'm not going to get into it now, but I'm doing a thing right after this. You know, s stands, uh, what is it?

No, the s in vibe coding stands for security. Oh, that's funny. And there is no s But anyway, um, let, let's talk, let's talk cloud security, though.

You're a big advocate, right? For how do we put this unwieldy monster to heal? Yeah.

Let, let's hear your thoughts on this, Christian. Um, every, every enter enterprise, every organization we've met with has something in the cloud, right? I don't think, you know, cloud, cloud solutions aren't novel per se.

Um, people are moving applications into the cloud or moving infrastructure into the cloud, or they're, um, subscribing to cloud services to make their lives easier in terms of spinning something up. Uh, and what we're seeing is that, uh, the more that enterprises move to cloud, right, the better adversaries acclimate themselves with, you know, what's in these cloud services. We, we, you know, Alan, I'm sure you're familiar with the way that we track the bad guys, the adversaries that are out there responsible for the attacks.

We're seeing, you know, ECR groups, nation state activist groups, for example. Um, and what we're seeing is that these adversaries are becoming ever very cloud conscious. So they understand how to navigate the control planes, and they understand the services that power, the fact that someone spun something up or workload or a container.

And so what we're seeing is that enterprises are having, you know, challenges defending against those attacks that are very cloud focused because they're also managing disparate tools. Or they may have a tool that's very much focused on endpoint and a tool that's focused on, again, the authentication mechanism that is GI giving access to that cloud environment. And then they're managing all of these different da disparate tools that lack context, or there's a lack of cohesion.

But adversaries are really using the cloud as a pivot point even to move back on premise. And so, you know, what we're seeing is that, uh, adversaries are very proficient in understanding those cloud services, and they're taking advantage of things like misconfigurations or they're taking advantage of systems that have been orphaned in these cloud services. And as a result, they're very successful with respect to how often they can actually successfully infiltrate an organization or enterprise because of those services that are misconfigured or, you know, IM policies that have been abandoned or have been excessive.

And that's, that's essentially a major area for us investment wise, even just to give our customers better visibility and better protection and consistent protection, I should say. Right? That analyst experience is extremely important in terms of consistency of dealing with a threat, whether it's on premise or if it's virtual or it's a hardware based box, or if it's an informal system, should be fairly consistent.

So Christian, I, you know, I've been in security 30 something years. 30 years, right? When I, I founded a few companies and helped found a few companies.

George Kurtz was at a company called Foundstone. Yeah. When I first got into security, and I knew George, we had a vulnerability management tool as well.

This has been a holy grail for a long time. Yeah, Yeah, Yeah. Right?

And, and it's funny 'cause when you look outside of security, right? com too, right? com, a new site cover platform engineering.

You, you, you look at all these other areas and the move to a platform is kind of the natural evolution, right? Mm-hmm. Instead of doing point solutions mm-hmm.

For developers, we, we develop an internal developer platform, an ITP, and we have platform engineers that lay this all out. And we have cloud native engineers that, you know, even there, like how we, 'cause I mean, Kubernetes was the hardest thing, one of the hardest things I've ever seen to use. How the hell it caught on?

I don't know, but, but it did, right? And everyone uses it, it seems so, but we have cloud native platforms. We've been trying to do unified security platforms.

Mm-hmm. For as long as I've been doing security. Why now?

Yeah, I think you're right. It's, it's not, uh, the concept isn't novel, right? To say the least.

Um, I think everyone has been trying to, uh, bring uniformity and consistency with, uh, a lot of disparate data sources. And I think what we're seeing is there have been, I, it's less of a trend and I think it's more of a necessity now. I think it's, you know, if you're asking why now, I think it's because it's very expensive to touch all of these disparate systems.

Uh, and it's also very risky, right? There's so many things that can happen in between those seams of these tools that people are, and organizations trying to stitch together, right? There's so many things that can fall through those little cracks and those, you know, an adversary only needs to be right one time, right.

In order to be successful. And so I think it, it comes down to the concept of the appetite for risk has been reduced significantly because the cost of a breach is also skyrocketing. And so, you know, companies can't afford to have these disparate systems that are, you know, poorly sits together, you know, act as their, their bedrock for security, right?

I think platform and platformization in grand scheme of, you know, what platform truly is platform is, you know, native capabilities that can show first party data seamlessly connected to together, right? So that that opportunity for the adversary to live in those cracks of those disparate tools now is going away, right? And we're be, we're making it harder for the adversary to become successful because everything is so well put together in a, in a true platform.

So I think it's less about a trend and it's more of around how do I ensure the, the future of my enterprise and the safety of my enterprise. And it's ensuring that, um, in the consolidation story, if you will, is, is, is more tied to how do I do my job as a, as a defender more effectively and more efficiently, uh, versus dealing with these disjointed tools. So, Christian, I, I think all that true.

Mm-hmm. But let me, let me put something else forth to you. We need AI to make this work.

'cause the, this coordination, this tightening up of, you know, it's like, you know, the SR 71 Blackbird plane you've seen pictures of, that's Aly badass looking plane, isn't it? Beautiful? You know, that used to leak fuel when it was first taken off or on the ground, it needed to go a certain height and speed and heat would build up that would fuse the, the metal there to make it air tight or fuel tight so the fuel wouldn't leak out.

AI is, it's the same kind of thing. We need AI to get that tightness right? So that we don't get the leaks out the little.

'cause it only takes, as you said, it only takes one time. Yeah. Only one time.

And it's only one crack. Absolutely. Yeah.

And that's all they need. But with ai, I feel like, you know, maybe we finally have the, the putty Yeah. To, to, to seal all these things and bring it together and make it work in real time.

'cause that's another piece of it, right? Oh, absolutely. We need this to work in real time.

Absolutely. And I, ai I think is so crucial to our success as defenders, right? I think what we've seen is, and, you know, data, um, if you were to ask me even like way, what, what is CrowdStrike, you know, today versus what we were, when I started here, we were very much talking about endpoint security then, but we were very much focused on how do we do more with this telemetry right?

On the endpoint? And then how do we expand and open up that aperture so we see more things. And now again, right.

30 modules later, right? We're seeing everything on, again, cloud identities, endpoint SaaS, I could go on even third party ingestion. Um, and so in order to get your arms around around the problem that's growing, right?

Data is the problem. Your, your business. So, so that's, that was CrowdStrike, right?

Like we, we opened up the aperture. We're collecting all this telemetry as a business and as an enterprise, you're also growing and your problems are becoming bigger. The more you start expanding on data and the more that you start to monetize even data, right?

And as your business. And so what happens is AI helps you get your arms around the problem. 'cause you can use AI to reduce remedial tasks, right?

You can augment your, so analysts and your defenders, you can start to assess large, uh, you know, areas of data to start making more sense of it, right? And you can start using that data to, to start getting answers faster. And I think that's really where on our side, AI plays a big role on augmenting your efforts as a SOC analyst, as someone that's looking into vulnerability management, someone, um, analyzing identities and authentication requests.

And then guiding you into what is gonna be the next best step for our organization based upon trends that we've seen that mimic this type of behavior or trade craft. Or what does a true outlier look like in my business based upon poorly written applications versus an adversary that has his, his or her hands on a system and they're removing laterally or dumping credentials or, you know, being persistent. And so I think AI allows you to start expanding your defensive measures in a similar fashion or a rate that is also matching what the adversaries are doing.

And also keeping up with the way that your business is growing. So you could probably spend another hour talking about AI use cases, but AI for us is expanding on allowing defenders to, to extend themselves. And, and there's a lot that, that, that we're, we're gonna announce soon around the AI capabilities in our platform.

But I, I think it's pretty, you know, it's pretty fun stuff. It's pretty geeky stuff. It is, it is.

Let me ask you a hard question about the AI, though. 'cause you know, when you first started talking about, we, when we first started talking, you mentioned, you know, and isn't it great to have sort of a personal advisor, a analyst to talk to and explain this to you? Is AI going to be that analyst soon?

Or do you think it's a person who's augmented by ai? I'd say for right now, it's a person that's augmented by ai. Uh, and the reason is, you know, your business is growing and processes change and adversary tradecraft evolves.

And, um, the, the AI model is as good as the data that it has a access to. And, you know, we, for example, at Crosscheck, we've trained our models based upon what human analysts are doing to triage systems and remediate systems. And so, you know, I think every, you know, you may have a business that has a, a SOC analyst that needs to come up with an answer faster, right?

And we don't necessarily plan on replacing that analyst, but if I can save that analyst five to 10 minutes for every detection right? Or event that they're analyzing, that adds up if we're talking about thousands of events, right? Right.

And so our goal is to have ai, uh, quite frankly, reduce things like alert fatigue, right? Or this platform in as a whole reduces alert fatigue because it's showing you contextually the anatomy of an attack from start to finish, from the identity side to the endpoint side, to the cloud side, to the SaaS side, for example. And then AI is basically guiding you on, these are the things you may wanna hunt for, or this is how this event was triaged, or let's start automating some of those tasks for you.

And having that human analyst validate that the AI is doing its job properly. So I think for right now it's augmenting, you know, there may be a feature state where AI's doing a lot of the work on behalf of that person, but that person would still be involved in some, some capacity. All right, we'll, we'll see.

Yeah, we'll see. Um, we'll see. Hey, Christian, I, you know, we're coming up in the black hat season.

I know CrowdStrike's out there totally random, but for folks watching this who maybe are going to Black Hat I'll, what can they expect from CrowdStrike? Yeah. Uh, I'll be there at, at Black Hat with, with the team.

Um, I think we're doubling down on showing some really great innovation coming out of, uh, our, our Charlotte AI models. Uh, so really excited to showcase some of those capabilities like our detection, triage, and our, uh, uh, our, our, our hunting, uh, capabilities and our workflow capabilities. Um, you know, there's a lot that we're doing in cloud security as well, right?

This ability to do, you know, runtime analysis and protection on, on of systems that are, again, workflows that are spun up in these cloud services. I think that's another big component. And I think we're, we're really doubling down on, uh, you know, what our customers are asking us for.

And it's just to protect, you know, everything, right? In terms of, you know, runtime and the, the, the configuration settings that you have across those cloud services or, you know, those AI workloads. A lot of our customers are asking us, like, how do we get better, better visibility into, um, AI workloads within our CI ICD pipeline?

Who's spinning up what instance? And is that going to lend itself to some type of new risk or some type of data loss? Uh, and so we're doing a lot on the AI SPM front for security posture management, and we're, we're excited to showcase some of that at, at Black Hat as well.

Cool. If you go into Black Hat, go check out the CrowdStrike booth. Good stuff going on.

Hey Christian, I want to thank you for coming here on Tech Trunk TV and, and talking a bit. Yeah, you got it. Come back.

Visit us soon. If we see out in Vegas, we'll we'll catch up there, but if not, back here on tv. Okay.

You got it. Looking forward to it. Yeah.

Alright, man. Ride good waves. Thanks so much.

Thanks Alan. Alright. Christian Rodriguez, CTO CrowdStrike here on Tech Trunk tv.

We're gonna take a break and we'll be back. Hey everyone, it's Alan Shilo. We're back here on Tech Drunk TV in beautiful Napa Valley at the Jfr Swamp Up event, continuing our Day two coverage.

There's a little bit of a break going on, you can't see, but out there people are eating ice cream and peanuts and potato chips. It's a little mid-afternoon break, but we're still here 'cause we've got a lot more to bring you. Let me introduce you to our next guest.

If you've been watching Tech Drunk TV over the years and any of our coverage of Jfr, you already know him, but I'm gonna pronounce his name right for the first time. 'cause it seems my pronunciation is a little old fashioned. So let me introduce you to OV Laman.

Perfect. Nice To meet you, Yoav. It's good to see you.

Yoav, of course, is a co-founder, one of the co-founders and CTO here at Jfr. Yoav a pleasure. How are you, Ben?

Likewise, Busy. Lots of announcements. This warm up.

Probably The most we're On that we have, right? Uh, lots have good, good feedback from customers and, uh, also some suggestions. So, uh, And that feedback's, That's the goal, actually.

You know, what they say, the feedback from this year's Swamp Up will be in the products for the, Hopefully even well Before ai. We have to. So, yo, we were talking, you know, before we got on, we have had a lot of people give us a piecemeal a piece here, a piece there, a piece there.

I'm gonna ask you, pull it all together for us, right? Give us the, a overview of all of these great announcements, all this great innovation mm-hmm. That was announced here at Swamp Up.

Okay. So I'll try to give the full umbrella of announcements that we made. So we started with Jeff Oak Fly and Fly is, uh, uh, our own disruption of the platform for, uh, a new agent, uh, repository based on our factory.

And that's, uh, that comes with, uh, a new user experience for managing software releases. Uh, so that was our first announcement. Then we went over to, uh, APPT Trusts.

And Abrus is, uh, the way to control your software supply chain based on three, uh, major concepts. First one is application that gives you ownership assignment for every release, every artifact, uh, in the JO platform. The second one is, uh, signed evidence.

And we announced, uh, partnership, uh, with many leading industry vendors, uh, such as GitHub, such as sonar, such as ServiceNow, uh, to, uh, uh, integrate their evidence, uh, into, uh, into the JO platform to accompany the the releases. And, uh, finally, uh, it's, uh, policies that, uh, allow you to use the information, uh, within the JO platform, use evidence in order to assign rules for the progression of the artifacts, uh, of lysis, uh, all the way towards, uh, production. Uh, so this is Apru.

It's a, it's a unified package that includes all these, uh, three main features, uh, ownership evidence and uh, uh, policies. Um, so that was, uh, another announcement. We also had a deep dive to our integration around evidence with, um, with GitHub to take the salsa provenance of GitHub, uh, workflow build and put them alongside artifacts in the JO platform, uh, as evidence, which is, when you come to think about it, it's the logical thing because, uh, you, it makes sure that, um, that the evidence itself is bound to the artifact and you can never get out of think.

And it's also the fact that Artifactory is the entity that is exposed through your production. So that's, uh, one thing that where we did a deep dive of UPT trusts. And the, uh, other thing is we announced on stage and integration with ServiceNow, uh, around UPT trusts as a, as a full, as a, as a whole.

Uh, and we show the, uh, synergy between applications that many of our customers are already managing in ServiceNow, and how change requests in ServiceNow are going to be, uh, reflected as evidence in, in, uh, in JO uh, and vice versa, how you can move between the platforms. So that was, uh, also, uh, part of the big announcement of apta. Yes.

Then, so it's a mouthful. Then, uh, we moved to, uh, uh, a new announcement, which is, uh, around machine learning and ai. This is AI catalog.

Yes. And AI catalogs, uh, allows you to have governance over, uh, models that are packages, but also models that are, uh, uh, sa like, uh, an and open AI and so on. Uh, under a single platform, you have a catalog where you can find the latest versions of the models and the metadata about them, like, uh, the security status, the, the licensing and, and, um, other metrics that have to do with the model health.

And then, uh, similar to what we have, uh, uh, in curation, uh, we elevated the same features for machine learning. So you can allow different teams to use different type of models. Um, for instance, you may allow a research team to use deep, but you never want to see that, uh, in a production facing, uh, uh, release.

And part of that when it comes to, uh, to SaaS models is, uh, also being a gateway between you and the SaaS models. So if you, for instance, if you're using open ai, uh, you will use it through the GO platform, and that allows you to have governance also over these type of models. Uh, so, so this is, uh, this is the gist about, uh, AI catalog.

Mm-hmm. And then we went into a bunch of security related, uh, announcement. I think I, I can mention two, uh, highlights there.

The first one is the support for ID extensions. Yes. Uh, and, um, basically it's a combination of artifactory acting as a proxy for your, uh, vs code extensions.

So we start with VS code, we will extend it to other IDs and, uh, curation allowing you, uh, to, uh, to, to control the, the, the, the, the extensions that your developers are able to install on their endpoints. And this is one of the most dangerous and overlooked, uh, risk that developers are, uh, currently facing because you basically install a software on your, on from the internet that everyone knows that it's wrong, but, uh, for some reason with the ID plugins, it's assumed to be safe. It's not.

And we demonstrated, uh, uh, a social engineering hack that's, uh, tempted, uh, developers. We read about 'em, we hear about it every other week, whether it's a docker container or from a repo component. It, Yeah.

So, so now we can apply this protection by, uh, pointing at, uh, Jeff Oga, your, uh, single source of record for, uh, for your ID plugins too. And another security related announcements that we made is around the gen remediation and, uh, what we've done there. So, uh, we do with modesty, we, we have one of the best, uh, research teams, uh, uh, in the world at Jeff o mm-hmm.

The security research team and our security advisories, our very accurate to a degree that you can, if you find a, um, a, a zero day in when you scan the code, the advisory that Jeff o gives you is, is one that if you take this advisor as a junior developer, it really tells you what the problem is. It gives you an example of how to fix it, and it goes into details of, uh, what exactly need to be changed in your code. And what we figured is that we can just give it to the LLM and we can prompt the LLM with the research data of jfo, and the LLM will remediate the, the vulnerability or, or the zero data that, uh, the JFO scans found.

We started with the integration with the, uh, co-pilot with the GitHub copilot, um, as part of the VS code integration. But we will extend it. And the, the user experience is you write your code, jfo is, uh, scanning your code continuously, it finds issues, and it's taking the research data of the JFO team to prompt the LLM and apply immediate, uh, uh, suggestions of how to fix that.

And you just have to accept it and, uh, and merge the changes. So, uh, that's the, the, uh, I think that's the last, uh, uh, big announce. No, I don't think we did.

Did we do fly? We, yeah. Yeah.

Started Fly, Fly, fly With Fly. Right. Okay.

I got a little confused. An ambitious and ambitious lineup. Yeah.

For one Swamp up. Yeah. Very ambitious.

And, uh, a team that works relentlessly on the, I mean, that breaking trust to, to our users. The theme around all of it though, yo Yoav, excuse me. Yeah.

Okay. Yoav, the theme around all of this is really the, the transcendence of ai. And, you know, how we're seeing this just totally upend the normal flow of, of, of progress, of, of it, of software development, of the software development, lifecycle insecurity in DevOps, in platform engineering, in, in everything.

It's, if you're not adopting this to as, as Shami said on the stage, if you're not adopting this, get outta the room. Get outta the room. Another important kind of theme here though was no one company can do this alone, right?

You j Fraud's, great company, you got a great research team, you got great developers, but the, we're talking about just upending entire Yeah. Ecosystems in, in a blink of an eye almost. And so you need a partners like a ServiceNow and an Nvidia and Sonar and some of the other ones that we've spoken about.

Definitely. How is it working? 'cause now you're not just working as one team, you've gotta work at the pace and in coordination with other engineering teams.

Yeah. How does that affect the pace of what you, you are doing at Jfr? So, first of all, like you said, we are in an ecosystem, but, um, I think we are in an ecosystem of, of platforms today.

Yes, there may be a few platforms in, in each domain, but still it's an ecosystem of lots of platforms that also makes the integration points. Once you figure out the integration points, uh, it, they, they are becoming very natural. So what we find out, first of all, we have great, great partners with us.

You mentioned ServiceNow and GitHub and Sono, but once you found out the logical integration points, it's very easy to get the teams together and, uh, create sort of a v team that works together and, uh, and creates the, the, the first level of the integration and then carries on to, uh, uh, to polish it. Uh, so it's actually surprisingly, maybe, but works exceptionally well once, uh, every once you have the clearance of, uh, how things are working together. Now, another thing that you mentioned is the, the impact of, uh, of ai.

So AI already made a huge change in how we code. Yeah. It's completely different now.

Nobody even is surprised by that. Maybe the next surprising thing, but this is also, uh, a reality today, is that you have coding agents living, uh, alongside the, the, the human developers. But I think that's the main gap is around.

So, so coding is kind of solved. It'll change a a lot I assume also, but, um, it's already, it's already happened. But I think where we still free, uh, see friction is around software delivery.

Because what's happening is that releases are being created in a much faster pace than ever. So it's a really a nonstop release train that is happening. And you cannot stop to, uh, think about irrelevant problems such as how do I version my release?

And what is the compatibility meaning compared to the, to the previous release? It's just an ongoing flow of, uh, of releases. With frameworks like UPT trusts, you will get the quality of the release so that you can trust.

It doesn't matter if, uh, it was an AI agent that created the release or, or, or a human or a combination of both. You have the gating, you, you have the governance to make sure that your release is, is ready for to be, to be deployed in, uh, in production, uh, and to be promoted, uh, across the different, um, um, policy gate. Uh, but at the end of the day, you need a new way to identify your releases.

Yeah. You need a new way to pinpoint them and, and, uh, and scale them up and roll them, roll back and identify issues with existing releases. And this is, uh, part of what, part of the change that we introduced with Fly, with the Gentech release as well.

I, I think between Fly and with, with the AI catalog, that's one of the sort of unwritten or underlying thing things, is that versioning is gonna change. Versioning, yeah. You will need a version because at the end of the day, you need to the down.

Yeah. But it doesn't need to be something that you, uh, take note of or remember. Uh, and it cannot be anything.

The trust is still not there to walk in a full semantic way with the releases, but it'll gather it'll Time. I'm sure it because trust, trust is a trailing indicator, never a leading indicator. You know what I mean?

You gotta earn it. Trust me. You Gotta earn it.

Yeah. But I think it'll also play out like that because of, uh, of agent to agent communication. Yeah.

So the negotiation of what kind of capabilities you have, it cannot be bound to a, to a specific version. It doesn't make sense anymore. It will be negotiated based on semantic, uh, between agents.

And speaking of that, we actually had, uh, uh, ya Janan on, uh, but the FPC server, he, he did a lot of great work on that. Yeah. Made sure to tell us.

So very proud of it. Yeah. Jonatan started the MCP server of Jeff Fog as a local MCP server.

As a, as a almost a, as a pet project. Yep. Uh, and then we, um, kind of, uh, upped the game and, and did a fully remote server.

Yes. Which is more, more difficult to do. But, uh, as a company, it allows, you have to have better control over security.

And also, um, you don't have to request clients to update the, uh, the MCP installation on the local machine. But it's a, it's a, uh, what's the word? A reference.

It's an indication, uh, a reflection. That's the word I'm looking for. It's a reflection of our times that before January, no one knew we didn't have MPC service.

Here we are, September MPC, here we are in September. And it is the standard. You must have it, you can't do without it.

Yeah. I think it's, uh, kind of a co common thing that we're seeing today that, uh, things are changing on a, on a, um, You Know, Today it's adequately new tomorrow. It's old hat.

Yeah. Well, MCP has a lot ahead of it. Like the, a lot of proposal of, uh, improving the standard and adding, um, soft, um, stronger authentication and, um, and the iden stronger identity and, and so on.

Well, I think there's also the A two A thing and There's the A two A thing, which are, we, we can argue whether the standards are Complement. The thing about A two A is now that's part of Linux, I believe. Foundation.

Yeah. That's some big names. And, uh, we'll see.

I mean, this is all gonna play out that the, the issue is for people like you and I are who've seen this, you know, we've seen these games. We've seen these plays before, never at this velocity. That that's the key thing.

The velocity here, the, the time crunch. Yeah. It's, uh, incredible.

It's the wall. Yeah. Yeah, yeah.

No doubt. What could we look? So next year in New York?

Yep. God willing. I'll be there.

It's my home. September 1st, We will be there. What do we, what You want to give us an early preview or too early?

I think it's too early, especially we just, uh, uh, wrapped up saying that, uh, things are changing so quickly actually. Yeah. So betting on, even betting on next year, uh, is hard.

I think you will see, uh, first of all, you, you will see there, there are some things that I can say that, uh, uh, you will definitely see like, uh, a lot of improvements on what we are bringing to market. Uh, today with Abrus, we have, uh, a few more things, uh, at our sleeve. And also, uh, we'd fly, uh, I think we will see a more, um, a more intention based way to do DevOps.

Yeah. Most, uh, um, vibe ops thing if you want. Yeah.

Vibe ops. Okay. Well, what ops dev, vibe ops.

'cause you gotta have the dev in the ops with something in the middle. No, but in, in, seriously, it's going to be much more intention. Oh, yeah.

Faced, uh, with, uh, a higher degree of trust. So I think that this Is this, you, I remember when HTML came out, all of a sudden I was a coder. I was never a coder, but H-T-M-L-I could do then.

Yeah. HTML 2 0 3, 0 4 oh CSS JS script, you know, all these things came on. All of a sudden I wasn't a coder no more.

I think we're gonna see a similar kind of thing. You'll have, everyone could be a, a developer with vibe coding. Everyone will with AI will develop something if they need, but there will be the tools that the pros use, right?

That vibe coating, refined vibe, coating squared, or whatever you want to call it, where it'll be for professional developers. And, and that's, you know, developers aren't going away. They're not gonna be replaced.

They're just gonna be empowered With this. I, I, I agree with you. I think we will have humans mainly for, uh, just expressing intention and providing, uh, feedback loops.

Uh, there's that. I, I'll tell you what else, and I've written about this. Uhhuh For, You'll Need Humans for the Creative Spark.

AI is very good at when you say, I wanna do this, I want you to do this for me, I want you to create that for me. But it doesn't create the ideas. Of course, The human brain still creates the ideas.

It's that spark of humanity that I think will always be The human is the guide. Yeah. The human is the guide.

Yeah. Uh, yeah. But I'm, but the reason I asked you about next year is because I didn't think you would know what's gonna be next year, otherwise why you should retire if you already know what's gonna be next year, retire.

But I would like to have you back on in July, maybe next year. We will talk about Swamp Up September 1st With pleasure. Alright.

Yoav, Yoav Laman, CTO Co-founder helping wrap up our day two coverage. But we're not done. We still have a few more.

So stay tuned. This Alan Shimmel for Text Drunk tv. We'll be right back.

Hey everyone, it's Alan Shimel. We're back here live at Platform Con Day in New York City. Of course, this is just the in-person day of a week long virtual event that's going on.

com. I forgot how many speakers and sessions there are, but there's a lot. And I encourage you to do so.

Let me introduce you to our next guest. His name is Sylvan Chu. Yes.

You got this right. A neighbor of mine from Fort Lauderdale. We're both up here in New York.

Um, Sivan, welcome to Tech Drunk tv. It's nice to have you on here. Thank You, Adam.

Um, not everyone watching this is gonna know who you are. Tell them a little bit about yourself. Yeah, So, um, I'm a former software engineer, was an SRE for nearly 10 years.

Um, then I was an entrepreneur, got an education training software engineer, and now I'm, uh, heading the rootly AI Labs. Um, so for this, we don't know, ROOTLY is an incident management and on-call platform. So our competition to PagerDuty, uh, which, you know, I think you probably heard of.

So we help businesses to manage their incident, and we're used by, um, you know, small companies and large businesses like Nvidia, Figma, Cisco, LinkedIn, and so on. Right. So we, we help, uh, a large, uh, large company, um, and the operation team to make sure that their incidents are, um, handled smoothly.

And my role, uh, truly is to lead this AI Labs. And the AI lab is a community led initiative where we work with team of fellows. So, um, we have people who are tech leader in the industry.

We have, as the head of platform engineering at Venmo, the former head of AI at Twilio and research students. And we work with these folks to really understand what can AI bring to the world of readability. And it's applied ai.

So we build prototypes, open source tools, we run, um, research and we write report, and we share all of this open source, um, on our GitHub with the community. So really the goal of this lab is like, how do you use AI for SES or platform people? Excellent.

So this then probably is a good audience for you. Yes, it is. It is.

Um, you mentioned you were on a, uh, a panel this morning. Yeah. So we were on the panel with, um, Google Sort work, um, and Nvidia.

And the goal was really to discuss, um, what's, uh, what's hyped with AI and what's reality applied to platform engineering. Right. Like, uh, I think we hear a lot from, uh, the executive and CEOs from uh, uh, model providers who are selling a GI or fully autonomous system.

I think we all agree that we're not there. And, um, I think especially for practitioner, which I think today is a lot of practitioner, they really want to understand what's true, what's maybe not there yet, and how can they really apply this in the, in their day to day job. Yeah.

Sivan, I think one of the big problems, especially for practitioners is every day it seems there's a new news story that some big tech company is doing a layoff and they're laying people off 'cause they're replacing them with ai. Yeah. I think as we sit here today, very few people are actually being replaced with ai.

Correct. I think what it really is, is that these companies overhired Yeah. During COVID and before.

Correct. And they need to cut back. They have too many people.

And rather than just saying that they kind of blame it on ai. Yeah. And so AI gets this thing of, oh, it's taking people's job.

Yeah, Yeah, yeah, yeah. You know, New York State is now setting up a tracker, jobs lost to ai, you know, and, and, and so you talk about separating reality from a hype. Yeah.

To me, that's a big, a big issue here. Now, I'm not saying that AI may replace people's jobs someday. I'm not saying that AI can help us or cannot help us.

Uh, to me today, AI is more of a co-pilot than a pilot. Mm-hmm. Does that make sense?

Yeah, it does. And, and so I wonder, now on the other hand, I'm talking to you, you run an AI lab. Yeah.

What do you see? What do you think? Yeah, so, you know, I think there are different type of AI labs.

Uh, if you look at, you know, the large stakes who are building these, uh, models, you know, these are like more like research, uh, researcher and PhD, and people who are like building all these LMS at ru we are really taking, um, a different stand where it's like really applied ai. So we use this tool to see how we can empower current practitioner, augment themselves, do their job better, and, uh, faster. And yeah, I agree with you.

It's like any tech new technology or tools, eventually it may replace some jobs, but maybe it's for the good. You know, like let's say before electricity, we had people going in the street and lighting the candle, uh, you know, for, for street lighting. Like, we don't use this anymore, but maybe that's a good thing.

Um, so for instance, um, uh, shortly we are really focusing on incident management, right? That's, uh, what we're about. And what we found is that you can really use the AI to help, uh, operation team to spend less time on managing incident because that's not something you want to do.

Right? Um, so I will share two main use cases where we saw, um, uh, you know, how this technology can help. One of them is incident, uh, triage and filtering, right?

You have all these alerts coming from a lot of tools, and you don't human to be looking at this, right? So here, LLMs can do a great job at like helping to filter and cut through the nose. And the second thing is, uh, root cause analysis.

So when you have an incident and you need to understand what's happening, what's wrong, a human may take 10 to 20 minutes to like, gather all the graph, look at GitHub to see what were the last commit, maybe go on Slack and see what conversations they were perhaps on the project. With LLM, you can reduce this by like 80 to 90%. So instead of spending 10 to 20 minutes investigating an incident, it can be done in like one to two minutes.

And that's a huge, that's, that's a factor of 10, right? It is. And, and I think that is, at least in the interim, that 10 x is the goal.

It is 10 ai, 10 x you. Yes. And we've been speaking about the TenX engineer for A long time.

A long time. It's finally coming. Absolutely.

Absolutely. So, and it's finally coming through. You know what we didn't mention Rootly.

What's the website? com. com.

Yeah. And, uh, the, the platform helps you. Basically, we sit at the center of your incident response, um, uh, efforts.

So you connect all your monitoring and logging, logging tools, uh, to our platform. So your data and Sentry. And, and we, we help help your R team to orchestrate a response to that.

So we will create for you, um, a team or a Slack channel, uh, spun up a Google meet or Zoom room so people can, uh, share. And then we embedded, um, a bunch of features that will help you to, uh, do the job faster. For instance, we have a bot that will listen to the conversation on Slack and audio, you know, and then if someone join an incident, you have a bot that you can ask, Hey, what's happening?

Can you gimme an update? Um, once an incident is solved, you have to write a postmortem or incident report. No one's likes to do this, so we automated this for you.

Um, so yeah, it's like a very, like, basically when something breaks, SREs go to Rotten. I love it. Ban.

It was a quick 15 minutes. It was quick. Indeed.

Thank you. Thank you, Alan. Thank you for telling us this.

Maybe we'll get together in person in Lauderdale, come into our studio. Yeah, I'm down. All righty.

Thank you. We're live at Platform Calm. We've got a lot more coming your way.

Stay tuned. We'll be back in a moment. Hello and welcome to the latest edition of the Techstrong AI Leadership Insight series.

I'm your host, Mike, er. Today we're with Brian Weiss, who's CTO for Hyper Science. And we're talking about, well, getting ready for Gen AI because it's a little bit more challenging than we imagined.

Brian, welcome to the show. Thanks, Mike. Really glad to be here.

We've seen everybody kinda launch one experiment after another, but I'm not quite sure that a lot of that is making it into production environments. And part of the issue seems to be is that it's not necessarily all about the technology, it's more about the rules, the regs and the cost and other factors that go into that. But what are you seeing?

Um, I see that a hundred percent and agree with not only sort of the, the stats, but the sort of trend that, you know, we start out with AI being kind of a, a, a solution looking for a problem. And while very, very promising for things like retrieval and summarization, the real rubber on the road now is, is is, um, data inside the enterprise, right? That actually, you know, tells you about the language of the business or a process.

Of course, as soon as you do that, you, you're into privacy. You're into understanding like where that data's being trained, how it's being used, how you get access to it. So we see, I see blockers in twofold to success of AI projects and adoption is one is is that sort of the, the, the concept of the unbridled use of AI to do everything and anything is, is, is, you know, it needs to be kind of reigned in a little bit.

Uh, and then sort of the realization that the, the hard stuff is actually in implementing to get you to get to the data and answer questions about the data you care about. So I think the last stat I saw was, you know, that, that over, you know, 60% of projects that have kicked off to do something with gen AI has stalled and they stalled for, you know, a lot of the reasons that, that you mention up front here. So we're seeing it in spades.

Um, you know, at hyper science we live inside the enterprise and we work extensively with really secure data. So things like veterans claims, things like, you know, information that, um, is very, very specific to individuals, uh, whether that's department of defense, uh, those kinds of things are, are, are mission and it's mission cri critical data where you can't be wrong, uh, when you're looking to get information out of a document set of that sort of thing. So there's the criticality of the information, the need for getting it right, that I, I think a lot of the early stage gen AI use cases are, are, are banging up against, right?

Hmm. One of the issues that I think we're now confronting is the sins of our data management past. And we all have structured data that, um, we manage reasonably well.

But most of these AI models are being, or need to be fed something that looks more like unstructured or semi-structured. And well, if it was unstructured, we tended not to manage it all that well. So are we revisiting all of that stuff now and kind of, you know, dealing with an issue we probably should have been dealing with for the last decade?

Uh, yes. Uh, part of, you know, a lot of what we're encountering right now feels a lot like the early days of enterprise search, to be honest, right? I mean, imprint enterprise search was, was not all about structured data.

It was about, 'cause I can do a SQL query on a row in a column. The question is what does this thing say, right? And how do I find the information in this 50 page document or a handwritten note?

So we have, I mean, TRA technology has traditionally struggled with all of that noisy information and it's kind of been a, you know, a North star that we've we're as you get more compute and now we have, you know, transformer models that can read things and do probability for what they understand and say and be able to respond. It's another chapter in that. But it is the unstructured data that it's, it's kind of the same problem, right?

That if you haven't put some guardrails and structure around that for who can see it, how you can use it, what you need to do with it, um, then bringing the technology sort of full, you know, full bore to that is, is can be a real problem. It's a struggle. Like I see a lot of the common struggles in gen AI use cases that, um, we're endemic to enterprise search, who gets to see the data.

If, if I, if I load all this, this stuff up into my enterprise ai and does someone get to say, Hey, who makes the most money at this company? Right? Um, but that the sort document level security, all of the problems that are associated with who can see what and what's available and how it's available are all now trip wires in some of these processes.

And while there doesn't seem to be a lot of regulations that's AI specific, uh, I hear folks will get down a path to a project and then suddenly they'll encounter something like HIPAA or whatever it is that they didn't think that they were gonna have to deal with. And suddenly they're like, oh wait, we can't do this 'cause it's gonna violate any one of 20 different regulations that are on the books. Um, how do we kind of navigate that so that we're not wasting time building things that are not gonna be used?

Uh, I have a really strong opinion about that. And that is you need to work with AI and modeling technologies that you control. So you control what goes into the model, what, how it gets used, and sort of the providence of that sovereign model.

So we, we work extensively in government industries, financial services, where that, that's predicated on that. And in fact, it has been the blocker to being able to use some of the broader capabilities of AI now where at hyper science, what, what, you know, that's kind of a non-negotiable. Like you have to be able to explain where you got the answer.

You have to be able to understand the ground truth data that is being used to, to fine tune or train the model, um, and be able to really own the outcome, uh, around secure data. So, so my, my, I I think there's a, there's a, you're right, there's a kind of a bifurcation happening here. There are models that don't do that, right?

Don't use them, right? Don't use them. You use a, use a platform which allows you to select and tune and train and, and, uh, models which are accountable to not only the data they use, but also to the answers they give.

Uh, and I I would say that they're, they're, you're sort of splitting two categories of models. Now there are those for which I can do that and those for which should look if I'm gonna use them, then I, I'm, I, I can't get that accountability or, or, uh, transparency. So we, we have been building models for many years, uh, for in, in-house, in some cases air gapped environments, right?

That look at, you know, say for example, uh, healthcare claims at the, at the Veterans Administration, like these are complex boxes of documents that have handwriting and all kinds of stuff all over them. There's no ter external modeling usable there, right? We need to be able to be on, on site at the va.

And, uh, I mean we're, we're, we've got models now that, that deliver, you know, AI results at 99% accuracy. And we've taken the, you know, the processing time from months down to days. Uh, but all of that's contained.

Like, like you can explain not only the answer, but also how it was trained and, and the way it's being used in combination with those techniques. So I, I see it, I see the market maturing and, but you're, you're absolutely right, Mike. There is a, there's lots of them that stall where people get excited about using a frontier model.

And then, um, look, the new, the new InfoSec gauntlet is, is your AI review committee. What model are you using and why? And what is it doing and who owns it?

And where's my data going? Like this is a, this is now the new normal, right? To have to really vet any kind of model inside an enterprise extensively.

The other issue, or at least one other issue that I keep hearing about too, is people will get through the pilot and then they'll go into production and they will have grossly underestimated the cost of running the thing. Yeah, Yeah, yeah, yeah. I look, that's another market maturity thing, right?

So if you think about it, the hyperscalers who are trying to, there's a sort of this big land grab to become the model that everybody loves and it's being underwritten, right? And as soon as you have to think about using that at scale, there is an underlying cost that's actually very, very hard to accommodate, right? So people get excited, like, I'm gonna use this giant model to do this task that used to, you know, but why would you use a helicopter to cross the street?

Like we we're like, I'll use that to cross the canyon, right? But if I, I'm, it's not, somebody's gotta pay for it at some certain point. So you absolutely see these things like, wow, it worked really great, and then you realize that you actually scoped it in a way which is just financially unreasonable.

Um, so I, I see that all the time. And, and you know what, what, what I'm focused on as, as sort of the, the composable platform at hyper science is using the right tool for the job. So, you know, let's use the CPU driven trainable models that are, understand your data and get you a really great result for the price.

And then I can then stack lots of things that are way more complicated, read, more expensive, GPU driven, all that kind of thing. But, but now I'm gonna start to decide like I want the best outcome for the right price using the models that are the most effective. But I see that all the time, super excited.

Let's use it. Let's use this giant model to do this thing. And you realize like, oh my God, I just took a helicopter across the street and I can't pay for it.

Like, why did I do that? You know, all the time. Do you also think that maybe, you know, to your earlier comment, will AI push more people to something that feels like a private data center, whether it's on premise or a private cloud or something?

And, um, we're all gonna be seeing a lot more of that activity rather than just relying on a public cloud. Uh, that's already happened and already happening, like this first wave of AI workloads. Most of our clients, most, 'cause we're dealing with government entities.

You're dealing with anybody who's, who's rightfully concerned about the providence of the data and the ai the workloads are going OnPrem, right? They're going, these workloads are going. So you, you see the industry responding to that problem.

But these init, this initial set of workloads on private done data that needs to be managed the carefully, um, is, is on-prem. Like, it's, it's, it is shifting, you know, that that sort of grand move that we all had to the cloud, like everything's gonna go from manage to just 100% public cloud and I'm gonna buy it by the minute and consume it. And, but now all of a sudden it's a, it's a very, very different, uh, um, process.

And, you know, we're in kind of a unique spot at hyper science. We, we deliver on-prem, we driven private cloud. We have a, a FedRAMP high secure SaaS environment.

Uh, we are, we are partnered with some, some, you know, the hyperscalers in, in, for example, Google's effort to provide a managed on-prem service of their, of their models right? Is also embedded with hyper science. But yeah, I, I think it's kind of the old is new again in that regard.

And, and it's, it's a hundred percent understandable. Uh, Is, is this therefore gonna become something of a rich company's game because you're gonna need to buy the infrastructure set up those data centers, get all that data managed. I mean, none of this stuff is inexpensive.

So, um, you know, what can a smaller company expect to be able to do versus a larger enterprise that has the resources to drive this thing? Yeah, I'll go back to my helicopters across the street. You don't actually need, you can achieve really high results, uh, and high performing results with narrow models on an ensemble which are actually cost effective for the task.

So you don't necessarily have to, you know, only the, only the, the really rich people can afford the the machine, which, which will get you the, the right and the perfect answer. Like that's actually, it's going the other way. What we're starting to see is that a composable architecture where, um, a combination of models that are cost effective and then you bring in the ones that are more expensive to, to do workloads that make sense.

You know, I've got a 300 page, uh, credit swap agreement with nesta tables and handwriting all over it, and there're gonna be chunks of that that are really relevant for a a, you know, a a large model. They're, at the end of the day, they're, they're, you know, they're, they're probability calculators for language and they're big ones, right? But I don't need the, I don't need it to tell me the square root of 32.

Like, I don't, I need it, I need a calculator from, from CVS to do that work. So don't ask that, right? Don't spend your money there.

But I, so I, I don't, and then, you know, the other point there, Mike, is that, um, look, the, the, the amount of innovation driving costs down, I mean, that's, it's, markets do that and then technology market does it, it's the same thing we saw with the virtualization of CPUs and, and you know, when cloud came out, like, you don't, so that, that is happening and you'll see the, the, you know, the compute get stronger and the price get driven down because there's so much pressure to make that possible. So over time, I, I don't, I don't think we end up being in a, in a, in a class warfare situation here with it, right? The market will respond.

And if people are smart about what you know about, about combining the right tools and not trying to get, you know, you know, buy helicopter to go everywhere, uh, I think gonna be all right. Will there be a shift in demand then, based on what you're saying? Because right now when I encounter people, you know, all they want is the latest and greatest GPU and they forget about the other GPU cycles, but before that, and those ones are a lot less expensive, and there's also other classes of processors.

So mm-hmm. Are we gonna get smarter about all this stuff? Uh, yes we are.

And I actually think that the, um, the market will create that for us. Um, so you're already starting to see companies that have sort of have bet the farm on, on being able to underwrite the GPU, but a fixed cost, but yet give the, if you can, if I'm, if I, if you get unlimited use of A GPU at a fixed cost, what is my business model if I can't charge you by the minute? And you can use it as much as you want.

There's a, there's a sort of a, a vicious cycle that I have to get ahead of, so that, that trend to sort of, everybody wants the latest and greatest, and there are some artificial pricing things happening right now with, you know, fixed cost against unlimited use and things like that, that are, and you know, we're seeing people go outta business, um, as a result of that. So I think what happens is that, that everything just kind of, that you can't keep that, that you can't stay ahead of that curve. It's almost a Ponzi scheme, right?

So what will happen is we'll end up with, um, we'll, we will end up with like, well, yeah, the good, the good enough model is, is good enough because I can afford it, right? As soon as the really good ones I can't afford anymore, and we kind of artificially are being able to think we can afford them. And then once all of that shifts, I think people will quickly say, yeah, yeah, I, I I need the right answer for the right price.

Not, not I need to use the greatest thing on the planet to get the same answer. Uh, Um, we also, you know, once again, seem to be thinking about security as an afterthought here, and a lot of the deployments that I've seen so far have, you know, significant vulnerabilities and there's all kinds of new ways to hack into these AI models. Are we waiting on some sort of catastrophic event before we get serious about AI securing?

Uh, I hope not, but maybe I hope not, but maybe, I mean, it, it's the, um, and I think, uh, uh, you can go back to other sort of inflection points in, in technology that've been somewhat similar, the internet, right? Things like that. And, and each time that happens, you create this surface area, uh, attack surface for, um, for bad actors.

And I, I do think we are, it is running very, very fast. Mm-hmm. You're not wrong.

It's running fast. And I do, I do agree. And I I, I am concerned there's risk in that.

Do I think that, um, you know, it'll, it'll take a major event to, uh, to snap everybody in the line. I really hope not. I, the trend I see in large enterprises is they are getting very serious with, uh, AI governance boards and security committees to try and keep up with it, uh, in the security industry.

I mean, it's a, it's a whole new world of, of vulnerability. So I see enterprises reacting and trying to be sure of that and get ahead of it. It's also, you know, there's so much new every day that is risk.

Mm-hmm. Also, early on, at least it seemed to me, most of these AI projects were led by so-called Tiger teams, and they pulled everybody together and they even had dedicated IT people who knew about infrastructure. But is more and more of this AI workload gonna just be shifted over and managed by traditional IT teams?

Or will we always need tiger teams? Yeah, you're seeing the impact of innovation, right? On business structures.

Uh, and it, i it will level out. I mean, you don't, you don't have to be an AI expert in order to be able to then look at hyper science. We, our platform is developed for just ordinary business users to be able to train models, um, and get high performing results.

So the data sciencey part of this is, is being democratized and productized, number one. Um, and then as we find those use cases that really matter to a company, I mean, there's a lot of stall light, you know, AI projects that are about, you know, building agents that'll do magical things someday. If you can figure out how to justify the ROI and then there are are folks and who were in this category where you're actually just creating hard ROI in the business.

And so as, as businesses find the use cases that, that deliver, you'll, you will, you will see it codify into a sort of run rate it function. I don't think the Tiger team who specializes in all things AI is a, um, is a new and permanent sector of the enterprise. Hmm.

So having considered all these things, what's your best advice to folks? What should they be thinking about right now to kind of avoid what are se what are some serious pitfalls? Um, look, the first one I say is, is know your ROI target when you go in, um, there are a lot of phishing expo just like, you know, science experiments and things like that, that end up being an, and we all know the downfall of that is that, you know, you're, you can't justify what you've done or how you've done it.

So there are, there are drill sites for real value with AI driven, uh, um, you know, opportunities. And, uh, I would start there. That's the first thing is let's just sort of understand the outcome and justify the business outcome and know how you're gonna get there.

So I, I'd say that's the first one. Um, I would say, uh, if anyone tells you that one, one model, or you know, one vendor's got the magic thing that's gonna do everything, then you prob just like think twice about that, right? Um, the, you know, using models, plural, using AI in a com, in a, you know, composed way is gonna get you better results and also more control.

Uh, so I'd I'd say look at a composable approach. Um, and then the last one, I think you called it already, which is data security and, and the providence of what that model is working with, what you're using it for. Uh, and, and ensuring, particularly if like, we're, we're, we're in the, we're in the really, really high fidelity mission critical data business, right?

So, um, it's one thing to ask a, you know, a a model to go do research for me and summarize things and all that kind of stuff. But if I'm asking it to, you know, process information, uh, and make decisions potentially that are mission critical on data, you can't really be wrong. Um, so you gotta think about the right, the right tool for the job.

What is that, what is that model supposed to be doing? And, and I would say above all, you know, managing security and privacy, et cetera, is that model accountable for when it's wrong, right? So when it's wrong, do I know, can I solve for it?

Will the model in the process or whether the platform itself help me solve for that? Um, now look, maybe you don't care. Maybe maybe error rates are great, right?

It doesn't matter. So you get the sentence a little bit different. And what if you're just sort of generating content and reading books and doing summarization?

It's something. But when you're into mission critical data process, you really sort of look at the whole picture of, of what, what do you do with wrong? Or what does that pro that model, that process tell you about wrong?

And does it give you the tools to bring people, say, for example, to solve it and sit next to the model? I think that because the frontier of this is not models do everything, or AI does everything, it's a combination of, of governed models working with the right slice of the, of human intervention to ensure you have not only data quality, but security and governance and, and transparency of the outcome. So I, I gave you a lot there, Mike does that, I'm probably gonna have to go to ask GPT to summarize all that for me into the main points.

But You know, I, I think you heard it here, folks. I think based on the goals, the risks and the cost, you gotta make sure the AI price is right. Hey, Brian?

Yeah. Being at the show, I guess that makes me a human, GPT too. There you go.

Thanks Mike, everybody. Thanks for watching the episode. You can find this episode and others on our website.

We invite you to check them all out. Until then, we'll see you next time. Data Centers on Fire.

No, literally Data Centers on Fire. You're watching Textron Gang. Hey everyone, happy Wednesday.

Welcome to our Textron Gang show for a hump day. We've got a lot to talk about. It's been a busy week.

I feel like we put a week in already. It's been a busy week, but we've got some interesting stuff to talk about with some interesting people. Let me introduce you to them.

We've got our friend Chris Blak, the one and only Kate Scarcella, Dan o' Dan O'Brien, and joining us from Barcelona, Mike Ard. Hey, Mike. Okay.

You able to get the Yankees out there? I'm gonna find out in a couple hours, but yeah. Wonders of the internet being that was yesterday.

Yeah, This is true. All right, let's jump into it. So, you know, we spent yesterday talking about the, uh, exciting news on a MD and, and, uh, open AI the day before that.

More exciting news, it seems every day there's exciting news around AI and data centers, right? I was talking to Dan o yesterday or the day before. You know, something like half of the growth in the U-S-G-D-P this year can be attributed to AI and data centers.

JP Morgan put out a report that for the first time in a very, very long time, consumer spending isn't driving the GDP growth. It's actually AI and data centers are bigger than consumer spending. Consumer spending is usually 70% of GDP, you know, are we in danger of, of kind of getting in over our skis here?

Can, can AI momentum outpace our ability to, to, to build out the infrastructure? Mike, what, what's the deal? All right.

Well, we have two things to talk about here. One is there's a report from our friends over at volter who confirms that well, demand for AI is far ahead of the available AI infrastructure, and we're not able to keep pace. And I don't think that's gonna surprise many people.

However, at the same time, Fujitsu and Nvidia have been talking about a new design for AI applications and AI agents in general, and a lot more integration at the silicon level. So you would see GPUs more tightly integrated with CPUs, in their case, their arm processors, but also there'd be ASIC and networking and infrastructure all built into the system at a much lower level. And Dan, I'd love to get your impressions on this.

'cause I wonder if a lot of the stuff that we're using today might seem fairly antiquated soon as we look at these new platforms that people are building. And are we finally building platforms that are made for agentic ai? Yeah, well, I think we're getting closer, Mike.

Um, you know, we've thrown, you know, really general purpose. GPUs a lot of work on the software side as well as the hardware side over the past several years on really building efficiency and right, the move to rack scale systems, you know, getting to the point where we can get much higher utilization rates out of existing GPUs. Uh, that was a lot of what was behind the news yesterday with, uh, open AI and a MD was, that was really the first AMD's first rack scale system, which integrates this all much more tightly.

Um, and I think, you know, there's, there's certainly a mindset out there. I mean, you look at the power, yet, the power efficient power efficiency of the human brain, we're a really long ways away from that. Um, and so, you know, I do think we're gonna continue to see more and more efficiency gains.

You know, there's a lot of people out there that believe, and I probably count myself as one of 'em, that we need a fundamentally different architecture to kind of scale a high to the point where we're eventually gonna get to. Um, and I think, you know, a lot of the announcements here of like immediate Fujitsu, you know, they're all incrementally moving the ball forward on that map. Um, I mean, just, you know, recently as we moved from, you know, LLMs and the more the agent world, you know, you're seeing 10 X tokens and under x growth in compute.

So, you know, the, the type of AI we're doing is we're getting more into reasoning. Um, and some of these, you know, and more advanced styles of gen ai, uh, the compute demands are just off the charts, right? You know, we're, we're years away from, you know, catching up on the supply side as we build out this infrastructure.

And, you know, there's a couple gating factors there that we're working with too, right? We're hearing more and more than it's really about the power supply, uh, but also, you know, there's not enough chip making capacity to build all the chips we're gonna need for this stuff. And I think, you know, TSMC being somewhat of a monopoly in the market, you know, is pretty good at kind of policing that overbuild on that side of things.

Feels like we're really gated by the power at this point. Uh, but an enormous amount of compute we'll still need to drive over the next several years to get supply, to catch up with demand. But lots of incremental innovations that will happen over time.

And I think, you know, ultimately a fundamental architecture ship, I think we're, you know, right now it kind of the system level design thinking, but, um, still using a lot of the, the same architectures we have that that will certainly have to evolve over the coming years if we ever wanna get power efficiency down to the point where we can actually afford all of the supply that we, you know, see to need on the compute side of things from a power perspective, You, you know, Dan, what you've just described is exactly the issue. Why we always want software versus hardware. Because software as we improve it, you just update your software, right?

It's a download. It's a di the old days we put a disk in, and you got, you got the next version. It's more efficient, it's better, it's faster, it's everything else.

With hardware, you can't just update your hardware, especially when, you know, you're, you're literally putting hundreds of billions of dollars, trillions of dollars into building out these hardware factories. And even though it's incremental, the drag time between how fast the human brain can design AI assisted, of course can design new hardware, you know, uh, con configurations versus the time it takes us to actually build those configurations. It always seems like we're going to be driving last year's model next year, right?

We're building like the world's fastest appreciating asset. Yeah. But we're gonna need to shrink, you know, the IRS is gonna have to come in here and let us, you know, write it off after 24 months 'cause it's obsolete.

Um, Just feel like there's some sort of financial engineering bomb around the depreciation schedules of this stuff that will eventually hit the industry. But Yep. Seems like knock came down the road for now, You know?

And what about, you know, how tech people are? We always, we want the fastest shiniest, you know, trinkets. Now, we don't wanna wait.

What, what's, what's a tech bro to do? Chris, I saw you had your hand up. I, I'm just thinking, you know, I, I love this.

You know, for me, we do this show once a week, right? And we've been doing it, it long enough. And with all the terabytes of video, much less the exabytes of stuff we're talking about as I'm listening to this, I'm thinking of these tiny little attestations between us.

Like literally Dan, you remember a, a power conversation a month or so ago talking about the grid, you know, and we're all, you know, adamantly agreeing, we need more power and all these little spots, you know, week to week through this year stitched together. And the the answer is to these questions are, are in there, right? You know, so whether it's, you know, from the power issue, yes.

The way we're scaling power, you know, is queued up for, for this segment. The grid actually doesn't match the expectations. But we're going to do those things.

And you know, we, we've had a couple conversations about the architecture and workflows and workloads. You know, we're, we're putting everything in the massive ai, you know, to do math or, or to, you know, change the capitalization in, in a paragraph. And on this, as I've mentioned before, you know this tiny little laptop back here.

There is LLM running, this is for Clifford Brewery, shout out to Darlene and Clifford, who keeps the, manage it there, keeps it all running, and of the benefits of all this stuff. So we've been working up and down the, the supply chain down to the micro business, the underserved who has a million different applications. And all the stuff we talk about here hits them coming and going.

They can't see it coming. But these systems we have now allow us to serve that need. And when we talk about this ai, we're talking about taking little bits of human text and turning into data and vice versa.

So people like Darlene and operations like Clifford can do what they they do. And if it works at the foundation like that, then maybe these esoteric things we get to talking head about here actually matter, because the, the physical grid is not moving faster than, like you said, Alan designing things physically implement. These are huge, long pro uh, projects.

A lot of the tra trajectories that we're predicting, living, experiencing right now don't match those. What are we gonna do? Maybe some of the things we've talked about on the show are, are right, we'll spread it out.

We'll get more intelligent and we'll do all these things. 'cause Oh yeah, we're, we're, we're embedding AI in everything. The benefits are too huge here.

So something that we've talked about though on this show has been the idea that we are building for tomorrow, um, on yesterday's thought, on yesterday's architecture, right? And Dan, you've talked about that a lot, about the architecture. Um, AI is sprinting on a runway.

We haven't finished paving. And I think with that being stated, I I do think the architecture is, is changing. I think it will actually become more distributed, uh, for a lot of different reasons from a sovereign perspective, right?

Um, having, you know, to have agency over, you know, geopolitical, our, our, our data, right? And I think the whole architecture of a cloud model, it, it, it's not gonna work for, for the chips that we're building. And, and it doesn't even really make sense.

0, um, distributed model, I think. And yeah, I mean, and it will make us, um, more resilient and everything. Again, I think we're building on, on, it's almost like we really have to start thinking differently.

And we're not there yet. I I think we're, we'll be there to do that. So, So Kate, follow up that thought for a second.

If you would, do I, as a developer, let's say I'm building some sort of agent AI app and I'm starting it now, do I assume a level of compute capability that might be available a year and a half from now? Or do I kind of not trust that and I just go with what I know I personally would, would go with the chips that we're not sure what are what? 'cause we're not utilizing them that, that we know for sure we're not utilizing those chips to their fullest.

And I would go for it. I think the one who goes for it is the one who's gonna win. Oh, I was just gonna say to me, the, the long pole in the tent here feels like power.

You know, the, the total capacity and the location of it. And to me, that's ultimately gonna drive the constraint based thinking across the industry. You know, there will be somewhat a power ceiling we need design for, and, you know, I think we can count on innovations down at the hardware level and then spilling down to the software level that allow us to essentially accomplish, you know, kind of what we need within that envelope.

Uh, that to me is, you know, kind of the right way to think about this. Um, I think the chip market will, you know, basically build, you know, to the point where, you know, you can deploy it and you've got the power to run it. Um, that's kind of the constraint on the upside of the chip side.

Um, but, you know, there's, there's a lot of levels at which, you know, we can optimize here. You know, we're, we're very much still in the build and move as fast as we can phase, eventually we're gonna hit a little bit more of a plateau. And that's when you tend to see those real efficiency gains kind of come through, right?

Because the data centers we're building today, they're where the power is and they're where the power is gonna be, but that power is gonna sit there and that power will be relatively consistent for a very long period of time. You know, ultimately, like tens of generation of chips will probably come through those data centers and, you know, think about the amount of compute you're gonna be able to get outta that data center 10 years from now with that same power envelope versus the, you know, amount of compute you're able to get it out to out of it today. Um, that, that's gonna be a, a tremendous increase.

And I think, you know, right now, early in the cycle, a little bit harder to see the light at the end of that tunnel. But I think once we get a little bit more mature and that efficiency starts to really compound, that's kind of what we'll get out of this, you know, spiral that we're at right now And, and all, all of this because, because now we finally can, you know, I like looking for patterns that, that are, they've always been there. They're sort of inevitable to fill in.

And then whether it's cybersecurity, the internet writ large, I mean, Kate, you've touched on one of the key things, and when he said that he, the, the resonance, like literally Darlene, yeah, I told you this one, right? Um, data sovereignty, because to date, I mean, if you're, most companies, much less small companies have 'em help you, the millions, tens of millions of micro businesses, your data is out there with somebody else. But it doesn't have to be because Moore's Law, blah, blah, blah, you know, the, the amount of data, you know, you actually care about as an organization, as a person, whatnot, tended not to be that much.

And no, it doesn't have to be in the cloud, and it doesn't have to be somewhere else. You could actually have it all right here because storage is cheap and compute is cheap. So you follow these curves up and you find us doing things because we finally can't.

Not because it's radical, it's because the way we've been doing things aren't really the way we wanted to. We had no choice. You know, 1999, 2000, 2001, I I'm out in Colorado near Boulder.

I'll, I'll give you some names. Interlochen, anybody ever been to Interlock in Colorado? Right outside of Boulder, right?

Uh, I forgot the hotel there where they have some good, the Omni the Omni Hotel. They had some good conferences there. com company.

And when you got to Interlochen, it's in the foothills. And as far as the eye can see, you saw two and three story buildings with companies with names like level three. And, and, uh, was it Quest, be Pack Bell West became Quest or something?

They, they had a few names, eventually Century Telecom or something. And there were, there were all of these fiber carriers, and then there were, there was the storage companies, mixed storage, storage, tech storage, this store that, as far as the eye can see, because we were building out data centers all over the place. This is when Northern Virginia became the data center capital of the world.

We had a beautiful one over in Tysons, you know, and it was the same thing. We've gotta build storage centers. And back then it wasn't, power wasn't the limit, it was bandwidth.

We couldn't, we couldn't, you know, you could build the biggest g*****n machines you could with big racks, but if you only had a straw to send the data through, it wasn't really doing you much good. And, and we had these same conversations, same conversations we're having now, and then the bubble burst. So it Took 10 years.

I would go for so far to say though, like if I look at the last decade, hardware was kind of boring commodity. And then GPUs came along The last decade, the last 20 years, 25 years, when was the last time hardware wasn't boring, But, but I'm surmising or at least postulating that hardware might be cool again. And as I look forward to see what people are starting working on and all these engineering things that people are talking about, for example, what Fujitsu was saying, that stuff's all gonna be crucial.

So I don't know Dan Hardware cool again, or what Hardware's very much cool guy. I think it became cool again, probably during the COVID Pandemic when everybody realized how critical chips were to pretty much everything you make, you had a hundred thousand dollars cars sitting on the lot, the lot 'cause of a Tencent by car controller. Um, and I think it's, you know, it's become more mainstream, you know, ever since.

But I think you're right with, there's a ton of innovation happening, right? You know, whether it's, um, you know, some of the, you know, other instruction sets like risk five, um, you know, new IP out there on the compute side, you've got in-memory compute, you've got analog AI techniques which use extremely low power relative to the type of stuff we're seeing today. There's a lot of innovation that's in been incubating during this, you know, unsexy period for hardware.

And I think it's gonna have its stay sometime in the next, you know, next decade or so. Alright. I was, I was, I was just going, I was just going for cool there.

I didn't go all the way to sexy, but okay. I went back to a hardware is cool, cool. Now, because you know, it, you were talking about chips and, and you know, to be clear, almost all chips are sitting there doing nothing almost all the time because our systems, we just, you know, we've overloaded the hardware because we had no choice.

But now we're smart enough and we can make complex enough systems that go back and use the old hardware to do things, you know, you could never imagine. We're doing it now. It's fun.

Good stuff. Hey, we we're a little over on this segment. I'm gonna need to pull the plug on this one.

Let's take a quick break here on the gang, and we're going to come back and as I promised you at the outset, data centers on fire. Wow. You're watching Text Gang, 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. You 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 cloak, digital executive protection, defending the new attack surface your personal life. Hey folks, we're back in. Yes, this is Alan alluded to, we're having an issue with fires and data centers.

It turns out that there are lithium batteries that tend to blow up and systems get overheated. And well, maybe the most important thing that any AI company can do is make sure there's a firehouse nearby. But Chris, does any of this kind of surprise you?

And are we gonna see any more of this stuff before maybe it gets better? I don't know. You know, the metaphor kind of writes itself.

You know, AI just doesn't burn cash and carbon anymore. It'll burn your actual wrecks. Um, it, it, it comes, you know, as we said in the last segment, you know, it, it comes back to these issues we keep talking about.

You know, the demand is real. You know, the demand is real. The market and cash will, will land and market and cash lands, heck will build things.

And is the, are the actual, is the infrastructure altogether. We're sort of, well, well, al al you know, my, you know, my boating experience. Like let's build boats while you're sailing 'em and, and so on and so forth.

See what happens. That's what we're doing. So yeah, sometimes things are gonna burn down, right?

You know, when your ambition is melting your racks, right? That's just, you know, hubris with a warranty, but, uh, hubris with a warranty. Yeah.

Nothing particularly surprising here. You know, look at it as indicators, right? You know, as a, a individual, you know, fire event is always a bad, bad thing.

But looking nationally, globally, you know, we're building 8 billion of these. Some of them pop because heck, there's no infrastructure there. And we didn't think this through very well.

Maybe there's steering lessons. How many of you have actually managed a data center? Well, I worked for a company managing data center, so I I have too, one of some of my funnest times was checking out the fire systems, right?

The fire suppression systems. What's the stuff they use? It's not, you know, it takes the oxygen out of the rooms.

It is a chemical hey lot, right? I thought it was, it's, Hey, look, you know, I thought we've licked this problem already. Apparently not.

Because I think part of the issue is the lithium batteries that are being put into the backup systems are at the core of the issue. I think, Dan, Yeah, no, I think you're right, Mike. It's the batteries.

We've introduced these, you know, really, really difficult and SRE devices in as part of the backup storage here. And the big takeaway for me is, you know, fire safety comes in. It's an architectural decision.

When you build a data center, you don't, you don't plan for a fire by having, you know, brave souls show up on a truck. Like this is true. We built into the design, right?

You know, compartmentalizing, you know, keeping servers further away from batteries, having fireproof barriers between them. I mean, it's really built into the industrial controls of the data center itself. Um, but, you know, beyond just the, you know, the prevention of fires for, you know, all of the, the risk it poses the, you know, health and human safety and all that, but the economic impact that these things going down is tremendous.

I mean, it reminds me of similarly when a, a semiconductor found goes, you know, as the firing goes offline, I mean, it's hundreds of millions, billions, maybe tens of billions of dollars in, in, you know, lost opportunity for these things to come off, right? I mean, you know, what Jenssen's quantified, uh, you know, a gigawatt, uh, data center is, you know, roughly equivalent to look at, you know, $50 billion in value. Um, in terms of, you know, the amount of compute capacity you can sell out of it.

Um, these things coming offline is not only gonna be really economically damaging, but, you know, a lot of mission critical, you know, critical services for, you know, our economy and our society are gonna be running in these things. And if you add a massive data center come offline, that's gonna ripple through and it, you know, and impact everyday people, um, at a certain level. So I, I think it's definitely a topic that we gotta take seriously.

And, you know, the right time to think about this stuff is when reporting concrete and doing the building, uh, because it truly does have to come down to an architectural decision to make these things as safe as possible. And, and you bring up a good point, Dan, and that is, um, you know, it is actually the potential of costing lives becomes a part of this equation as well. And something that we, we do need to take seriously.

I mean, is it time to update the fire coats? You know, this is, again, this is not a new problem that we've dealt with at in societal levels. It's the same way after Hurricane Andrew.

They changed the building codes down here in South Florida, or in Florida in general. Maybe we need to re-look at the, uh, fire regs. And, and Dan, to your point, you know, what, what do you need to get a c of o certificate of occupancy to get, you know, an approval here?

What, what fire suppression systems do you have? What architectural decisions do you have? You know, this, this seems, hey, maybe AI can help us with this, right?

Reasonable to revisit it. I'm certainly no fire code expert, but, you know, my concern is it does feel like, you know, actually at a a federal level and even at a state level, there's a lot of deregulation and kind of skirting around some of the, you know, processes and procedures we have around this stuff. Um, you know, really, 'cause I think, you know, the knock on America is we can't build fast enough, you know, whether it's fast, it's data centers, Well, it'll take what, what That's way that we're actually deregulating amiki easier to do this stuff versus, you know, versus potentially harder by, you know, really kind of clamping down on, you know, rules and policies.

What was the fire where the women burned in the Garin district, uh, in Chicago to, I i the Linen? It was, think it was the linen factory in New York, right? No, but the, there's a name, I don't remember the name, but it, it took one, The triangle shirt factory.

The triangle shirt factory, correct, Mike. It takes one of those and then people will, you know what Dan, they'll say, okay, maybe we shouldn't, we shouldn't cut those corners. You know, maybe we should take that stuff seriously.

Unfortunately, it may, it may take a tragedy like that. So a lot of the guys I grew up with are firemen, and they all frequently remind me, they say, you know, Mike, when your house is on fire, we're not coming to save your house. We're coming to save your neighbor's house.

'cause your house is already toast. The data center's pretty much the same thing. I mean, they're just gonna come there and try to contain it, but I don't think they're there to save it.

I, I think if you're relying on the firemen, it's too late. I think you, you've gotta, you've gotta design for fire, if you will. So last, last week, last Wednesday, I was on the, uh, at, at sector, the, the, uh, black Hat Canada event.

Uh, shout out Ja, Jamie Arland was moderating the fail panel, and I got to be the old person on it, right? You know, with two younger folks. And Jamie had brought props for us.

I got to, to hold the 19 89, 9 and a half inch floppy that had the industrial control system, Dan from the Orfield water, uh, water system, uh, back in 1989. And the folks next to me had a CD ro and a five and a quarter, three and a half. And we talked about, uh, the, are are we getting better?

Classic, I think it been 19 years they've been running this fail panel, right? So it literally long enough itself to, to be an example of, of what we're talking about. And, and I, I guess the point that I was trying to make there is the same one I couldn't generally make here, you know?

And because we're moving along a path, we're at this point, this is not new. You know, everything from, you know, I think everybody here, you know, said parts of this in, in this segment. Um, we've been down this path before, you know, we're having a lot of things catch on fire.

What do we know about fires back in 89 where that DiSette was around, you know, I worked a, in, in, uh, South Carolina at a, a data general var, and we actually had the raised floor in the HA lawn system. And yeah, we kind of forgot about that and, you know, but, but it's not that we need to reinvent the wheel, like it's so many other things. We need to now, we can both do what we've already done before, put that back in place and connect things, uh, appropriately and fast enough to, to solve the, the current issue.

And, and maybe as we cycle forward, you know, raise our qualities and have different levels of failure in the future, which is the whole goal. Yeah. To fail more fun, right?

But just not the same way over and over again. I always wanna be introduced to small modular nuclear reactors to power these. Like, we've also talked a lot about that.

Uh, but it's something Why, what could go wrong there? Things Always go Wrong. But I, but, but here's, here's the, the, I I don't wanna use the word irony, but it's ironic.

It's not a question of could we be designed them safer? Of course we could design them safer. I think the question is what Dan said, which is, do we have the appetite to design them safer if it means delaying putting them online?

Well, and that's the whole, we don't weigh everything together, right? You know, and fail failure is, is a wonderful thing, right? You know, like if you're not failing enough, then maybe it's really stable, that's fine.

Um, but it's certainly not a growth area. And this is, you know, in data integrity of, of IAM there's so many different areas. I think we've been driving towards this.

How do we get to perfect and we need to understand, stop trying that don't, there's no such thing, you know, there's, to your que question, Alan, it's a balance. If the benefits are X and the cost is walkable, data centers burn down. Maybe let the data centers burn, but know you've made the choice and put it together.

Sounds like a song burn. Baby burn. Yeah, that's what I was thinking.

I was like, okay, somebody who paid for that data center's gone. Yeah. You know, that data centers tens of billions of dollars.

Well, anyway, hey, you know what? This is gonna be, I, I think an issue we're gonna see, unfortunately, uh, you know, until something bad happens and, and you know, it's what, what my mom used to say, funny till someone loses an eye or something, right? Mm-hmm.

And, and that, that's what you got. With that, let's take a break though on this one. We're gonna come back and let, let's turn over to managed insecurity.

Discover Textron 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 back. And we're talking about, well, the number of instances involving security breaches that get traced back to some sort of third party provider of an IT service or security itself seems to keep climbing and folks are getting a little bit worried about that.

But there's a converse to this argument. Some folks would say that the providers of those services know more about security than the folks who run the internal ITE systems. So you're better off relying more on those folks.

But Alan, I know you put a post together over on digital CXO talking about this in terms of the risk and how we should be thinking about this. But where do you fall in this debate? 'cause it kind of sounds like you're saying don't rely on the third party.

So look, Kate, Chris, I'm sure you guys are gonna have a lot of thoughts on this one as well. And remember, this wasn't on Security Boulevard, this digital CXO, 'cause it's, it's a leadership discussion. Of course, you're gonna rely on a third party security's too damn hard, right?

For, for, except for a handful of companies. Maybe the Fortune 50, maybe not even all of the Fortune 50. Security is too hard to do by yourself.

Almost by definition, you're gonna have third party providers helping you. We live in a cloud world where we have, by definition, seated some of our, uh, capability for security to the cloud providers 'cause they control the stack. But the real point of my article is whether you are outsourcing, Taylor, I'm hearing a echo.

The real point of my article is whether you are outsourcing your security or giving it, sharing it with a cloud provider or some third party MSSP or, or an AC provider who you give access to your network, as in the case, I think it was a target, right? That led to the target breach. You can outsource the work.

You can't outsource the responsibility when the stuff hits the fan and your customers are impacted. They don't want to hear, it was one of our partners, they don't want to hear it was a third party. It's not our fault.

The the buck stops on your desk. And so you gotta remember that whenever you bring in a third party who is in any way remotely responsible for your security posture, for your compliance, I'm not saying don't do it, but you've gotta be thoughtful and, and kind of follow best practices and make you making sure those third parties have, are doing what you think they should be doing and what you feel is reasonable and managing that risk. Because when the, when the music stops and you don't have a chair, you are the one in trouble.

You can outsource operations, but not accountability. And if you think about, um, you know, hey, we lived in Florida for 20 years. I'm sure you know Chris and, and Alan, you remember, um, the, the deep, um, the water oil, the deep water oil issue, and it was, you know, bp, but, you know, who were they blaming?

They were blaming Halliburton and, uh, Transocean and it, but at the end of the day, it came back to bp. So we see that over and over again, right? That, you know, you gotta, it's gonna go to the brand at the end of the day.

And we see that. Absolutely. I, I'm gonna, I am going to, uh, bring the, uh, late grade admiral gr amazing Grace Hopper into this, you know, because she was real admiral.

And and for those of you who don't know the name, read a book, but, uh, she, she, uh, in the Smithsonian, there's a log book with a Beatle taped into it, you know, that was caught in a, in a relay in a computer, and she said, found bug. So if you talk about bugs, you know, amazing grace, and she's famous reverend, all sorts of great, uh, sayings. And what I'm always loved as a sailor myself is sail.

You know, a a shipping port is safe, but that's not what ships are for. Sail out to sea and do new things. And to be clear, she's not just, that's, that's an analogy.

She's an admiral. And as I was building the, the, the twains, my two silly solar power boats and sailing 'em up, uh, from the keys to St. Augustine, remember explaining this to some of the, uh, uh, uh, the Coast Guard folks in the water.

You know, to be clear, my lifeboat is another boat, and both lifeboats have lifeboats, right? So, you know, you know, you, you guys may have to come get me someday, but to be clear, I'll just be standing here. I'm not gonna be swept out to sea.

And the nice thing about that sort of thing is what, what, what, uh, Grace Hopper was meaning is that take responsibility and, you know, there's nothing else after that. There's not taking responsibility, just taking responsibility. And then whatever happens you're responsible for.

So, you know, whether you're literally sailing out to sea or whether you're, you know, building an oil rate deep horizon, right? Uh, Kate and or running a data center or anything else, it's your responsibility. So you can outsource your cybersecurity.

You know, Alan, your intro is exactly right. It's all very complicated stuff. You shouldn't be doing it unless you are a security company.

But if that is your one link, if you decide to have one fragile link, you know, again, have a lifeboat, have a second boat, have a second lifeboat, it depends. But you've decided to have one brittle link, and that's your, your choice. So Chris, I I hear redundancy in five nines.

Yeah. Redundancy to your redundancy. Well, when it's life critical, right?

You know, I literally Backup. I love it. Yeah.

Right? You know, at worst case, on those boats, I'll be, I'll be in there in the ocean with an anchor, you know, floating, but the storm passes that I'm still there, right? So you, you gotta think this all the way through.

It's your fault, period. I wanna think a little different than an approach here, right? I, I think what we're ultimately saying is you can't outsource the business impact, right?

Ultimately, something goes wrong, it's your business that will impact everything in between is just the blame game, right? And, you know, whether you have security totally inside, how many CEOs have survived an incident by throwing their CSO under the bus, right? That's a short shelf fly job if you ever seen, right?

So, you know, you're running the business as a CEO. Ultimately something goes wrong. It is your business gonna be impacted?

The rest is just the blame gate. Do you wanna, do you wanna have a third party outside? You can blame, alright.

You know, BP probably got off, uh, little bit easier because Halliburton took some of that blame forth, right? Kate, you know, to go back to your analogy, um, but, you know, end of the day it's your business at the end of the day that's gonna be impacted. And, you know, whether you, you you decide to do it yourself or you outsource it, you know, that's really much more a matter of cost, risk mitigation and skillset.

The term outsourcing is getting more nuanced though, because what a lot of the folks are doing these days is they're calling them co-managed services, and it's a, a better blend of an external service provider and an internal team. So everybody's kind of got skin in the game, but it's not like I'm dumping all of it onto one provider and hoping for the best. I think that's gonna be the better approach, and frankly, it's taken a long time to get there.

Much to my amazement. No, that's a good point, Mike. And, you know, co-managed, um, models with shared telemetry and clear SLAs, um, are the future.

So it it has taken though way too long. Yeah, go ahead. This Part.

Good resiliency strategy. Yeah. You know what, when, when CrowdStrike had the incident, which is just about a year ago now, I believe, right?

Um, the blue screen of death and Delta went down. I didn't see all those delta passengers in the airport saying, God, Don CrowdStrike. No, they didn't.

They blamed Delta. Delta Tried as much as they could To do everything they did. They sued them.

I, I don't know what happened with that lawsuit, but they did Sue CrowdStrike, but people were p****d at Delta. That's that, it's your business. And when you outsource you, Kate, you said it, I think in the beginning, you could outsource what not resp, uh, capability, but not Outsource operations, but not accountability.

Not accountability. That's, that's the bottom line here. And then if you're a digital leader out there, I'm not saying do all your security in-house, you probably be more danger to yourself doing that.

But you've gotta recognize that you still are accountable to your customer. It's still your business. And, and the ironic, wonderful thing about it is once you do that, it turns out you get more business.

Right? You know, and, and it, this, this, I'm sure I, I'm old, I repeat my stories all the time, so forgive me, but, right. You know, in a vc, early VC life, for me, giving a pitch, um, after the, the, the Cisco firewall thing, um, one of the in investors present, leaning forward and saying that, that, you know, billion dollar, uh, uh, firewall thing, how'd you do that?

And I just honestly said, well, we built, built a team that died for each other. We took responsibility with the market and built trust. And this, this individual leaned forward and said, no, you can tell me.

And it key. So the key to what we did, literally, I think, uh, Danny said it, uh, somebody said a minute ago, we had this horrible thing happened that was totally our fault. So instead of saying, oh my God, you know, we immediately went out and said, we're gonna go, I'm Chris Blak.

I'm responsible. It's our fault. It's as bad as it looks.

It's in fact worse, and we're going to do everything we can to fix it. And here's our plan, and we did it. And that actually works.

You people trust you. They give you more money, they give you more trust, as opposed to saying, you know, the whole blame game thing. We all look at the, the, the looks on our face when we bring that up.

It's like every human being goes, oh my God, it's blame game. But if you could actually develop trust, that's business. That's money, that's efficiency.

But you have to actually mean it. Oddly enough, you know, your, your grandmother told you these things, right? Remember back to when you were a kid?

It's true. You know, I think George Kurtz, so I know George a very long time, probably 25 years. Um, George K is the CEO founder of CrowdStrike.

George did, that's exactly what George did. He went before Congress, he went before the cameras, he went before everyone, black hat. He got on the stage of black hat and said, yeah, we, we screwed up.

We messed up. And here's what we're doing to make sure that don't happen again. And, and I think that is better than playing the blame game.

You're, you're right, Chris. That's how you build trust. That's how you build a business.

To, to be fair, Microsoft had a hand in this, but they blamed CrowdStrike too, so, well, I'm not gonna throw dirt on them, but yes. But Microsoft's been blaming everyone for this, for their security issues for a long, long time, unfortunately. But, but they had their moment too, right?

There was the time where Bill Gates stood up and said, security's gonna be important. We're doing trustworthy computing. They hired the best security people in the world.

And there was that period of time from maybe 2005 to 2010 or 12, where Microsoft was stellar, stellar security. And now, And, and now the president of Microsoft shows up in front of Congress once a year and doesn't a coppa and that's that 10 Hail Mary's make and like, make a donation at the box on the way out. You know, that's, that's security today.

But anyway, it is, it's important for leaders out there. It's something as security people I know we, we deal with all the time because we know, we, we don't, you know, look, if you are using a cloud, you're, you're, you're, you're outsourcing security to a certain extent. Absolutely.

There's no way around it. So anyway, gang, what a ter, terrific, terrific discussion today. Of course, we had a double dose of ai, a little security thrown in.

Doesn't get better than that. Uh, as usual, we have a full Textron TV show following our Textron gang here today. If you're watching it during our stream, if not, whether you're watching this on OT OTT or Textron tv, or text on tv, YouTube channel, thank you for watching.

We have a lot more content out there. Go check some of it out. Mike, good luck to you in, uh, Barcelona.

And then of course you'll be at Amsterdam. You'll be, I, I'm sure, recording videos and filing stories like mad, uh, Chris, Kate, great seeing you. We'll see you on the gang soon.

Dan, though, I'll be talking to you I guess in a little bit, hopefully, but keep doing what you're doing. Enjoy your day, everyone. Stay tuned for Text Drunk tv.

This is Alan Shemel, we're out. Hey everyone, welcome back here to Techstrong tv. I'm really happy to have our next guest on.

Let me introduce you to Christian Rig Rodriguez. Christian is a CTO over at CrowdStrike. When he is not surfing, it looks like.

Hey, Christian, how are you man? I'm doing, I'm doing great, Alan. You know, thanks for having me on.

So I always like to find out a little bit about and share with our audience a little bit about, talk to us about the boards. Oh, so, and I just got into surfing, uh, last year. My fiance's a big surfer.

She's kind of surfed all over the globe. And, um, I basically am at this point now where I'm learning to not drown, uh, and also look fancy during that process. And so I've had the, the, the pleasure of going, uh, through like Central America and a few parts in on the East coast, do some surfing, getting used to, you know, standing up and, um, you know, not embarrassing myself as I get better at it.

But, uh, yeah, that's, that's, Yeah, no style Style absolutely counts in surfing, so It sure does. It absolutely does. Yeah.

No, good for you, man. Enjoyed It's a great hobby. Yeah.

Um, so Christian, when you're not surfing, and before you were CTO at CrowdStrike, talk to us a little bit about kinda your, you know, give us the Christian Rodriguez story. Oh, wow. I, we may need another hour.

Um, and so, um, you know, not we got time not including all the, the, the, the inner child work and meditation, um mm-hmm. But, but, uh, Haven't we all, Don't we all Exactly. Uh, I've been with CrowdStrike for almost 11 years.

I started here as a, um, sales engineer, probably part of the first 120 employees when we, we basically had like three offerings of the company from media r to intelligence and, and hunting. And you know, now we're, we're, we're, there's 30 plus modules, but, um, but my background even prior to coming to CrowdStrike is big in the solution architect slash systems engineer, sales engineer roles. I've worked for the likes of, of Websense and Fishnet and Zimperium.

Uh, so roughly 20, 21 or 22 years in the industry, you know, working for sure for cyber companies in some capacity. Yeah. Gary's an old friend of mine as well.

Oh yeah, I know Gary. Yeah, he's good people. Yeah.

Yeah, absolutely. One time we'd get a chance off camera, my first time to Kansas City to fishnet way early this early, early two thousands. And the first time I met Gary Fish and Jody Perel was the CTO, then is now CEO at Fire Mark.

But yeah, all, all good people. Yeah, great people. Um, So you, you've got some real, you know, cyber, cyber creds as they say.

How long have you been CTO over at CrowdStrike? So, Uh, feel CTO of the Americas for roughly three years. Um, where, um, I feel very fortunate this, this role is, I feel, um, I just love doing what I'm doing, right?

I get to, to understand the way we think of problem solving, but then I also get to, uh, work with all of these enterprises on understanding their vision and how they plan on maturing different disciplines in their security practices and, you know, how they think of problem solving as well. And then figuring out how we align our strategy with their strategy and their vision with our vision. And, um, it's just a really great opportunity to just speak with a very wide array of different customer types.

Uh, and then also just evangelize what we're doing here at CrowdStrike and what we're seeing in terms of new threats and, you know, what, what, what thought leadership is based upon customer feedback and how we're building new things. So Very cool. Yeah, very cool indeed.

Um, you mentioned when you first came on to CrowdStrike, you know, they had three modules. Yeah. Yeah.

And they've got 30. Yeah. Yeah.

There's a little something for everyone. It seems that when it comes to security and CrowdStrike, but you know, if we, if we, we don't have to go through all 30 modules 'cause we really will be here an hour, but, you know, Christian, gimme like big garbage cans, if you will. Yeah.

You know, break it into big cans. Yeah. The, the areas.

Yeah, I think, I think they're, they're really, um, I, man, it's, there's some there I would say, uh, four to five kind of major pillars, if you will. Um, you know, naturally we, we, the flagship capability of the, of the company or the nucleus, if you will, of the company's success has been around the endpoint, right? So we, we do endpoint security, uh, extremely well, right?

And that's, you know, the EDR, the next gen av, anything that we can see on the endpoint has been kind of the big, you know, thing for Crosswork for since our inception. And then we got into identity as well, doing identity and authentication analysis and behavioral analysis on who's logging into what, why are they doing that, you know, what do we do to prevent something that could be bad? And then there's cloud kind of the third pillar, if you will, uh, where we can understand, um, ephemeral workloads or the configurations around your systems and the likes of any of the major CSPs, like AWS or, or Azure, uh, or GCP or OCI.

Um, and then there's also even the, the, the SaaS side of the house, right? Where we, we we plug into doing SaaS analysis and authorization of analysis across SAS applications. Uh, and then even if you were to think about, you know, those kind of four pillars, then there's this AI model that sits on top of all that very high fidelity to telemetry that we're capturing where we can start to help our customers augment their stock efforts by having AI help them with decision making or hunting, or, you know, just doing analysis and triage.

Excellent. I think that was a nice way of kinda compartmentalizing Absolutely. Into some people wrap their heads around, man.

Yeah, absolutely. Yeah. All right.

com if anybody wants to go to the website and explore all 30 modules and Yeah. And jump into all that. There's Lot, there's lots of brows.

The Website. Yeah, there is. It'll keep you busy while, yeah.

Hey, look, it's job, it's job security for someone who does the website, right? Totally. Um, well, let's AI replaces them soon.

Who the heck knows? That's whole nother, that's another podcast episode. Get into that if you want.

Yeah, I was gonna say, Alan, that'd, that's a podcast episode where we, we, we, you know, we'll, we'll commiserate Exactly. I mean, like, keeps me busy just talking and writing about that particular subject. But anyway, you know, well, I'm not gonna get into it now, but I'm doing a thing right after this.

You know, s stands, uh, what is it? No, the s in vibe coding stands for security. Oh, that's funny.

And there is no s But anyway, um, let, let's talk, let's talk cloud security, though. You're a big advocate, right? For how do we put this unwieldy monster to heal?

Yeah. Let, let's hear your thoughts on this. Christian.

Um, I, every, every enter enterprise. Every organization we've met with has something in the cloud, right? I don't think, you know, cloud, cloud solutions are, aren't novel per se.

Um, people are moving applications into the cloud or moving infrastructure into the cloud, or they're, um, subscribing to cloud services to make their lives easier in terms of spinning something up. Uh, and what we're seeing is that, uh, the more that enterprises move to cloud, right, the better adversaries acclimate themselves with, you know, what's in these cloud services. We, we, you know, Alan, I'm sure you're familiar with the way that we track the bad guys, the adversaries that are out there responsible for the attacks.

We're seeing, you know, ECRM groups, nation state, hacktivist groups, for example. Um, and what we're seeing is that these adversaries are becoming ever very cloud conscious. So they understand how to navigate the control planes, and they understand the services that power, the fact that someone spun something up or workload or a container.

And so what we're seeing is that enterprises are having, you know, challenges defending against those attacks that are very cloud focused because they're also managing disparate tools. Or they may have a tool that's very much focused on endpoint and a tool that's focused on, again, the authentication mechanism that is GI giving access to that cloud environment. And then you're managing all of these different da disparate tools that lack context, or there's a lack of cohesion, but adversaries are really using the cloud as a pivot point even to move back on premise.

And so, you know, what we're seeing is that, uh, adversaries are very proficient in understanding those cloud services, and they're taking advantage of things like misconfigurations or they're taking advantage of systems that have been orphaned in these cloud services. And as a result, they're very successful with respect to how often they can actually successfully infiltrate an organization or enterprise because of those services that are misconfigured or, you know, IM policies that have been abandoned or have been excessive. And that's, that's essentially a major area for us investment wise, even just to give our customers better visibility and better protection and consistent protection, I should say.

Right? That analyst experience is extremely important in terms of consistency of dealing with a threat, whether it's on premise or if it's virtual, or if it's a hardware based box, or if it's an informal system, should be fairly consistent. So, Christian, I, you know, I've been in security 30 something years.

30 years, right? When I, I founded a few companies and helped founded a few companies. George Kurtz was at a company called Foundstone.

Yeah. When I first got into security, and I knew George, we had a vulnerability management tool as well. This has been a holy grail for a long time.

Yeah, Yeah, Yeah. Right. com too, right?

com, our new site cover platform engineering. You, you, you look at all these other areas and the move to a platform is kind of the natural evolution, right? Instead of doing point solutions mm-hmm.

For developers, we, we develop an internal developer platform and ITP and we have platform engineers that lay this all out. And we have cloud native engineers that, you know, even there, like how we, because I mean, Kubernetes was the hardest thing. One of the hardest things I've ever seen to use.

How the hell it caught on? I don't know, but, but it did, right? And everyone uses it, it seems so, but we have cloud native platforms.

We've been trying to do unified security platforms mm-hmm. For as long as I've been doing security. Why now?

Yeah. I think you're right. It's, it's not, uh, the concept isn't novel, right?

To say the least. Um, I think everyone has been trying to, uh, bring uniformity and consistency with, uh, a lot of disparate data sources. And I think what we're seeing is there have been, uh, it's less of a trend, and I think it's more of a necessity now.

I think it's, you know, if you're asking why now, I think it's because it's very expensive to touch all of these disparate systems. Uh, and it's also very risky, right? There's so many things that can happen in between those seams of these tools that people are, and organizations trying to stitch together, right?

There's so many things that can fall through those little cracks. And those, you know, an adversary only needs to be right one time, right. In order to be successful.

And so I think it, it comes down to the concept of the appetite for risk has been reduced significantly because the cost of a breach is also skyrocketing. And so, you know, companies can't afford to have these disparate systems that are, you know, poorly sits together, you know, act as their, their bedrock for security, right? I think platform and platformization in grand scheme of, you know, what platform truly is platform is, you know, native capabilities that can show first party data seamlessly connected to together, right?

So that that opportunity for the adversary to live in those cracks of those disparate tools now is going away, right? And we're beca we're making it harder for the adversary to become successful because everything is so well put together in a, in a true platform. So I think it's less about a trend, and it's more of around how do I ensure the, the future of my enterprise and the safety of my enterprise?

And it's ensuring that, um, in the consolidation story, if you will, is, is, is more tied to how do I do my job as a, as a defender more effectively and more efficiently, uh, versus dealing with these disjointed tools. So, Christian, I, I think all that true. Mm-hmm.

But let me, let me put something else forth to you. We need AI to make this work. 'cause the, this coordination, this tightening up of, you know, it's like, you know, the SR 71 Blackbird plane you've seen pictures of, that's Aly badass looking plane, isn't it?

Beautiful? You know, that used to leak fuel when it was first taking off or on the ground, it needed to go a certain height and speed and heat would build up that would fuse the, the metal there to make it airtight fuel tight so the fuel wouldn't leak out. AI is, it's the same kind of thing.

We need AI to get that tightness right? So that we don't get the leaks out the little. 'cause it only takes, as you said, only takes one time.

Yeah. Only one time. And it's only one crack.

Absolutely. Yeah. And that's all they need.

But with ai, I feel like, you know, maybe we finally have the, the putty Yeah. To, to, to seal all these things and bring it together and make it work in real time. 'cause that's another piece of it, right?

Oh, absolutely. We need this to work in real time. Absolutely.

I, ai I think is so crucial to our success as defenders, right? I think what we've seen is, and, you know, data, um, if you were to ask me even like way, what, what is CrowdStrike, you know, today versus what we were, when I started here, we were very much talking about endpoint security then. But we were very much focused on how do we do more with this telemetry right?

On the endpoint? And then how do we expand and open up that aperture so we see more things. And now again, right.

30 modules later, right? We're seeing everything on, again, cloud identities, endpoint SaaS, I could get on even third party ingestion. Um, and so in order to get your arms around around the problem that's growing, right?

Data is the problem. Your, your business. So, so that's, that was CrowdStrike, right?

Like we, we opened up the aperture. We're collecting all this telemetry. As a business, as an enterprise, you're also growing and your problems are becoming bigger.

The more you start expanding on data, the more that you start to monetize even data, right? And as your business. And so what happens is AI helps you get your arms around the problem.

'cause you can use AI to reduce remedial tasks, right? You can augment your Slack analysts and your defenders. You can start to assess large, uh, you know, areas of data to start making more sense of it, right?

And so you can start using that data to, to start getting answers faster. And I think that's really where on our side, AI plays a big role on augmenting your efforts as a SOC analyst, as someone that's looking into vulnerability management, someone, um, analyzing identities and authentication requests. And then guiding you into what is gonna be the next best step for our organization based upon trends that we've seen that mimic this type of behavior or trade craft.

Or what does a true outlier look like in my business based upon poorly written applications versus an adversary that has his, his or her hands on the system and they're moving laterally or dumping credentials or, you know, being persistent. And so I think AI allows you to start expanding your defensive measures in a similar fashion or a rate that is also matching what the adversaries are doing. And also keeping up with the way their business is growing.

So you could probably spend another hour talking about AI use cases, but AI for us is expanding on allowing defenders to, to extend themselves. And, and there's a lot that, that, that we're, we're gonna announce soon around the AI capabilities in our platform. But I, I think it's pretty, you know, it's pretty fun stuff.

It's pretty geeky stuff. It is, it is. Let me ask you a hard question about the AI, though.

'cause you know, when you first started talking about, we, when we first started talking, you mentioned, you know, and isn't it great to have sort of a personal advisor, a analyst to talk to and explain this to you? Is AI going to be that analyst soon? Or do you think it's a person who's augmented by ai?

I'd say for right now, it's a person that's augmented by ai. Uh, and the reason is, you know, your business is growing and processes change and adversary tradecraft evolves. And, um, the, the AI model is as good as the data that it has a access to.

And, you know, we, for example, at Wick, we've trained our models based upon what human analysts are doing to triage systems and remediate systems. And so, you know, I think every, you know, you may have a business that has a, a SOC analyst that needs to come up with an answer faster, right? And we don't necessarily plan on replacing that analyst, but if I can save that analyst five to 10 minutes for every detection right?

Or event that they're analyzing, that adds up if we're talking about thousands of events, right? Right. And so our goal is to have ai, uh, quite frankly, reduce things like alert fatigue, right?

Or this platform in as a whole reduces alert fatigue because it's showing you contextually the anatomy of an attack from start to finish, from the identity side to the endpoint side, to the cloud side, to the SaaS side, for example. And then AI is basically guiding you on, these are the things you may wanna hunt for, or this is how this event was triaged, or let's start automating some of those tasks for you. And having that human analyst validate that the AI is doing its job properly.

So I think for right now it's augmenting, you know, there may be a future state where AI's doing a lot of the work on behalf of that person, but that person would still be involved in some, some capacity. All right, we'll, we'll see. Yeah, we'll see.

Um, we'll see. Hey, Christian, I, you know, we're coming up in the black hat season. I know CrowdStrike out there, totally random, but for folks watching this who maybe are going a black hat I'll, what can they expect from CrowdStrike?

Yeah. Uh, I'll be there at, at Black Hat with, with the team. Um, I think we're doubling down on showing some really great innovation coming out of, uh, our, our Charlotte AI models.

Uh, so really excited to showcase some of those capabilities like our detection, triage, and our, uh, uh, our, our, our hunting, uh, capabilities and our workflow capabilities. Um, you know, there's a lot that we're doing in cloud security as well, right? This ability to do, you know, runtime analysis and protection on, on of systems that are, again, or workloads that are spun up in these cloud services.

I think that's another big component. And I think we're, we're really doubling down on, uh, you know, what our customers are asking us for. And that's just to protect, you know, everything, right?

In terms of, you know, runtime and the, the, the configuration settings that you have across those cloud services or, you know, those AI workloads. A lot of our customers are asking us, like, how do we get better, better visibility into, um, AI workloads within our ci cd pipeline? Who's spinning up what instance?

And is that going to lend itself to some type of new risk or some type of data loss? Uh, and so we're doing a lot on the AI SPM front for security posture management, and we're, we're excited to showcase some of that at the, at Black Hat as well. Cool.

If you go into Black Hat, go check out the CrowdStrike booth. Good stuff going on. Hey, Christian, I want to thank you for coming here on Tech Trunk TV and, and talking a bit.

Yeah. And you got it. Come back.

Visit us soon. If we see out in Vegas, we'll, we'll catch up there, but if not, yeah. Back here on tv.

Okay. You Gotta looking forward to it. Yeah.

All right, man. Ride. Good waves.

Thanks So much. Thanks Alan. Alright.

Christian Rodriguez, CTO CrowdStrike here on Techstrong tv. We're gonna take a break and we'll be back. Hey everyone, it's Alan Shilo.

We're back here on Textron TV in beautiful Napa Valley at the Jfr Swamp Up event, continuing our Day two coverage. There's a little bit of a break going on, you can't see, but out there people are eating ice cream and peanuts and potato chips. It's a little mid afternoon break, but we're still here 'cause we've got a lot more to bring you.

Let me introduce you to our next guest. If you've been watching Text Drunk TV over the years and any of our coverage of Jfr, you already know him, but I'm gonna pronounce his name right for the first time. 'cause it seems my pronunciation is a little old fashioned.

So let me introduce you to Yoav Laman. Perfect. Hey, nice To meet you, Yoav.

It's good to see you. Yoav, of course, is a co-founder, one of the co-founders and CTO here at Jfr. Yoav a pleasure.

How are you? That was Busy. Lots of announcements.

This one, probably The most we're a Few up that we have, right? Uh, lots have good, good feedback from customers and, uh, also some suggestions. So, uh, And that feedback's, that's the, That's the goal, Actually.

You know, what they say, the feedback from this year's Swamp Up will be in the products for the next, Hopefully even Well before with ai. We have to. So, yo, we were talking, you know, before we got on, we have had a lot of people give us a piecemeal, a piece here, a piece there, a piece there.

I'm gonna ask you, pull it all together for us, right? Give us the, a overview of all of these great announcements, all this great innovation mm-hmm. That was announced here at Swamp Up.

Okay. So I'll try to give the full umbrella of announcements that we made. So we started with Jeff O Fly and Fly is, uh, uh, our own disruption of the platform for, uh, a new agent, uh, repository based on Artifactory.

And that's, uh, that comes with, uh, a new user experience for managing software releases. Uh, so that was our first announcement. Then we went over to, uh, UPT Trusts.

And UPT Trusts is, uh, the way to control your software supply chain based on three, uh, major concepts. First one is application that gives you ownership assignment for every release, every artifacts, uh, in the JO platform. The second one is, uh, signed evidence.

And we announced, uh, partnership, uh, with many leading industry vendors, uh, such as GitHub, such as sonar, such as ServiceNow, uh, to, uh, uh, integrate their evidence, uh, into, uh, into the JO platform to accompany the, the releases. And, uh, finally, uh, it's, uh, policies that, uh, allow you to use the information, uh, within the JO platform, use evidence in order to assign rules for the progression of the artifacts, uh, of Theis, uh, all the way towards, uh, production. Uh, so this is Apru.

It's, uh, it's a unified package that includes all these, uh, three main features, uh, ownership evidence and uh, uh, policies. Um, so that was, uh, another announcement. We also had a deep dive to our integration around evidence with, um, with GitHub to take the salsa ance of GitHub, uh, workflow bills and put them alongside artifacts in the JO platform, uh, as evidence, which is, when you come to think about it, it's the logical thing because, uh, you, it makes sure that, um, that the evidence itself is bound to the artifact and you can never get out of think.

And it's also the fact that Artifactory is the entity that is exposed through your production. So that's, uh, one thing that where we did a deep dive of Frust. And the, uh, other thing is we announced on stage and integration with ServiceNow, uh, around Atrust as a, as a full, as a, as a whole.

Uh, and we show the, uh, synergy between applications that many of our customers are already managing in ServiceNow, and how change requests in ServiceNow are going to be, uh, reflected as evidence in, in, uh, in JO uh, and vice versa, how you can move between the platforms. So that was, uh, also a part of the big announcement of APTAs. Yes.

Then, so it's a mouthful. Then, uh, we move to, uh, uh, a new announcement, which is, uh, around machine learning and ai. This is AI catalog.

Yes. And AI catalogs, uh, allows you to have governance over, uh, models that are packages, but also models that are, uh, uh, SaaS like, uh, an and open AI and so on. Uh, under a single platform, you have a catalog where you can find the latest versions of the models and the metadata about them, like, uh, the security status, the, the licensing, and, and, um, other metrics that have to do with the model health.

And then, uh, similar to what we have, uh, uh, in curation, uh, we elevated the same features for machine learning. So you can allow different teams to use different type of models. Um, for instance, you may allow a research team to use deepy, but you never want to see that, uh, in a production facing, uh, uh, release.

And part of that, when it comes to, uh, to SaaS models, is, uh, also being a gateway between you and the SaaS models. So if you, for instance, if you're using open ai, uh, you will use it through the GO platform, and that allows you to have governance also over these type of models. Uh, so, so this is, uh, this is the gist about, uh, AI catalog.

Mm-hmm. And then we went into a bunch of security related, uh, announcement. I think I, I can mention two, uh, highlights there.

The first one is the support for, uh, ID extensions. Yes. Uh, and, um, basically it's a combination of artifactory acting as a proxy for your, uh, vs code extensions.

So we start with VS code, we will extend it to other IDs and, uh, curation allowing you, uh, to, uh, to, to control the, the, the, the, the extensions that your developers are able to install on their, uh, endpoints. And this is one of the most dangerous and overlooked, uh, risk that developers are, uh, currently facing because you basically install a software on your own from the internet that everyone knows that it's wrong, but, uh, for some reason with the ID plugins, it's assumed to be safe. It's not.

And we demonstrated, uh, uh, a social engineering hack that's, uh, tempted, uh, developer. We read about them, we hear about it every other week, whether it's a docker container or from a repo component. Yeah.

So, so now you can apply this protection by, uh, pointing at, uh, Jeff oga, your, uh, single source of record for, uh, for your ID plugins too. And another security related announcements that we made is around the gen mediation and, uh, what we've done there. So, uh, with, with modest, yeah.

We, we have one of the best, uh, research teams, uh, uh, in the world at Jeff o mm-hmm. The security research team and our security advisories are very accurate to a degree that you can, if you find a, um, uh, a zero day in when you scan the code, the advisory that Jeff o gives you is, is one that if you take this advisor as a junior developer, it really tells you what the problem is. It gives you an example of how to fix it, and it goes into details of, uh, what exactly need to be changed in your code.

And what we figured is that we can just give it to the LLM and we can prompt the LLM with the research data of jfo, and the LLM will remediate the, the vulnerability or, or the zero that, that, uh, the jfo scanners found. We started with the integration with the, uh, co-pilot with the GitHub co-pilot, uh, part of the VS. Code integration.

But we will extend it. And the, the user experience is you write your code, jfo is, uh, scanning your code continuously, it finds issues, and it's taking the research data of the jfo team to prompt the LLM and apply immediate, uh, uh, suggestions of how to fix that. And you just have to accept it and, uh, and merge the changes.

So, uh, that's the, the, I think that's the last, uh, uh, big announcement. No, I don't think we did. Did we do fly?

We, yeah. Yeah. Started Fly.

Fly was fly, right? Okay. I got a little confused.

An ambitious and ambitious lineup. Yeah. For one Swamp up.

Yeah. Very ambitious. And, uh, a team that works relentlessly on the, I mean, that bring trust to, to our users.

The theme around all of it though, yo, Yoav, excuse me. You okay? Yo have the theme around all of this is really the, the transcendence of ai.

And you know, how we're seeing this just totally upend the normal flow of, of, of progress, of, of it, of software development, of the software development, lifecycle insecurity in DevOps, in platform engineering, in, in everything. It's, if you're not adopting this to as, as Shami said on the stage, if you're not adopting this, get out of the room. Get outta the room.

Another important kind of theme here though, was no one company could do this alone, right? J F's a great company, you got a great research team, you got great developers, but the, we're talking about just upending entire Yeah. Ecosystems in, in a blink of an eye almost.

And so you need a partners like a ServiceNow and an Nvidia and Sonar and some of the other ones that we've spoken about. Definitely. How is it working?

'cause now you're not just working as one team, you've gotta work at the pace and in coordination with other engineering teams. Yeah. How does that affect the pace of what you, you are doing at Jfr?

So, first of all, like you said, we are in an ecosystem, but, um, I think we are in an ecosystem of, of platforms today. Yes, there may be a few platforms in, in each domain, but still it's an ecosystem of flaps of platforms that also makes the integration points. Once you figure out the integration points, uh, it, they're, they're becoming very natural.

So what we find out, first of all, we have great, great partners with us. You mentioned ServiceNow and GitHub and Sono, but once you found out the logical integration points, it's very easy to get the teams together and, uh, create sort of a v team that works together and, uh, and creates the, the, the first level of the integration and then carries on to, uh, uh, to polish it. Uh, so it's actually surprisingly, maybe, but works exceptionally well once, uh, every once you have the clearance of, uh, how things are working together.

Now, another thing that you mentioned is the, the impact of, uh, of ai. So AI already made a huge change in how we code. Yeah.

It's completely different now. Nobody even is surprised by that. Maybe the next surprising thing, but this is also, uh, a reality today, is that you have coding agents living, uh, alongside the, the, the human developers.

But I think that the main gap is around. So, so coding is kind of solved. It'll change a a lot, I assume also, but, um, it's already, it's already happened.

But I think where we still free see friction is around software delivery, because what's happening is that releases are being created in a much faster pace than ever. So it's a really a nonstop release train that is happening. And you cannot stop to, uh, think about irrelevant problems such as how do I version my release?

And what is the compatibility meaning compared to the, to the previous release? It's just an ongoing flow of, uh, of releases. With frameworks like UPT trusts, you will gate the quality of the release so that you can trust.

It doesn't matter if, uh, it was an AI agent that created the, the release or, or, or a human, or a combination of both. You have the gating, you, you have the governance to make sure that your release is, is ready for, to, to be deployed in, uh, in production, uh, and to be promoted, uh, across the different, um, um, policy gates. Uh, but at the end of the day, you need a new way to identify your releases.

Yeah. You need a new way to pinpoint them and, and, uh, and scale them up and roll them, roll back and identify issues with existing releases. And this is, uh, part of what, part of the change that we introduced with Fly with Gen releases.

Well, I, I Think between Fly and with, with AI catalog, that's one of the sort of unwritten or underlying thing things, is that versioning is going to change. Versioning. Yeah.

You will need a version because at the end of the day, you need to for the same down. Yeah. But it doesn't need to be something that you, uh, take note of or remember.

Uh, and it cannot be, and I think the trust is still not there to walk in a full semantic way with the releases, but it'll get there. It'll Take time. Sure.

It, because trust, trust is a trailing indicator, never a leading indicator. You know what I mean? You gotta earn it.

Trust Me. You gotta earn it. Yeah.

But I think it'll also play out like that because of, uh, of agent to agent communication. Yeah. So the negotiation of what kind of capabilities you have, it cannot be bound to a, to a specific version.

It doesn't make sense anymore. No. It will be negotiated based on semantic, uh, agents between agents.

And speaking of that, we actually had, uh, uh, Janni Janin on, uh, but the PC server, he, he did a lot of great work on that. Yeah. Made sure to tell us.

So very proud of him. Yeah. Jonatan started the MCP server of Jeff Fog as a local MCP server, as a, as a almost a, as a pet project.

Yep. Uh, and then we, um, kind of, uh, upped the game and, and did a fully remote server. Yes.

Which is more, more difficult to do. But, uh, as a company, it allows you, uh, to have better control over security. And also, um, you don't have to request clients to update the, uh, the MCP installation on the local machine.

But it's a, it's a, uh, what's the word? A reference. It's a indication, uh, a reflection.

That's the word I'm looking for. It's a reflection of our times that before January, no one knew we didn't have MPC service. Here we are, September NPC, like here we are in September.

And it is the standard. You must have it, you can't do without it. Yeah.

I think it's, um, kind of a co common thing that we're seeing today. That, uh, thing is our changing on a, on a, You Know, so Today it's radically new Tomorrow it's old hat. Yeah.

Well, MCP has a lot ahead of it. Like, there a lot of proposal of, uh, improving the standout and adding, um, so, um, stronger authentication and, um, and the ident stronger identity and, and so on. Well, I think there's also the A two A thing, and There's the A two A thing, which are we, we can argue whether these standards are Picking complementary.

Well, but that the thing about A two A is now that's part of Linux, I believe. Foundation. Yeah.

That's some big names. And, uh, we'll see. I mean, this is all gonna play out that the, the issue is for people like you and I who've seen this, you know, we've seen these games.

We've seen these plays before, never at this velocity. That that's the key thing. The velocity here, the, the time crunch.

Yeah. It's, uh, incredible. The warp.

Yeah. Yeah, yeah. No doubt.

What could we look? So next year in New York? Yeah.

God willing, I'll be there. It's my home. September 1st, We will be there.

What do we, what? You want to give us an early preview or too early? I think it's too early, especially we just, uh, uh, wrapped up saying that, uh, things are changing so quickly actually.

Yeah. So betting on, even betting on next deal, uh, is hard. I think you will see, uh, first of all, you, you will see there, there are some things that I can say that, uh, uh, you will definitely see like, uh, a lot of improvements on what we are bringing to market, uh, to today with Abrus, we have, uh, a few more things, uh, at our sleeve.

And also, uh, we'd fly, uh, I think we will see a more, um, a more intention based way to do DevOps. Yeah. Almost, uh, um, vibe oping if you want.

Yeah. Vibe ops. Okay.

Well, what dev vibe ops. 'cause you gotta have the dev in the ops with something in the middle. No, but in, in, seriously, it's going to be much more intention Oh, yeah.

Faced, uh, with, uh, a higher degree of trust. So I think that This is, this is, you know, I remember when HTML came out, all of a ci I was a coder. I was never a coder, but H-T-M-L-I could do then.

Yeah. HTML 2 0 3, 0 4 oh CSS JS script, you know, all these things came on. All of a sudden I wasn't a coder no more.

I think we're gonna see a similar kind of thing. You'll have, everyone could be a, a developer with vibe coating. Everyone will with AI will develop something if they need, but there will be the tools that the pros use, right?

That vibe coating, refined vibe, coating squared, or whatever you want to call it, where it'll be for professional developers. And, and that's, you know, developers aren't going away. They're not gonna be replaced.

They're just gonna be empowered With this. I, I, I agree with you. I think we will have humans mainly for, uh, just expressing intention and providing, uh, feedback loop.

Uh, there's that. I, I'll tell you what else, and I've written about this Uhhuh For, You'll Need Humans for the Creative Spark. AI is very good.

It's when you say, I want to do this, I want you to do this for me. I want you to create that for me. But it doesn't create the ideas.

Of course, The human brain still creates the idea. It's that spark of humanity that I think will always be The human is the guide. The human is the guide.

Yeah. Uh, yeah. But I'm, but the reason I asked you about next year is because I didn't think you would know what's gonna be next year, otherwise, why you should retire if you already know what's gonna be next year, retire.

But I would like to have you back on in July, maybe next year. We will talk about Swamp Up September 1st with Pleasure. All right.

Yo have Yoav, Laman, CTO co-founder helping wrap up our day two coverage. But we're not done. We still have a few more.

So stay tuned. This is Alan Shimmel for Tech Drunk tv. We'll be right back.

Thank. Hey everyone, it's Alan Shimmel. We're back here live at Platform Con Day in New York City.

Of course, this is just the in-person day of a week long virtual event that's going on. com. I forgot how many speakers and sessions there are, but there's a lot.

And I encourage you to do so. Let me introduce you to our next guest. His name is Sylvan Achi.

Yes. You got this right. A neighbor of mine from Fort Lauderdale.

We're both up here in New York. Um, Sivan, welcome to Tech Drunk tv. It's nice to have you on here.

Thank You, Adam. Um, not everyone watching this is gonna know who you are. Tell them a little bit about yourself.

Yeah, so, um, um, I'm a former software engineer, was an SRE for nearly 10 years. Um, then I was an entrepreneur, got an education training software engineer, and now I'm, uh, heading the rootly AI Labs. Um, so for this, we don't know, ROOTLY is an incident management and on-call platform.

So we competition to PagerDuty, uh, which, you know, I think you probably heard of. So we help businesses to manage their incident, and we're used by, um, you know, small companies and large businesses like Nvidia, Figma, Cisco, LinkedIn, and so on. Right.

So we, we help, uh, large, uh, large company, um, and their operation team to make sure that their incidents are, um, handle smoothly. And my role, uh, truth, uh, is to lead this AI labs. And the AI lab is a community led initiative where we, uh, work with team of fellows.

So, um, we have people who are tech leader in the industry. We have, as a head of platform engineering at Venmo, the former head of AI at Twilio and research students. And we work with these folks to really understand what can AI bring to the world of readability.

And it's applied ai. So we build prototypes, open source tools. We run, um, research and we write report, and we share all of this open source, um, on our GitHub with the community.

So really the goal of this lab is like, how do you use AI for SES or platform people? Excellent. So this time probably is a good audience for you.

Yes, It is. It is. Um, you mentioned you were on a, uh, a panel this morning.

Yeah. So we were on the panel with, um, Google search work, um, and, uh, Nvidia. And the goal was really to discuss, um, what's, uh, what's hyped with AI and what's reality applied to platform engineering.

Right. Like, uh, I think we hear a lot from, uh, the executive and CEOs from, uh, uh, model providers who are selling a GI or fully autonomous system. I think we all agree that we're not there.

And, um, I think especially for practitioner, which I think today is a lot of practitioner, they really want to understand what's true, what's maybe not there yet, and how can they really apply this in the, in their day to day job. Yeah. I sivan I think one of the big problems, especially for practitioners is every day it seems there's a new news story that some big tech company is doing a layoff and they're laying people off 'cause they're replacing them with ai.

Yeah. I think as we sit here today, very few people are actually being replaced with ai. Correct.

I think what it really is, is that these companies overhired Yeah. During COV and before. Correct.

And they need to cut back. They have too many people. And rather than just saying that, they kind of blame it on ai.

Yeah. And so AI gets this thing of, oh, it's taking these people's jobs. Yeah, Yeah, yeah, yeah.

New York State is now setting up a tracker. Jobs lost to ai, you know, and, and, and so you talk about separating reality from the hype. Yeah.

To me, that's a big, a big issue here. Now, I'm not saying that AI may replace people's jobs someday. I'm not saying that AI can help us or cannot help us.

Uh, to me today, AI is more of a copilot than a pilot. Mm-hmm. Does that make sense?

Yeah, it does. And, and so I wonder, now on the other hand, I'm talking to you, you run an AI lab. Yeah.

What do you see? What do you think? Yeah, so, you know, I think there are different type of AI labs.

Uh, if you look at, you know, the large stakes who are building this, uh, models, you know, these are like more like research, uh, researcher and PhD, and people who are like building all these LMS at ru we are really taking, um, a different stand where it's like really applied ai. So we use this tool to see how we can empower current practitioner, augment themselves, do their job better, and, uh, faster. And yeah, I agree with you.

It's like any tech new technology or tools, eventually it may replace some jobs, but maybe it's for the good. You know, like let's say before electricity, we had people going in the street and lighting the candle, uh, you know, for, for street lighting. Like, we don't use this anymore, but maybe that's a good thing.

Um, so for instance, um, uh, shortly we are really focusing on incident management, right? That's, uh, what we are about. And what we found is that you can really use the AI to help, uh, operation team to spend less time on managing incident because that's not something you want to do.

Right? Um, so I will share two main use cases where we saw, um, uh, you know, how this technology can help. One of them is incident, uh, triage and filtering, right?

You have all this alerts coming from a lot of tools, and you don't want human to be looking at this, right? So here, LLMs can do a great job at like helping to filter and cut through the nose. And the second thing is, uh, root cause analysis.

So when you have an incident and you need to understand what's happening, what's wrong, a human may take 10 to 20 minutes to like, gather all the graph, look at GitHub to see what was the last commit, maybe go on Slack and see what conversations they were perhaps on the project. With LLM, you can reduce this by like 80 to 90%. So instead of spending 10 to 20 minutes investigating an incident, it can be done in like one to two minutes.

And that's a huge, that's, that's a factor of 10, right? It's, and, and I think that is, at least in the interim, that 10 x is the goal. It is 10 ai, 10 x you.

Yes. And we've been speaking about this 10 x engineer for A long time. A long Time.

It's finally coming. Absolutely. Absolutely.

So it's finally coming through. You know what we didn't mention Rootly. What's the website?

com. com. Yeah.

And, uh, the, the platform helps you. Basically, we sit at the center of your incident, ments, um, uh, efforts. So you connect all your monitoring and logging, logging tools, uh, to our platform.

So your Datadog and Sentry. And, and we will help, help your authority team to orchestrate a response to that. So we will create for you, um, a team or Slack channel, uh, spun up a Google meet or Zoom room so people can, uh, share.

And then we embedded, um, a bunch of features that will help you to, uh, do the job faster. For instance, we have a bot that will listen to the conversation on Slack and audio, you know, and then if someone join an incident, you have a bot that you can ask, Hey, what's happening? Can you gimme an update?

Um, once an incident is solved, you have to write a postmortem or incident report. No ones likes to do this, so we automated this for you. Um, so yeah, it's like a very, like, basically when something breaks, SREs, go to Rotten.

I love it. Sava, it's a quick 15 minutes. It was quick indeed.

Thank, thank You. Thank you for telling us this. Maybe we'll get together in person in Lauderdale, come into our studio.

Yeah, I'm down. All righty. Thank you.

We're live at Platform Calm. We've got a lot more coming your way. Stay tuned.

We'll be back in a moment. ai Leadership Insights series. I'm your host, Mike, er.

Today we're with Brian Weiss, who's CTO for Hyper Science. And we're talking about, well, getting ready for Gen AI because it's a little bit more challenging than we imagined. Brian, welcome to show.

Thanks Mike. Really glad to be here. We've seen everybody kinda launch one experiment after another, but I'm not quite sure that a lot of that is making it into production environments.

And part of the issue seems to be is that it's not necessarily all about the technology, it's more about the rules, the regs and the cost and other factors that go into that. But what are you seeing? Um, I see that a hundred percent and agree with not only sort of the, the stats, but the sort of trend that, you know, we start out with AI being kind of a, a, a solution looking for a problem.

And while very, very promising for things like retrieval and summarization, the real rubber on the road now is, is is, um, data inside the enterprise, right? That actually, you know, tells you about the language of the business or a process. And of course, as soon as you do that, you, you're into privacy.

You're into understanding like where that data's being trained, how it's being used, how you get access to it. So we see, I see blockers in twofold to success of AI projects and adoption is one is is that sort of the, the, the concept of the unbridled use of AI to do every, everything and anything is, is, is, you know, needs to be kind of reigned in a little bit. Uh, and then sort of the realization that the, the hard stuff is actually in implementing to get you to get to the data and answer questions about the data you care about.

So I think the last stat I saw was, you know, that, that over, you know, 60% of projects that have kicked off to do something with gen AI has stalled and they stalled for, you know, a lot of the reasons that, that you mentioned up front here. So we're seeing it in spades. Um, you know, at hyper science we live inside the enterprise and we work extensively with really secure data.

So things like veterans claims, things like, you know, information that, um, is very, very specific to individuals, uh, whether that's department of defense, uh, those kinds of things are, are, are mission. And it's mission cri critical data where you can't be wrong, uh, when you're looking to get information out of a document set of that sort of thing. So there's the criticality of the information and the need for getting it right, that I, I think a lot of the early stage gen AI use cases are, are, are banging up against, right?

One of the issues that I think we're now confronting is the sins of our data management past. And we all have structured data that, um, we manage reasonably well. But most of these AI models are being, or need to be fed something that looks more like unstructured or semi-structured.

And well, if it was unstructured, we tended not to manage it all that well. So are we revisiting all of that stuff now and kind of, you know, dealing with an issue we probably should have been dealing with for the last decade? Uh, yes.

Uh, part of, you know, a lot of what we're encountering right now feels a lot like the early days of enterprise search, to be honest, right? I mean, enterprise search was, was not all about structured data. It was about, 'cause I can do a SQL query on a row and a column.

The question is what does this thing say, right? And how do I find the information in this 50 page document or a handwritten note? So we have, I mean, technology has traditionally struggled with all of that noisy information and it's kind of been a, you know, a north star that we've we're as you get more compute and now we have, you know, transformer models that can read things and do probability for what they understand and say and be able to respond.

It's another chapter in that. But it is the unstructured data that it's, it's kind of the same problem, right? That if you haven't put some guardrails and structure around that for who can see it, how you can use it, what you need to do with it, um, then bringing the technology sort of full, you know, full bore to that is, is can be a real problem.

It's a struggle. Like I see a lot of the common struggles in gen AI use cases that, um, were endemic to enterprise search, who gets to see the data, right? If, if I, if I load all this, this stuff up into my enterprise ai and does someone get to say, Hey, who makes the most money at this company, right?

Um, but that sort of document level security, all of the problems that are associated with who can see what and what's available and how it's available are all now trip wires in some of these processes. And while there doesn't seem to be a lot of regulations that's AI specific, uh, I hear folks will get down a path to a project and then suddenly they'll encounter something like HIPAA or whatever it is that they didn't think that they were gonna have to deal with. And suddenly they're like, oh wait, we can't do this 'cause it's gonna violate any one of 20 different regulations that are on the books.

Um, how do we kind of navigate that so that we're not wasting time building things that are not gonna be used? Uh, I have a really strong opinion about that. And that is you need to work with AI and modeling technologies that you control.

So you control what goes into the model, how it gets used, and sort of the providence of that sovereign model. So we, we work extensively in government industries, financial services, where that, that, that's predicated on that. And in fact, it has been the blocker to being able to use some of the broader capabilities of AI now where at hyper science, what, what, you know, that's kind of a non-negotiable.

Like you have to be able to explain where you got the answer. You have to be able to understand the ground truth data that is being used to, to fine tune or train the model, um, and be able to really own the outcome, uh, around secure data. So, so my my, I think there's a, there's a, you're right, there's a kind of a bifurcation happening here.

There are models that don't do that, right? Don't use them, right? Don't use them.

You use a, use a platform which allows you to select and tune and train and, and, uh, models which are accountable to not only the data they use, but also to the answers they give. Uh, and I I would say that they're, they're, you're sort of splitting two categories of models. Now there are those for which I can do that and those for which should look if I'm gonna use them, then I, I'm, I, I can't get that accountability or, or, uh, transparency.

So we, we have been building models for many years, uh, for in-house, in some cases air gapped environments, right? That look at, you know, say for example, uh, healthcare claims at the, at the Veterans Administration, right? These are complex boxes of documents that have handwriting and all kinds of stuff all over them.

There's no ter external modeling to usable there, right? We need to be able to be on, on site at the va. And, uh, I mean we're, we're, we've got models now that, that deliver, you know, AI results at 99% accuracy.

And we've taken the, you know, the processing time from months down to days. Uh, but all of that's contained. Like, like you can explain not only the answer, but also how it was trained and, and the way it's being used in combination with those techniques.

So I, I see it, I see the market maturing and, but you're, you're absolutely right, Mike. There's a, there's lots of 'em that stall where people get excited about using a frontier model. And then, um, look, the new the new InfoSec gauntlet is, is your AI review committee.

What model are you using and why? And what is it doing and who owns it? And where's my data going?

Like this is a, this is now the new normal, right? To have to really vet any kind of model inside an enterprise extensively. The other issue, or at least one other issue that I keep hearing about too, is people will get through the pilot and then they'll go into production and they will have grossly underestimated the cost of running the thing.

Yeah, yeah, yeah, yeah. I look, that's another market maturity thing, right? So if you think about it, the hyperscalers who are trying to, there's a sort of this big land grab to become the model that everybody loves and it's being underwritten, right?

And as soon as you have to think about using that at scale, there is an underlying cost that's actually very, very hard to accommodate, right? So people get excited, like, I'm gonna use this giant model to do this task that used to, you know, but why would you use a helicopter to cross the street? Like we we're like, I'll use that to cross the canyon, right?

But if I, I'm, it's not, somebody's gotta pay for it at some certain point. So you absolutely see these things like, wow, it worked really great, and then you realize that you actually scoped in a way which is just financially unreasonable. Um, so I, I see that all the time.

And, and you know what, what, what I'm focused on as, as sort of the, the composable platform at hyper science is using the right tool for the job. So, you know, let's use the CPU driven trainable models that are, understand your data and get you a really great result for the price. And then I can then stack lots of things that are way more complicated, read, more expensive, GPU driven, all that kind of thing.

But, but now I'm gonna start to decide like I want the best outcome for the right price using the models that are the most effective. But I see that all the time, super excited. Let's use the, let's use this giant model to do this thing.

And you realize like, oh my God, I just took a helicopter across the street and I can't pay for it. Like, why did I do that? You know, all the time.

Do you also think that maybe, you know, to your earlier comment, will AI push more people to something that feels like a private data center, whether it's on premise or a private cloud or something? And, um, we're all gonna be seeing a lot more of that activity rather than just relying on a public cloud. Uh, that's already happened and already happening, like this first wave of AI workloads.

Most of our clients, most, 'cause we're dealing with government entities. You're dealing with anybody who's, who's rightfully concerned about the providence of the data and the ai the workloads are going on-prem, right? They're going, these workloads are going.

So you, you see the industry responding to that problem. But these init, this initial set of workloads on private done data that needs to be managed, uh, carefully, um, is, is on-prem. Like, it's, it's, it is shifting, you know, that that sort of grand move that we all had to the cloud, like everything's gonna go from managed to just 100% public cloud and I'm gonna buy it by the minute and consume it.

And, but now all of a sudden it's a, it's a very, very different, uh, um, process. And, you know, we're in kind of a unique spot at hyper science. We, we deliver on-prem, we driven private cloud.

We have a, a FedRAMP high secure SaaS environment. Uh, we are e we are partnered with some, some, you know, the hyperscalers in, in, for example, Google's effort to provide a managed on-prem service of their, of their models, right? Is also embedded with hyper science.

But yeah, I, I think it's kind of the old is new again in that regard. And, and it's, it's a hundred percent understandable. Uh, Is, is this therefore gonna become something of a rich company's game because you're gonna need to buy the infrastructure set up those data centers, get all that data managed.

I mean, none of this stuff is inexpensive. So, um, you know, what can a smaller company expect to be able to do versus a larger enterprise that has the resources to drive this thing? Yeah, I'll go back to my helicopters across the street.

You don't actually need, you can achieve really high results, uh, and high performing results with narrow models on an ensemble which are actually cost effective for the task. So you don't necessarily have to, you know, only the, only the, the really rich people can afford the the machine, which, which will get you the, the right and the perfect answer. Like that's actually, it's going the other way.

What we're starting to see is that a composable architecture where, um, a combination of models that are cost effective and then you bring in the ones that are more expensive to, to do workloads that make sense. You know, I've got a 300 page, uh, credit swap agreement with nesta tables and handwriting all over it, and they're gonna be chunks of that that are really relevant for a, a, you know, a a large model. They're, at the end of the day, they're, they're, you know, they're, they're probability calculators for language and they're big ones, right?

But I don't need the, I don't need it to tell me the square root of 32. Like, I don't, I need a, I need a calculator from, from CVS to do that work. So don't ask that, right?

Don't spend your money there. But I, so I, I don't, and then, you know, the other point there, Mike, is that, um, look, the, the, the amount of innovation driving costs down, I mean, that's, it's, markets do that, and then technology market does it, it's the same thing we saw with the virtualization of CPUs and, and, you know, when cloud came out, like, you don't, so that, that is happening. And you'll see the, the, you know, the compute gets stronger and the price get driven down because there's so much pressure to make that possible.

So over time, I, I don't, I don't think we end up being in a, in a, in a class warfare situation here with it, right? The market will respond. And if people are smart about what you know about, about combining the right tools and not trying to get, you know, you know, buy a helicopter to go everywhere, uh, I think gonna be all right.

Will there be a shift in demand then, based on what you're saying? Because right now when I encounter people, you know, all they want is the latest and greatest GPU and they forget about the other GPU cycles, but before that, and those ones are a lot less expensive, and there's also other classes of processors. So are we gonna get smarter about all this stuff?

Uh, yes. We are, and I actually think that the, um, the market will create that for us. Um, so you're already starting to see companies that have sort of have bet the farm on, on being able to underwrite the GPU at a fixed cost, but yet give the, if you can, if I'm, if I, if you get unlimited use of A GPU at a fixed cost, what is my business model if I can't charge you by the minute?

And you can use it as much as you want. There's a, there's a sort of a, a vicious cycle that I have to get ahead of so that, that trend to sort of, everybody wants the latest and greatest and there are some artificial pricing things happening right now with, you know, fixed cost against unlimited use and things like that, that are, and you know, we're seeing people go outta business, um, as a result of that. So I think what happens is that, that everything just kind of, that you can't keep that, that you can't stay ahead of that curve.

It's almost a Ponzi scheme. Right? So what will happen is we'll end up with, um, we'll, we will end up with like, well, yeah, the good, the good enough model is, is good enough because I can afford it, right?

As soon as the really good ones I can't afford anymore. And we kind of artificially are being able to think we can afford them. And then once all of that shifts, I think people will quickly say, yeah, yeah, I, I I need the right answer for the right price.

Not, not, I need to use the greatest thing on the planet to get the same answer. Uh, Um, we also, you know, once again, seem to be thinking about security as an afterthought here. And a lot of the deployments that I've seen so far have, you know, significant vulnerabilities and there's all kinds of new ways to hack into these AI models.

Are we waiting on some sort of catastrophic event before we get serious about AI security? Uh, I hope not, but maybe I hope not, but maybe, I mean, it, it's, um, and I think, uh, uh, you can go back to other sort of inflection points in, in technology that've been somewhat similar, the internet, right? Things like that.

And, and each time that happens, you create this surface area, uh, attack surface for, um, for bad actors. And I, I do think we are, it is running very, very fast. Mm-hmm.

You're not wrong. It's running fast. And I do, I do agree.

And I I I am concerned there's risk in that. Do I think that, um, you know, it'll, it'll take a major event to, uh, to snap everybody in the line. I really hope not.

Uh, the trend I see in large enterprises is they are getting very serious with, uh, AI governance boards and security committees to try and keep up with it, uh, in the security industry. I mean, it's a, it's a whole new world of, of vulnerability. So I see enterprises reacting and trying to be sure of that and get ahead of it.

It's also, you know, there's so much new every day that is risk. Mm-hmm. Also, early on, at least it seemed to me, most of these AI projects were led by so-called Tiger teams and they pulled everybody together and they even had dedicated IT people who knew about infrastructure.

But is more and more of this AI workload gonna just be shifted over and managed by traditional IT teams? Or will we always need tiger teams? Yeah, you're seeing the impact of innovation, right?

On business structures. Uh, and it, i it will level out. I mean, you don't, you don't have to be an AI expert in order to be able to then look at hyper science.

We, our platform is developed for just ordinary business users to be able to train models, um, and get high performing results. So the data sciencey part of this is, is being democratized and productized, number one. Um, and then as we find those use cases that really matter to a company, I mean, there's a lot of stall light, you know, AI projects that are about, you know, building agents that'll do magical things someday if you can figure out how to justify the ROI.

And then there are are folks, and we're in this category where you're actually just creating hard ROI in the business. And so as, as businesses find the use cases that, that deliver, you'll, you will, you will see it codify into a sort of run rate IT function. I don't think the tiger team who specializes in all things AI is a, um, is a new and permanent sector of the enterprise.

Hmm. So having considered all these things, what's your best advice to folks? What should they be thinking about right now to kind of avoid what are se what are some serious pitfalls?

Um, look, the first one I say is, is know your ROI target when you go in, um, there are a lot of fishing expo just like, you know, science experiments and things like that that end up being an, and we all know the downfall of that is that, you know, you're, you can't justify what you've done or how you've done it. So there are, there are drill sites for real value with AI driven, um, you know, opportunities. And I would start there.

That's the first thing is let's just sort of understand the outcome and justify the business outcome and know how you're gonna get there. So I, I'd say that's the first one. Um, I would say, uh, if anyone tells you that one, one model or you know, one vendor's got the magic thing that's gonna do everything, then you prob just like think twice about that, right?

Um, the, you know, using models plural or using AI in a com, in a, you know, composed way is gonna get you better results and also more control. Uh, so I'd I'd say look at a composable approach, um, and then the last one, I think you called it already, which is data security and, and the providence of what that model is working with, what you're using it for. Uh, and, and ensuring, particularly if like, we're, we're in, we're in the, we're in the really, really high fidelity mission critical data business, right?

So, um, it's one thing to ha ask a, you know, a, a model to go do research for me and summarize things and all that kind of stuff. But if I'm asking it to, you know, process information, uh, and make decisions potentially that are mission critical on data, you can't really be wrong. Um, so you gotta think about the right, the right tool for the job.

What is that, what is that model supposed to be doing? And, and I would say above all, you know, managing security and privacy, et cetera, is that model accountable for when it's wrong, right? So when it's wrong, do I know, can I solve for it?

Will the model in the process or what the platform itself help me solve for that? Um, now look, maybe you don't care. Maybe maybe error rates are great, right?

It doesn't matter. So you get the sentence a little bit different. And what if you're just sort of generating content and reading books and doing summarization?

It's something. But when you're into mission critical data process, you really sort of look at the whole picture of, of what, what do you do with wrong? Or what does that pro that model, that process tell you about wrong?

And does it give you the tools to bring people, say, for example, to solve it and sit next to the model? I think that because the frontier of this is not models do everything, or AI does everything, it's a combination of, of governed models working with the right slice of the, of human intervention to ensure you have not only data quality, but security and governance and, and transparency to the outcome. So I, I gave you a lot there, Mike does that, I'm probably gonna have to go to ask GPT to summarize all that for me into the main points, but You know, I, I think you heard it here folks.

I think based on the goals, the risks and the cost, you gotta make sure the AI price is right. Hey, Brian? Yeah.

This being at the show, I guess that makes me a human, GPT too. There you go. Thanks Mike, everybody.

Thanks for watching the episode. You can find this episode and others on our website. We invite you to check them all out.

Until then, we'll see you next time.