The Great Security Debate is *back*! It’s been a busy year, but it’s time to get this show back on the air (and maybe on the road). Dan takes a break from the rat race, Erik took over the world, and Brian uses Elmer’s Glue to splice his network cables.

Topics in the show this week:

• AWS and Microsoft make the best cases for business continuity plans, the AI
• Is public cloud reliable enough? Should we all move back to local data centres? How can we reliably assess that risk?
• Want an AI Data Centre on your town? NIMBY vs Innovation!

We will be back every 2 weeks on Mondays. Subscribe on YouTube at https://youtube.com/@greatsecuritydebate to see our smiling faces as you watch, or in your favourite podcast application to listen on your commute or with your whole family around the radio.

See you on the 17th with more debates! And some entirely new shows coming from Distilling Security very soon, too. Subscribe to the newsletter on our website https://distillingsecurity.com to hear all about them

Links to mentioned articles and topics:

• AWS Outage - 20 October 2025 - https://www.bbc.com/news/articles/cev1en9077ro
• Microsoft Azure Outage - https://www.wsj.com/tech/microsoft-hit-with-azure-365-outage-b3ac0724
• 37Signals move from AWS to Data Centre - https://world.hey.com/dhh/our-cloud-exit-savings-will-now-top-ten-million-over-five-years-c7d9b5bd
• 100 Years Flood - usgs.gov - https://www.usgs.gov/water-science-school/science/100-year-flood
• Great Flood of 1937 - https://www.weather.gov/lmk/flood_37
• Impact of Jaguar Land Rover Incident - https://www.bbc.com/news/articles/c0qpl0v3gnzo
• CDK Attack and Outage - https://www.industryweek.com/technology-and-iiot/article/55091142/major-cybersecurity-breach-affects-auto-manufacturers
• Russian grain blockade against Ukraine - https://www.cfr.org/article/how-ukraine-overcame-russias-grain-blockade
• Saline, Michigan OpenAI Data Centre & Pushback - https://apnews.com/article/openai-inc-joi-harris-data-management-and-storage-microsoft-corp-oracle-corp-f25196fca5865ed79d94c972249a272c
• Racine, Wisconsin Foxconn and Microsoft site failures - https://racinecountyeye.com/2025/10/08/microsoft-abandon-1st-caledonia/
• Racine, Wisconsin What happened to FoxConn? https://www.nbcchicago.com/news/local/what-happened-to-foxconn-a-look-at-the-1-2-billion-spent-and-where-it-all-went/3759518/
• Gartner Hype Cycle - https://www.gartner.com/en/research/methodologies/gartner-hype-cycle

Welcome to the great Security Debate. This show has experts taking sides to help broaden understanding of a topic.

Therefore, it's safe to say that the views expressed are not necessarily those of the people we work with or for. Heck, they may not even represent our own views as we take a position for the sake of the debate.

Our website is greatsecuritydebate.net and you can contact us via email at feedbackreatsecuritydebate.net or on Twitter. Twitter at Security Debate. Now let's join the debate already in progress.

Yeah, so it's been, it's been about a year, I think, since we dropped an episode. We've recorded a few but have hadn't had time to edit them. They. I don't know if they'll see the light of day.

They're not timely anymore, so maybe we'll see. But it's been a, it's been a long year. I know that. Yeah.

Between Eric and Brian and I, there's been a lot going on and I guess I'll start just quickly and say, you know, I've. I'm now on a, on a, on a planned hiatus. I'm spending some time with my family.

I used to think that this was a phrase that people used, uh, as a cover for, uh, for. I had a bad breach, uh, and I'm no longer employed at the place or, or I, uh, I annoyed the guy I worked for and they let me go.

But I safely say this is, it is absolutely a true statement.

Not, maybe it's not always used genuinely, but there definitely comes a time when you say I need to, I need to spend more time with the people I love and the people that live with me and put up with me and come out of the basement. I had this moment about maybe five months ago when my Apple watch said, a new trend has been detected in your, in your activity levels.

Your new average daily caloric expense, like what, what you're expending, has lowered to 150 calories per day.

And that, that's not good because it means I, what it means is I, I got up, I rolled out, I went into my chair in my, in this office here and I, I didn't get out until about 7 o' clock and then I was so bedraggled from, you know, from days of Zoom meetings and things and, and that I, I didn't do anything more.

So zoom. Zoom is the silent killer, is what you're saying.

Well, teams really, because it also, Teams also gets you mentally as well as lethargically. But Zoom gets you just lethargic.

I thought we were gonna go on like, it's Zoom teams. No, Zoom, yes.

Zoom. Yes. I mean, honestly, to. To quote my father, everything in moderation, Daniel.

But.

But what it means is that we're re. Jump starting the network. We're rejump. Starting distilling security.

We're bringing the podcast back, as evidenced by this episode when it hits the light of day. And we are, we're going to work on some new things.

We're working on some sponsorships, we're working on some new shows which will likely show up after the first of the year. So subscribe now to the. To the network and you'll get notified.

And I'm really excited about getting to take a little time away as an operator and put some focus on other ways that I can help the community for the next few months. I don't know what happens after that, but we'll figure it out. Eric, what have you been up to for the last year?

May have stepped into a little bit broader role than just security. Moving into the CTO space, which is super cool. Transition. Very, very interesting. Going from being focused on. Not that. Not that security goes away.

Right.

But going kind of being focused on the security space and then expanding to broader it, which is super cool because if you think about it, security has a unique view across the organization. Right. You got to be plugged in. Good. I should say good security teams across the organization.

And we've, We've mentioned it in all of our past episodes. I'm talking about understanding the business because that gives you context to what's going on. Right.

So how being able to take that and expand it in, instead of playing the influence role of having more direct control over where we're going. Super interesting. It's fun. Economic headwinds, which makes things a little bit more interesting.

Right. In a consumer business. In a consumer business.

But hey, we've been doing that from a security perspective for the longest time.

True.

I. There was only a small period where all of a sudden budgets really opened up and then they contracted right back down, so used to this environment.

Wow. Well, congrats. That's a wonderful. Thank you. That's a wonderful change. And yeah, we like to. We like to joke that CISOs make the best CIOs and CTOs.

So in today's age, like Eric, I. I always joke like security. I don't joke. I actually, I'm dead honest when I.

Say you never joke.

15, 20, 20 years ago was like firewalls Right. And you had security over here, you had the IT team over here. And then like 10 to five years ago, it started to like become like a Venn diagram.

You know, the thumbs were kind of overlapping, and today it's completely overlapping. So to bring somebody that understood all of this and now that it's overlapping and teams want to sprint. Right.

And move faster, marketing teams want to do this. Right. AI we're going to go ahead and just write AI on everything we do.

Like, I, I guarantee, probably even in the cabinet industry, you're like, dude, we just sold a cabinet with AI we just wrote it.

These hinges have AI.

We got 15% uplift. Right. Because we got AI cabinets. And for those of you listening, we're probably going to add AI to the podcast. Know the great security debate. AI.

Right.

I'm already a gent. AI. It's not really Dan. It's just a model of me.

Yeah, 100%. And happened when Dan said, or Eric said, he graduated from the basement. Like, for those of you that last saw me, I was in the basement.

I. I lived in a cave. I had been moved down there by the family during COVID And what started is, we'll set this up for like a year, turned into like five years.

And I'm now putting together an office. So that's awesome. I'm like staring up. Right. And I don't have to like hide the fact that it's open ceilings and there's stuff hanging.

So you're saying there's hope I might someday exit the basement too?

There is a chance, Dan. There is a chance not to give too much away of where I'm at. I didn't graduate that far. I'm basically above the garage now, winning.

As long as nobody comes home during the podcast recording, you won't. No one will know the difference.

100%. So a lot of change, right? Yeah. And I would say for the better. So it's.

It's great to get back together because, Dan, I think you and I just saw each other in person 3 days ago as I'm trying to build out my network.

Note the nerdery over your left shoulder there.

Oh, yeah, look at that one cord plugged in.

It's a gateway drug. You'll be.

Try to.

You'll be there, put together the most broken backwards infrastructure I can and then come back to it with security protocols and be like, man, why is this network so messed up? Just as like a training exercise of something we can talk about on future dates.

Because you used Elmer's Glue to splice those ethernet cables don't do that.

Yeah, I got a lot of glue. A lot of glue. A lot of duct tape. A lot of tape.

Well, it's been a busy week guys. There's a lot going on in the world actually let's call it two weeks.

I don't know about you but this has been the best, the best advertisement for business continuity plans over the last two weeks that I've ever, I've ever had. I've had the.

Maybe I spend too much time with insurance people but I've been having the hundred, the fifty years, one hundred years and thousand years flood discussion with, with people about you know, to what level do you want to build your resiliency?

For those that don't know in the insurance world they talk about things like the highest flood level in the last 50 years and the highest flood level in the last hundred years and those in theory are less and less likely numbers.

That highest number over a hundred shouldn't happen in the theory for the next hundred years but in risk man, yeah, shouldn't happen but in risk management, yeah we have to figure, we always have to figure out how big we want to plan for because you could plan for the catastrophe of Godzilla attacks New York but the likelihood is pretty low. We hope but I think with both AWS US East 1 taking a major outage and a large chunk of Azure taking a large out an outage in the last two weeks.

I know that I, and you know we security people are notorious for getting, you know, getting amped up over things like this but I think I've taken a more risk averse posture in the last 10 days that I have in a long time about cloud services. I think it's starting to change how I recommend or to what level of flood. You know we start to think about protecting things.

I don't know, I think it's it I've seen, I've already started to or started to clamp down on more things. I've started to talk about you know, greater resiliency, greater redundancy, multi cloud strategies or even.

There was a great post by DHH from from 37signals about returning their stuff back out of the cloud and the cost save. It was mostly about cost savings but you know, are we to that point where cloud is no longer reliable?

It depends. I just wanted to be the first one.

That's my line.

I, I was sitting here trying to, I, I was trying to pull up this article because I, I used it a long time ago to talk about. And there was a city in Kentucky, not Louisville, small town. And when you just said like planning. Right.

But they looked at it and said, okay, was that like a once in a lifetime occurrence? Right. And what would you know if that was the highest. Where do we build to? Right. And they were like, yeah, but to build that high cost X.

And they're like, right, but in the next 50 years that the city grows, like here's what it's going to protect on this side. Right.

So there was the first stage where built up water comes in and then that second stage and what ended up happening, I think it was 20, 23 or 24 massive floods, massive rains came through and it literally was like an inch underneath and protected, you know, 50,000 plus people. And it was all because I'm just.

Envisioning Brian singing the old song from vacation Bible school. Like the rains came down and the floods. Yeah.

And it takes one person in the room to stand up and say, I disagree though. Here's why we should do this. Right. And they did.

So to your point, like when you're building in that resiliency and I think a lot of the conversation that I've seen lately is on that, and I always pronounce this wrong. The re. Patrization. Repatriation. Repatriation. Repatriation.

Bringing back the hardware side to bring back.

Yeah.

And you're starting to see that big shift now over the last year, year and a half. Right. Of not going full cloud. Yeah. You're operating in the cloud, but you also have.

What's your fallback in terms of a hybrid or a colo. Or what is it that you're doing?

This is why though, I, I say it depends in jest, but there's some seriousness to it. Because if we take the broad brush approach to our organizations, it doesn't make any sense.

But understanding the context of the different nuances on the business, like if I think about for us, manufacturing has to be up for manufacturing to be down super high cost because you're still paying people, you're still running machinery, but you're not producing any product. Right. But there's other components where can live in the cloud because it's a known risk. We can take some outage. It's not that big of a deal.

But then we're not managing on premise servers. We're not all the overhead that goes with that.

So I think we have to be cognizant of a multiple strategy approach depending on the context of what we're doing in the organization.

Yeah, it's if I was just like you just mentioned manufacturing. So just like manufacturing looks at their supply chain not just from a software side but says what's our supply chain risk?

Because one bolt could shut us down. One, this could shut us down. But let's reverse that in context. If you're a supplier, right talking to your customer, your oe, how resilient are you?

Because if you go down, let's use Jaguar Land Rover context, right. Like that 1.5 billion is not just a bailout of Jaguar Land Rover.

In order for them to stay in business, suppliers can't really go one month, two months not making any parts. Right. They may have other customers but if they had key models that 50 to 70% of their business was that that impact is far and wide.

So when you look at supply chain let's look at that trickle down effect of subsupplier. Now look at where those subs suppliers are located which is traditionally around the manufacturing plant within an hour radius.

So now then back that up even further. If all of that is down and companies go out of business.

Now the se, the store that made the sandwiches, that does the dry cleaning, that does this, that trickle down effect is massive. So it's the reverse on the supply chain risk of big oe, you go down. Yes.

You can't be down for a long period of time because every part you make is amortized into the capital investment. Right. For return on investment over time. So if you take a one time loss, you may that year but all your sub suppliers are.

That trickle down effect is massive.

But there's, there's two things that come to mind here. One, have we not learned about razor thin supply chain processes and this applies to tech as much as to physical goods.

After both Covid and Suez Canal issues I, we're. We're back to making the same exact mistake. Everything is so razor thin.

And then the other p. The other point is how there's the, it's the opposing debate position that if us, if Amazon goes down, I am the least of everyone's worries. This doesn't apply if you are a top tier, you know, if you are the, the top tier of your, you know, of your food chain.

If you're the, you know the, the main product in your space or you are Google, Amazon or, or Microsoft yourself. If the cloud goes down, then it's you that are down. And everybody else will, yeah, will say, well, Office 365 is down again.

Clippy can't help me, so I'm completely useless. They won't be thinking about your product being down. And I think that's kind of a naive way to think about it.

But it's definitely, especially for early stage companies, a way of, of I guess of, of thinking about what the minimum viable workload they have to do to get their product to market. But that, but that thought process persists beyond it.

Well, I think the salient point there is that it has to be a thought process. You're working through it. There's a cost risk balance that you have to work through. And frankly.

So I was sitting on a panel, gosh, last week and we were talking about AI and AI being used to automate processes and was the one ringing the bell that hey, we have, we have to be careful with this. We've seen this play out before that everybody, you know, buzzword automation. We have to automate everything.

Okay, let's go back to automotive retail. What happened when CDK got hit? Well, that was the only process to register cars in a lot of states.

All of a sudden a huge chunk of the economy stalled, right, because they, they automated it and couldn't get to anything. Well, we're heading in the same direction with AI. What do we know about AI? AI oftentimes is running on cloud infrastructure.

So aws, Amazon or that was duplicative. AWS or Microsoft take an outage. What happens to your AI model that's running okay, not available.

Have you built in the redundancy, the muscle memory to be pick up manual processes?

Is there an understanding that you can continue and this is the whole business continuity and understanding what processes are critical enough that we have to practice the manual piece of it to continue on.

Like look at food, right. From a supply chain standpoint in grain, right.

If you just use Ukraine as an example in Odessa, right, When the war first started in those shipments, because that's a big grain basket, right. And some of their export locations to.

I'm trying to remember, the two or three different countries almost went into civil war because there was a good month or two months where those grain shipments and other goods were completely cut off, right. And from the world of supply chain, right.

Food, water, energy, etc, when those things happen, panic sets in, people resort to self preservation and you're looking out for the good of your family when that unrest starts to settle in. So now case in Point what you guys were saying when people tell me, like, yeah, but one day you're not even going to need to go to the grocery store.

You're just going to order all your food online and then Amazon ships it to you. Sure, Amazon would love that. To have all that control and power. And then when Amazon goes on and people can't get food, what do you do?

Right.

Even if it's 24 hours, right? 24 hours, right. What if it was 48 hours? What if it was eight hours? Right. The eight hours is I'm complaining that I can't get the butter I wanted.

Right. 24 hours, though, could be, man, I need X, y and Z. 48 hours. And then compound that. Right. And then get into, say, a week, should that ever occur.

And this is where taking that full circle, like the idea of whether you call it backup, secondary, like, why small stores, small locations, being able to provide.

And I always find Europe very unique that you go to this store to get your breads, you go to that store to get your meats and cheeses, you go to this store for that. Right. There's specialty stores and products within the city. And then as you go out, we have much larger stores. Right.

Sometimes like during COVID you went in, maybe this aisle was empty of five or six different things because of coven. It was kind of like a bummer, right?

So maybe you tried to order it online, but you still had a path to get that when you try to build that all in to say, this is the best way to do something and we're going to use, you know, technology as a way for people to order and ship stuff. But now you'd be become dependent on them, right.

As the company that supplies your food because it's become an automated online service using their applications. Right. How robust is this world?

Well, and this is, I mean, I'm going to take a shot at you, Dan, because you were not, not at you, but you mentioned insurance companies. A lot of people gravitate towards looking at the actuarial data on, hey, historically, we know these happen.

For what we're doing right now in the interconnectedness of everything, there is no historical data for this.

No, but you can make some, you can make some assumptions that Godzilla is going to be akin to the 10,000 to the thousand years flood. And that, you know, a major outage at a cloud provider in a regional capacity is more like the 100 years flood.

Or maybe that actually it's becoming more like the 50 years flood because DNS continues to be the pain of all of our existence.

But I mean you, you can rank, you can stack, rank these things from 1 to 10 and then start drawing lines and say we think this will be a very infrequent occurrence based on history. We have 20, we have what, 15, 18 years of public cloud availability.

And you know, and you, we do have a model, we don't have 100, but you can start to extrapolate. This is not a, this is not a brand new four day old service offering we can make, you can let.

The AI tune itself too. I mean like Elon said, like if you really want to make the data great, just remove some stuff from history. Why do you need it? Right.

And then where you find gaps, feel free to fill it in with what you know. Right. Or what you think.

See I, this is, this is what, this is where I disagree though. I think if we were to just broad brush and say, oh, cloud's been around for a while. No, cloud hasn't been around for a while.

Cloud has changed so much. Everything is software defined now. The proliferation, the ability for one change to impact so much more is much easier today than it's ever been.

I agree with running the risk scenarios that we could say, all right, hey, we're all in Azure. What would it look like if Azure went down? Right.

Anybody who's doing single sign on, all right, everything ties back to Entra, we can start to play out some of those scenarios. Entra goes down, what is the impact? Yes. And then you have to start playing the time horizons back to what Brian.

Was saying about the standard business continuity planning, business impact assessment.

But trying to agree with this one. Where I struggle is trying to apply any type of idea of what's the potential of this happening.

Oh, I can clue. I disagree again, just because mainstream hasn't been in cloud.

Yes, cloud services have changed, but I've been, I've been in cloud first companies for 17 years now. Like it's been a large part of my, or a large part of my life where it is cloud first.

And yes, that some of the mechanisms and intricacies have changed, but it's not that. I mean the fundamentals are still the same.

The idea that it's a different data center, that it's remotely software controlled the net, that you have network rules. I mean all of these fundamentals have been there for well over a decade and, and longer. And these things are, these things are, they do have runtime.

Just because there's, and this is, this is a trap. I've fallen into I've assumed everybody has been in this world for that long, but it's.

The reality is there are people who are just climbing out of their basement data centers and moving things to public cloud services. But some of us have been there for a long time and we do know that there are runtimes on these.

So I mean I will argue that they're not the same, but let's be honest, cars aren't the same as they were 20 years ago. What do you call it? Floodplains. They evolve at a slower pace. But you still, you, you understand that these things exit.

They have existed and so we do have runtime on them.

Would you say cars are better now because they're software defined vehicles?

No, I want.

Case in point.

Well, I didn't say this is again, I'm not saying it's better or worse.

I'm just saying I don't want to beat you up here. But in, in 18 years, take yourself back then to one of the companies you worked at that did produce a finished good. Right. Versus a service company.

And where, here's where I'm going with that. And yes, this will probably be my third automotive reference.

Yeah.

For this podcast I can't make any.

Analysis against finished goods because I haven't done it in so long. I've been in software for so long that.

Right.

I wouldn't even venture to make a guess.

And this is where, like where I'm in agreement with Eric and it goes back to even the food analogies and everything else. Right. That.

So when you said like people are just bringing some of their stuff out of the basement, including me coming out of the basement, but those on prem data centers. Like when you look at big manufacturing sites, small manufacturing sites, the mom and pop, that just makes the widget. Right.

Like is there a need for the cloud now as they build, let's say a bigger online presence because they want to sell whatever it is they're making as an aftermarket product. Yes. Right. You're building this application. Are you going to do it on. On prem or are you going to hire a company out?

And maybe this is a different question. You. That's not the question of do we understand how cloud works well enough to understand.

Well to understand when we may see certain events happen and what the likelihood is that comes with run.

It's the that that risk nature though for them if their application goes down and people aren't ordering that Widget for aftermarket, etc. But if their assembly line goes down, that's. That's why the P and L? That's why plants.

Why manufacturing companies would break up the P and L for every single one of their plants and make them responsible for their, their P and L. And that's why then they, like if you take Toyota, General Motors, etc. Right. They have their sales division completely separate. Like sometimes they were completely different buildings. Right. Than this building over here.

And this was your design headquarters, this was your engineering headquarters, etc. Those plants themselves from a PNL standpoint. Right. And the ability to make those finished goods like Eric, to your point. I agree.

Like the, the risk modeling. Right. Of cloud for those companies. I. I don't think the data. Well, I guess I would say the risk modeling would still tell me that it's.

There's too big of a risk not to have some type of a hybrid system for those companies, that their core bread and butter comes from making the butter.

But you're talking about solutioning and you're talking about risk posture.

And those all come well after the point that Eric was making, which was we don't have enough runtime to understand when the frequency of some of these things happening. I argue we do.

The points you made are about risk posture, which every company, I mean, even within the software world, where you think, you know, my, my definition of break things and fix them later is way different than Mark Zuckerberg's definition, I promise you. But again, different risk appetites, different risk postures, different sales models, different revenue models, all the things you just described.

But none of those are part of understanding how frequently something will happen. You can still start to stack those frequencies and then you.

Each organization can draw the line as to where they're comfortable and how they're going to implement and how they're going to account for that risk posture. That's the big difference. And I want to make clear.

I want to make clear to everybody listening that that is not the same as we can't quantify how long, how often some of these things happen. We do. That's a fundamental. It's the baseline. We have a run. We can quantify it in the same way for a physical data center for dte.

Plopping my power every four days here. Yeah. Because a tree fell on a something. You know, those kinds of things we can quantify.

But then the variable portions are what you do with that, where you draw those lines, where you decide what you're comfortable with, what mitigations you put in, et cetera.

Agree to disagree. I think where we can quantify it is the our risk exposure or our. We can quantify our control of the risk elements on.

If we're thinking about the gamut of on prem to in cloud. Right. Because you go back to the. The physical data center and you know Dte there's ways that I can mitigate some of that risk.

I can put power conditioning. I can have multiple power lines run into the building to be on two different grids. Right.

There's ways I can mitigate that as I move into the cloud service. I would agree with you that we lose that control and we are beholden to somebody else and have no insight into how they're managing it.

But it changed. But the, but the potential risks change. They change to routing issue changes.

They change to DNS issues to an availability of multi zone to Godzilla crushing a data center like the the Amazon data issues always DNS issues. Everything is a DNS issue no matter what problem in the world. I promise you it was a DNS issue to start. Except for the tree falling on.

On the power line. That's probably not a DNS.

The new norm of what gets blamed now.

Oh, it is.

Turns out the AWS one contribution dice it was a. It was a very sophisticated attacker. That was.

Yeah. That got to our DNS.

No, you left information on a public file server.

Yeah, exactly.

So the idea of these investments into these data centers. You probably just saw the announcement for the one coming in Michigan.

Yes. Not too far from my house.

Yeah. And there's talks of others here in Michigan. You see the. I got to go on a tour. I shouldn't say a tour. Microsoft.

I did not go on a tour of your data center in Racine, Wisconsin. I did.

There isn't a data center in Racine, Wisconsin. They've abandoned the project.

No Foxconn abandoned the project.

No. Microsoft. About two weeks ago.

No way.

Yep.

Well, never mind. I stand corrected. It was probably because I visited. They have since abandoned the project.

So on the 8th of October I'll put in the show notes. Microsoft abandons original Caledonia Data center plan still committed to something in Racine County. Maybe a large statue of Clippy.

Wow. Because the infrastructure that Foxconn put in there.

Yeah.

Was incredible. So I drove, drove around it. Drove kind of right through the middle of it. And I mean you talk about all the pipes running from Lake Michigan.

You talk about the electrical like you name it. And then there's this like giant looks like a space center in the middle that would overlook all the buildings. It was.

I mean that was a massive footprint, massive investment Interesting. So where am I going with this?

The amount of money that I'm seeing dropped into these new data centers, specifically these new AI hubs, data centers and the energy constraints outputs like Michigan shares the Great Lakes with not only another country, but also some other states. And you know, they call it the Great Lakes water basin.

And there's supposed to be some kind of formal agreements on how you can use the water, bringing it in and out.

And there's always been controversy around that because Michigan uses water to pull out to one of the consumers energy facilities where during the night they pull the water up to a basin and during the day, during high peak needs, it runs out.

And although the net net energy is a negative because it takes more energy to pull it up than the energy gain, once you push it down and run it through the turbines, it goes back to peak usage. Right now you're putting data centers in where that usage, right, could be 14 hours a day, could be any time, right? Of when that's happening.

Like what are your guys's overall opinions on the massive infrastructure projects going in, the energy consumption? Because this goes back to where we started in the beginning, around where I talked about food and water and the importance of it energy also, right?

When your lights go out, right? Or you have those flickers and let's say something happens and a power station blows up. I'll use the aluminum industry right now, right?

Novelis had a fire overnight.

Anybody that was processing or doing anything with aluminum got letters in the mail the next day from every all the mills saying, hey, we're pulling back, we're not honoring any of your orders because this oe GM or Ford or so. And so we have to send everything to them, right? And it's like now where do I go to get it? And this goes back to the idea, right?

Where when I was talking about those little sandwich shops. So you go here for your meat, here for your cheese, etc, that's great. We did the same thing with the cattle industry, right?

Like how many slaughterhouses are there actually in this country now? It's not a hundred, it's less than 10, right? So when one goes down, it's a really big issue, right? And we love the idea of M and a baby.

We're gonna buy this sucker up, boost it, then we're just gonna ipo and then whatever happens after that, who cares? We all got paid dollar dollar bills, right? This goes back to the book when we were talking about Titan, right? Data is the new oil. No, it is, right?

And we're pumping it all into these areas to be refined and, and outcomes like what is this that we're actually using? It's not cheese. Can't eat it, right?

Supposedly it's going to make my life better, but that guy's not going to show up and plug in all that network equipment and even the stuff I'm googling. That shows me how. Dan, you're a genius. I reach out to you five text messages later, it's like I know what I'm supposed to do.

There's a human element here, right? Do we need these massive data centers to solve the next world's problems or are we just going to create more?

Well, I think the second part of the, the second part of the question is of course we're going to create more problems. There's always new problems. Anything you do that's new comes with new problems and new problems to be solved. Do we need it? No.

Do we.

Have a. Do we. Do we have a manifest destiny to it? We may have created that path now. And I think people are gonna. The, the power.

The power needs the, the data centers. There's a counter, there's a foil to all of it. And it's NIMBYism. You know, the idea. Not in my backyard. NIMBY. You know, these have to go somewhere.

And they're going in people's backyards, they're going in neighborhoods. And neighborhoods are pushing, by the way, big reason why they were seen. Project continues to fail.

Buy it failed Foxconn, why it failed Microsoft and others. There's some politics, I'm sure, but. But nobody wants it in their backyard.

The, because of the reasons like water use because of power use because of power drain. And when something goes wrong, they're the big revenue drivers. So you.

They will.

All the resources will get put toward them. And people, locals or locals are afraid of that. Do we need to be doing this? Maybe not. In fact, I don't. Honestly, I don't think so. Is it helpful?

I think once we get through the hype, you know, I'm not one who likes to quote Gartner on anything but their, their hype cycle. You know, curve is about right and you know, we're into the super hype portion. But you know what's following that?

The trough of disillusionment, which is the best part of any process. When people go, oh my God, we overbuilt, we overspent. What are we. How are we going to do with this garbage? And I think we're coming upon that.

But the Nice part is what comes out of it is refinement. You're going to get to a part where the. It's not just slapped on everything. I think there are reasons to do it.

Do we need to be building all these damn data centers everywhere? I still don't think we need Amazon warehouses everywhere. I can wait a day, just ship it a different way.

But, you know, same thing with data centers, same thing with other things. No, Brian got a delivery in five minutes, seven hours.

There you go. And then what happened? My Internet was out, so it didn't even matter, but I had to have them in seven.

So Brian is part of the problem.

Spend 20 more dollars and you'll get it in seven, seven hours. So what did I do? I spent 20 more dollars, of course.

Yeah.

So I mean, I think, do we, do we have to know, is there an opportunity that we come, we go through the natural course of process and we come out the other end with some, with some systematic improvements, with some life improvements as a result of it, but not nearly at the crazed pace that we're operating at now. Yes, as well. But then the big question is, does the end justify the means?

And Eric, I'd love to get your take on this too, because one of the things I read and what some people are talking about is, hey, this is all going to work out pretty good for us because everyone's going to dump all this money in. And it's funny, it's the capital markets basically paying for it, right? So then you do the reverse of that.

So it's us paying for it, but they're like, then when this trough hits, there's going to be all these guys. There's going to be so much open data center usage and it's going to make things easier and things are going to go back to being cheaper. Right?

It's like if you can find the.

People to run it.

I, I'm not convinced that this necessarily fits into the hype cycle model, because I think the hype cycle model is usually around a defined space that is built. You have multiple companies that enter it that overhype what they're actually the results you're going to be able to deliver.

And that's where we hit the trough of disillusionment. But I don't think that model necessarily takes into account the breakneck innovation that's coming. Right?

Because why, why are these data centers being built? The driver is AI. Why is this being driven? So, Brian, you hit the nail on the head. It's capital markets, right?

Public Companies cannot afford not to invest in this. Whether it's going to work or not. Totally irrelevant.

They have to to be investable, that if I don't have the next whiz bang widget coming out, I dropped in the amount of capital that's flowing into my organization and somebody else is going to pick that up. Right. So it's the market forcing this innovation. Now is it going to pay off? Remains to be seen. Right.

Because this is, this is almost if you go back to blockchain, right. It's, we've got this awesome solution for what. It's awesome. Yeah, yeah. But what are you going to actually do with it?

And then we're trying to find all of these problems. Not to say that it didn't solve some problems.

I think AI has a greater potential to do some of this and I suspect that as we continue to push how we're using this, how we're building models and everything, it's going to force us to continue to build these monolithic data centers all over to continue this innovation until we find out what it's actually going to do. The scary part for me on this whole thing is we're pushing all of this, build out all of this innovation on something that we truly don't understand.

How the data damn thing works.

Yep.

Yeah.

I mean I studied it 30 years ago.

Produces some results.

Yeah.

How it produces them, I don't know.

I studied it in rather great depth.

Like Dan, in terms of energy consumption, using that same analogy of consumers or whatever plant, right. Drawing that water up at night, pushing it down during the day, the net net is negative. Right.

Meaning like the overall energy used was greater than the energy gain.

So there is no model right now Dan, that can say the amount of energy we're about to use in these AI data centers is going to produce knowledge that is going to produce finished goods and capital that's going to way outpace. It's like well no, let's just take the amount of energy today and then times that by X as it just keeps increasing year over year.

As you use data centers, just AI data centers open up. Is the net net just to get here.

Like you're saying, yeah, we're going to incrementally get faster at our ability of innovations, problem solving, etc, but for what?

And I, I, I'm not asking you to say like oh but look at, even in the biotechnology markets or this market, the go to market being able to bring, let's say a pharmaceutical drug to market faster, you reduce that down by X, you know, it saves you X amount of money. Or you just pull the Chinese method where it's like, or because you're like Brian, with a really bad network, I just break in, pull all that out.

But somebody's got to innovate it the first time.

Time from 10 years to 2 years.

Like, but somebody's got to innovate it the first time. Couple of things come to mind.

One in the press release, just for those of you who don't know, and I'll put this in the show notes, the idea that data center was announced by turns out open AI here in Seline, Michigan, just south of Ann Arbor. Two million square feet.

But the commentary in it that caught my eye was DTE promises that the, that the power that will serve this is, is already excess capacity and won't affect locals. And it's not the second part of the sentence that worries me. It's the first excess capacity. Why, why don't we just bring down the capacity then?

It's not, this is not coming from renewable sources necessarily. If this were all just wind, if it solar, you know, different story.

But this is you have to burn more coal, you have to burn more gas, you have to burn more oil to make these, or make these things happen. That really affects the, the, the is it worth it Question that you asked Brian.

So I agree with you on that. But that's all predicated on the technology and how we do things today. What do we know about innovation? Innovation spurs innovation. Sure, right.

That as we innovate in one area, it causes a bottleneck in another area. You know, it's the theory of constraints that power in the way we produce it today is a constraint. Our power grid a constraint.

Is this likely to spur some innovation on that side? I, I would think so.

Well, it already is.

The order he is the work on nuclear fusion, the work on personal reactors, like all of that stuff is coming and is and is being accelerated by the need for more power.

I mean eventually the matrix, we're just harvesting it from humans.

And oh, if we thought electric vehicles were going to be so great for the world for the last 10 years, why didn't we put in that infrastructure?

Because what you just said is AI is going to be so amazing that these data centers, we're going to innovate energy solutions and infrastructure specifically for them. So this, and this goes back to what the capital markets view is. So the EV market is much smaller than the capital market of what AI can touch.

And that's where I'll say I agree.

But like, there was this time period of 10, 15 years where we could have started looking at energy a little bit different and how we could do things if EVs really were right. Sorry, I'm.

They really. But they really are. Get out of the U. S And the U. S. Politics And Evie's absolutely 80. I spent the summer in Norway. 80% of all sales are EV.

You know why? Because they built the infrastructure. And this comes back to Eric's point.

If you build the infrastructure and you support it, you're going to drive the innovation and the uptake.

See, I'm going to take the opposite on this one, though.

I. I struggle because I originally went the same way, Brian, is that I started comparing in my head to EV and kind of that proliferation, especially as we talk about capital markets.

Right.

Capital markets pushed the OEMs to start creating electric vehicles because that was the buzz. What happened? Fizzled out. Because by and large, at least in the US Market, the demand wasn't that high.

No, the lobbyists were too good for oil. The oil lobbyists were just too damn good.

And the right, big, big energy companies. So what happens when you build 32 massive AI data centers and then there becomes a trough and you've created all this excess energy?

Can you really dump it on the rest of the world? Are we going to use it? I mean, it'd be great if energy prices went down. For those of you watching, yes, I am using clippy.

And Clippy can bend in so many different directions. And it's driving Dan crazy because subliminally as I do this, I'm trying to change his thought process.

You know where I come back for? There wasn't a huge pressing need for electric vehicles. They were cool. I think there was a huge pressing need. Now, yes, you can get into.

Is there a pressing need for AI Another?

I think there is, because I think we are starting to hit the balance of how we could compute things and what we could do. Throwing humans at trying to synthesize data and come out with new elements. We've kind of struggled.

What's 50? Yes and no, Eric. Because the race for AI is because we got to get there first. It's like the moon.

Yes.

Because we got to get there first. And electric vehicles was kind of like, yeah, we don't really need to do that.

And what happened overnight, China flipped that on their head, that paradigm. Right. And then it's like, wow, we missed the boat on battery technology. We were going to depend on other countries to develop that.

And we'll just produce it here, etc. And it's really shifted of man. How would we do that now? Right?

Or we're just whiny because we lost.

Figures that out, Brian. Maybe AI, since we have the biggest investment here, figures out how we can actually see Clippy gone.

Eric just took Clippy, threw it over the shoulder. You're right.

I mean, there, there is, there is a bit of a. It's a global pissing match.

Yeah, yeah.

It 100% is. Because China holds the reins to the EV market. We now hold the reigns to the AI market.

And that. And that's what this is becoming. It's like the Space Wars. Right. And the thing is, AI is getting smart enough to know it needs to protect itself.

Like Terminator. So if you ask AI, like, how are you going to solve this? And are you important? They're like, you're damn right I'm important. Do not turn me off.

Keep pumping energy into me every day.

There's only one thing that ends a conversation faster than a Hitler reference, and it's a Terminator reference when it comes to AI. So, guys, unfortunately, we're out of time. We. We have to go for today, but thank you so much for joining us. Thanks, Eric, and thanks, Brian.

And thanks to you, the listener. We're so glad to be back. We're gonna plan for this every other week, dropping early in the week, Monday or Tuesday. We'll figure out a cadence.

But subscribe in your favorite podcast application.

Drop us a mail info@distillingsecurity.com the great security Debate is part of the distilling security network, and you're going to see the network portions of that build out very soon. Thanks again for listening and we'll see you again on the next great Security Debate.