Shawn Cordner (00:11.694)

Hey everyone and welcome to the Amplitude of Tech Podcast. I’m Shawn Cordner, Chief Marketing Officer of Amplix. AI and data centers have become controversial. It’s the topic that you’re finding in the news and that you’re hearing in political campaigns. Today I sat down with Ed Loveless and he’s from Zayo. He talked to me about the truth around AI and data centers and how it’s impacting communities. Listen to this one. It’s an important topic, and I think you’ll feel better about AI in the future.

 

Shawn Cordner (00:45.806)

All right, Ed Loveless, welcome to the podcast.

 

Ed Loveless (00:48.952)

Thanks, Shawn, for having me. Appreciate it.

 

Shawn Cordner (00:51.224)

So maybe take two seconds who you are or what you do.

 

Ed Loveless (00:54.958)

Sure. so I have a bit of a background. I have an engineering degree, started in systems and computer engineering, found my way, stumbled into telecom, and spent the last eighteen years working in the telecom space. first in network architecture and technical architecture, so really in the weeds of how things work, really learned to love and understand all this hidden infrastructure that exists in the world around us. And s over my career I’ve moved into more service and product management and today I’m in a product development role.

 

hitting up product development for Zayo for network connectivity and managed services. So really interested in how all the infrastructure becomes an enablement for businesses and customers and really passionate about all the different aspects of doing that at quality for customers at at scale for a national and international fiber provider.

 

Shawn Cordner (01:41.856)

So we’re gonna have a conversation today about a really hot topic. We’re gonna talk about data centers. before we get into it, why should people listen to your opinion on data centers?

 

Ed Loveless (01:52.142)

I think we have a lot of perspe interesting perspectives over time. data centers have been around for a long time. They’re not a new thing. it’s very interesting that they are becoming more visible and it’s more you know, top of mind. It’s in the news now. It used to be sort of this uninteresting thing that was just a dis non-discrete building in a in a part of town that kind of did its thing and no one worried about. We all rely on them today. it’s kind of very exciting to be able to look at the infrastructure that’s existed for a long time, that’s now becoming visible and talk about it and really think through where we can

 

How do we do this well? How do we do it at scale? we’re we’re going from, you know, the electrical era, the power generation era, the telecom era. We’re now, you know, entering the so called intelligence era where we’re looking at how do we enable intelligence and AI at scale for for the world? Everyone’s excited about this new technology. That’s bringing a lot of visibility to data centers. So we’re taking what’s traditionally, I think, a telecom problem and a data center problem and mash them together and doing it at at scale and doing it in a way that, you know.

 

good or bad, the public now is very interested in. so I think there’s a lot of relevance to what’s been done in the past to what needs to be done in the future and then all these worlds are kinda colliding together.

 

Shawn Cordner (02:59.606)

So you said everyone’s excited about AI, but that’s actually why we’re talking today is because not everyone is excited about AI. I think the public opinion has turned pretty quickly against AI and data centers specifically. So let’s start off with you know, you you kinda hinted at this. Telecom infrastructure for a long time has been part of our world.

 

I remember being young and you would see physical plant outside all the time. We’ve managed to bury most of it and as we’ve transitioned over to to fiber. But the central office was always part of the neighborhood. I didn’t know what it was until I was in my twenties and working in this position, but that nondescript you know, brick building with hardly any windows and, you know, they all look the same. that was the the Bell Atlantic you know, neighborhood point of presence or now it’s

 

Verizon, but this this has always been part of the conversation and cloud computing has been around for a really long time and that means that data center growth has been happening for a long time as well. So you know, you alluded to it, we I don’t want to bury the lead here. This is about AI, but why is this starting to saturate television and politics and the conversation that normal people are having as opposed to us, you know, technology weirdos?

 

Ed Loveless (04:25.646)

Yeah, and I think not everyone’s excited, positive about AI, but everyone’s excitable about AI. and I think you’ve you’ve got a dichotomy of people that are terrified of what AI could become, and you got a group of people that are completely over the moon at what they’re able to do with AI. it’s it’s unlocking a lot of capabilities. it’s it’s it’s fascinating the last couple of years, going from, you know, hey, this AI thing, it kind of can do some stuff kind of okay, it’s kind of cool, to I can’t fe I don’t feel like hey, take the weekend off.

 

following what’s happening because things are going to outpace my development and where the trends are going. It’s very, very hard to keep, you know, on on track with where everything’s going. I I think AI’s growth has brought a lot of visibility, a lot of kind of big like, what is my value in the world questions to people as they look at, you know, I’ve always done this job, this task, this is how I do it. And suddenly this thing comes in that can do it faster than me. There’s a lot of, you know, there’s a threat there that comes with that. And there’s a perception of threat, whether it’s legitimate or not.

 

doesn’t matter. There’s a there’s a feeling of uncertainty that is driving a lot of kind of fear. you have a you know economic, you have different job situations. So there’s a there’s a new technology coming in. There’s a you know in the world this change and where, you know, is AI threatening my job, those questions are starting to come up and all these things come together, you get a perfect storm of this new thing is the bad thing and I’m gonna blame all my, you know, troubles on the new thing rather than understand the nuances about it. And

 

No, I think there’s a lot of benefits. There’s also a lot of things that will change. And this is not the first time that’s happened, right? We go back in in in history and you know, there’s a lot of things that changed. not many people use the operator anymore. But thirty years ago, people were all up in arms about, you know, do we want to, you know, start turning down the operator servers? You go back a hundred years, there’s a guy whose job was to go tap on windows every morning as an alarm clock. You know, no one’s looking for the alarm clock guy role to be kept alive. technology and innovation has changed how things work.

 

which changes how we work. So I think there’s a lot of these factors, a lot of change. And change is scary for people, right? Even if it’s good change, it’s terrifying because what you’ve known and not trusted is now changed. So I think there’s a lot of this and just the pace of change is so incredible that doesn’t matter what you’re doing, it’s scary. and it might be the best thing for you ever. but it’s it’s it’s faster change than I think people can, you know, as humans we can adapt to quickly. And I think that change kind of makes everyone

 

Ed Loveless (06:48.792)

terrified. And then when you start seeing construction, you start seeing things where we’re now scaling. We’re going from this this cool novel idea, and it’s starting to actually show up in neighborhoods and cities. And people are saying, hey, there’s this huge new development and it’s tied to this tech company. It must be AI. and AI is a thing that, you know, my friends, you know, brothers, cousins lost the job for AI. This, you know, new thing, bad thing together, you bring it together. It’s very quick to bring, you know, sum it all up into something that’s very, very, you know,

 

high level and just it’s becomes the bad bad thing that’s happening, it becomes the scapegoat essentially. And I think that’s driving a lot of visibility and, you know, we’re a social creature. So a as these things get visibility we start to talk about it and it starts to spiral. Next thing you know, it’s on the news and people are out pro out protesting at nondescript buildings that may or may not be data centers. But they’re in their neighborhood and it’s something to go at physically to kind of express that frustration and that fear about as well. And so I say the fear’s not legitimate, but it

 

It’s a lot of change, a lot of things at once. And sometimes we need something to to look at and point to as a concrete thing we can represent that frustration with. Whether or not it is it is actually the thing we think it is, it’s often where we go to first, I think. That’s my opinion.

 

Shawn Cordner (08:03.168)

agree with your opinion. I the scale and the speed of this and I think also the the potential ubiquity of it. I mean it how many alarm guys were there that were actually impacted, you know, that were knocking on the windows to be your morning alarm? Not that many, right? But the the promise of AI is a promise of increasing productivity, reducing cost, improving performance of almost everything. Right. So we don’t really know what those boundaries are and they seem like

 

there aren’t many boundaries of who won’t get impacted by this in some way. Yeah. So I think that could raise the fear. Yeah, I know you didn’t intentionally set this up this way, but you you said it like there was two camps. There are people that are afraid of AI and then there are people that are excited about how they can use AI. And I think that there’s a lot of people that are both, right? There’s a lot of people that are vocally

 

scared of AI or rejecting AI, but at the same time they’re using AI and whether they realize it or not, I mean they might be using it in their daily excuse me, their daily life. They might be using it in their job. they might not realize that it’s being used in a lot of the technology that they are using day to day, right? So there’s just a who’s not being impacted by AI today. and I I think it’s a rational thing to

 

both be excited about the promise of AI, but also afraid of what the future could bring for AI. So let’s let’s try to separate some of the the fear that’s rational versus maybe some of the irrational fear of this. So maybe just kind of starting off with what I think is really the genesis of of where the public rejection of AI came from and that was the founders of these frontier models. They were out there very publicly.

 

Talking about how AI is gonna take over people’s jobs. We’re not gonna need to work anymore. There won’t be any jobs. We’re gonna need universal basic income. Then they started talking about the dangers of AI and how this could potentially kill us all or take over the world. And they’re really using those those hyperbolic terms. I I consider them hyperbolic, but in in your opinion, do you think that they kind of set the stage for how people are viewing AI today by how they talked about it before?

 

Ed Loveless (10:23.594)

yeah, for sure. Like these guys, their job is to grow as a new industry, a new technology as fast as possible, right? They are excited, they’re over the moon. This is all they do with all their time, right? So yeah, they’re and they’re all trying to outpace each other. So it’s not just one one firm that’s doing all this stuff that’s, you know, using this language. It’s a bunch of competitors that are all trying to out compete each other, you know, and outscale each other and you know

 

This is one of those things where all news is good news for them and any kind of, you know, visibility kind of shows the importance of it and builds and scales the demand for their business. So I do think there is a bit of a bias there to say whatever you gotta say to get visibility to these things. I do think there’s some truth to that stuff as well. It’s but I think there’s a bit of both and there’s maybe an, you know, unbalanced incentive to show all these different extreme options to get more and more attention and more, you know, people

 

going behind it and people going against it to really just bring visibility to it, l to really ramp the whole technology sp space up. And then, you know, one goes faster, the next one’s gotta go faster. You’ve got this, you know, ever climbing rate of speed that they’re trying to meet to outpace the competitors. So the language they’re gonna use is going to be very strong. and people in that space, you know, you get burned out, you look at it, y you can get very overwhelmed very quickly if you’re living in that space as well. So I think there is a natural inclination to to overemphasize potentially.

 

but it doesn’t mean that these aren’t real outcomes that can happen too. There is potential for these things. I don’t think we’re there yet, but I think the speed of change is what’s making everyone go from a that sounds really unrealistic to, wow, if they did what they pulled off last week and we extrapolate three years out, I don’t know where they’re gonna be. That’s scary. so I think it’s it’s a very natural reaction to be overwhelmed why by it. I think it’s also a very natural habit to speak in grandiose terms about capability and things like that as well.

 

and we are seeing these inklings of things in production that are doing incredible things we didn’t think was possible. And that’s you know, all those things together, you know, you get large language talking about it and you got some pretty large fear coming out of that and we don’t all know what’s true yet. until time kind of proves that out to us. So

 

Shawn Cordner (12:39.542)

I think you kinda hit on something there too where it’s it’s the unknown of the magnitude of the impact of this. I mean you said i you know, they’re extra extrapolating and I think extrapolating is exactly the problem here. You can’t really extrapolate. There is no comparable technology that we’ve seen, right? So I I don’t think you can I I mean, literally every I every time I ask a question of an AI expert, you know, to forecast what’s gonna happen in

 

I would start off by saying, What’s gonna happen in the next two to three years? And then it’s like next one to two years and then I they still wouldn’t answer the question. So then it’s you know, six months. Now it’s like two weeks. What’s gonna happen in the next two weeks? ‘Cause it seems like we’re on a one or two week news cycle where all these crazy headlines are are about AI are are hitting media.

 

Ed Loveless (13:29.346)

Yeah. I I think if you had someone that under an NDA, they might be able to tell you the next two, three months. Everyone else is guessing in weeks, right? There’s we don’t know what three years l looks like for us. It’s you know, it’s completely unknown. And I think very few people called where we would be at today six months ago or a year ago, right? So I think that’s further causing the the the issue because we’re we’ve seen we’re wrong in the past and we underestimated as capability. And now we’re looking in the future and saying, Well, I was wrong in my prediction then it

 

went so much faster than I expected. And maybe it went as fast as they said it would go f you know. But as we look forward, it’s hard to predict what that’s going to be and what that’s going to change for us. And when you have a a good demo of something that works flawlessly, like anyone that’s used AI, you go in there like, wow, this is incredible. And then you you dig into it and use it again. And it’s like, wait second, this was completely raw. I used it the other day and I gave it all the good data. Like this looks phenomenal. Then I I I checked some numbers and like, hey you didn’t use the map you didn’t give the

 

The hourly rate I gave you, you didn’t use it. It’s like, sorry, I should have done that. It’s like, okay. We’re we’re talking with this huge thing that can do all this incredible stuff, and then we’re catching these, you know, stupid little errors that any human would have caught if they just follow basic instructions. So there’s this weird like, wow, it’s capable of everything, but a sanity check of it’s got some really basic low level challenges still to be worked out. And it’s just kind of like, are those gonna be solved? Or are we just gonna be doing crazy things at scale with, you know, crazy errors that no, oops, sorry, I didn’t mean to

 

do this terrible thing to this person because I, you know, forgot your instructions kind of thing. And the more we see outbreaks of AI doing things it shouldn’t do, you know, one on one side it’s like, wow, that’s incredible, it did that. And the other side it’s wow, how are we having like common sense like with humans, we have rules, we have, you know, safeguards, we have all these things we put in place. How are we doing that stuff, right? And we’re looking at this, you know, these two extremes like they’re not coming in, they’re not converging in balance right now. And I think that’s what’s driving a lot of fear is just like, hey, we’re letting these things go nuts and we don’t

 

fully know how they do it or what they do and they’ve got no safeguards, and yeah, that’s gonna instill a lot of fear. I think that’s just a ver that’s a very rational response.

 

Shawn Cordner (15:35.01)

Yeah, a hundred percent. I and you know, there is no regulation right now. There’s a l a patchwork of state level regulation.

 

Ed Loveless (15:41.134)

So how do you regulate it, right? It’s not like we’re choosing not to put these safeguards in place. It’s we’re we’re kind of figuring this technology. It’s it’s it’s it’s weird because everything else has been this very prescriptive, well figured out technology road where we kind of we get to the point and we know the guy that wrote the code and we go talk to him, so yeah, I I made a decision around that. This is, you know, something on its own inferring things and we gotta somehow teach it, you know. It’s almost like a a toddler that we have to train and and somehow discipline or something like that to keep it in line and

 

We’re still trying to figure out what that looks like and how to actually enforce that. So different approaches, but I think that’s causing a lot of the the issue. It’s it’s not that it’s like this cool technology that’s well constrained, it’s it’s cool c technology that’s unconstrained. and we’re kind of liking what it can do, but we’re also a little terrified of what it could do.

 

Shawn Cordner (16:26.082)

Yeah, absolutely. I I think something else that’s going and I guess I want to make the case for what you said, which is that essentially data centers are a physical manifestation of this fear and anxiety around AI in general, right? So it’s something that is physical, that is present in a lot of people’s communities that they’re able to point to and say, This is the problem. This is this is what I don’t like.

 

So to that end, I think there’s also been a bit of a movement to reject billionaires and to reject the tech industry, right? And I I think it’s a a specific type of billionaire that people are really pointing to and they are from the tech industry and maybe it’s it’s that bravado that they have or, you know, how present they’ve become in our media. But a lot of people don’t like that and a lot of people point to them and they’re associating with, you know, AI with

 

those billionaires. And so there is I I think that is also setting the stage for this anger and this fear towards data centers.

 

Ed Loveless (17:34.114)

Yeah, I think most people don’t have a billionaire and speed dial rate. It’s very impersonal. It’s very easy to blame someone you don’t know. Right. You don’t have to go and face them at the family dinner or anything like that where you have to go and kinda have a really well thought out relationship with that person. It’s easy to just kind of point to them and say, Well, let’s blame these guys ’cause they have lots of money. We don’t have lots of money and they’re doing the new stuff and it’s you know, it’s scary and no one knows them anyways, so no one can back them up. so it it makes a lot of sense. It’s I think it’s a very normal reaction to that stuff for sure.

 

Shawn Cordner (18:02.478)

And I do think that people feel a little bit burned from social media. I mean, we we had a very similar kind of thing with social media where it was a technology that rapidly grew and expanded and became ubiquitous without regulation and without really understanding what it was or how it was going to impact people and society. And it didn’t necessarily have the most positive impact on people and society. I’d say that we can all agree that there’s

 

A lot of upside to social media. There’s a lot of benefits, but also a lot of downside that I think are are creating challenges in our society today that I don’t know that we’ve got a way to get the cat back in the bag for these things. So here we are on the precipice of another one of those technologies that we don’t really understand. We don’t know what it does to us. We don’t know how it’s going to impact society and yet we’re full force ahead.

 

Ed Loveless (18:55.468)

Yeah, and if you go down that road too, you’ve got generative AI, you’ve got social media. Social media was bad enough where everyone had a a curated perfect life, right? Where, you know, they’ll spend in two hours getting that perfect shot and post it online as if it’s just a regular accidental photo that got taken, right? Now you’ve got generative AI that can automate that whole procedure. So it’s not even just look at my friend’s life, it’s curated, it’s look at my friend’s friend’s life, it’s all AI generated, but I don’t know that because it’s done so well. so it’s

 

you know, we’re we’re also losing this trust in everything that’s not physically in front of us, which I think is also undermining all of this stuff too. It’s not just, you know, this is cool and it’s machine learning, it can predict things better. It’s this thing can generate an alternative reality that I can’t tell apart. And that’s very very

 

Shawn Cordner (19:38.508)

Yeah. So you’ve got this, I think we both agree, somewhat rational. There are definitely some irrational, right? But some rational fear and concern about AI and a lot of unknown and a lot of confusion. And so media has started to report on this because real community concerns have started to bubble up onto their radar. And I think they found that when they wrote articles about this

 

It got a lot of clicks, got a lot of views. So this is the incentive structure for our media. Like it, don’t like it. That’s not what we’re here to debate. It is what it is. So I think you get some sort of media contagion here where these headlines start to propagate and grow. And now this is something that the media talks a lot about. And then when you have media talking about a subject and when you have that subject resonating with constituents, you have politicians.

 

start to pick up on this, particularly populist politicians. And so now this is becoming one of the defining political issues for the upcoming midterm elections. So I I think this is how we’ve gone from this new thing to a div a a a divisive political topic in, you know, maybe record time. What do you think?

 

Ed Loveless (21:02.272)

For sure. It’s it’s I mean it’s to flip it on its head, it’s easy to take for politics to take whatever the latest hot topic issue is and make it about that rather than about their own political positions on things because it’s the new thing and you can take the the stand that gets the most visibility and the most reaction from your cons constituents. I I don’t think it it necessarily should be. I don’t think that politics needs to not be involved in that either. I think there’s a lot of there’s a lot of things that we gotta figure out very quickly here.

 

as we figure about how do we do this with balance that is good for the cons you know, the people in the community, the people that might be physically impacted or potentially in that area, good or bad. And then how do we also enable technology to grow and the evolution happy? Because I don’t think I don’t think is as much as everyone’s afraid of this stuff, I don’t think anyone is also saying, let’s stop all of it. We want to just be as we are and let the rest of the world continue around us as well. Because we we know that there is kind of an unspoken arms race globally around this stuff. We want to be part of it.

 

we just don’t necessarily want it in our own backyards or too much impacting our lives. You know, we want to kind of see it successful, but from a distance, I think. And yeah, I think politics is going to be a big play in that. and I think there’s a bit a lot to be said about prudent politics in that space where there’s a lot of forethought and, you know, how do we do this properly at long term? I mean, if you if you look back,

 

I I when we were kind of talking the other day and I I did some digging into early telecom and early power and I think there’s a lot to be learned from oil and gas, power and telecom because they’ve all gone through these these things of, hey, we found a new thing. It’s it worked really cool in this guy’s backyard. We bro did a proof of concept. It was really great. And then we just started going nuts trying to service everybody and we caused a lot of chaos, right? The the early power systems, they were just direct lun runs down the street to everyone’s h house or

 

our factory. and it was great. It was exciting. Everyone came to check it out. But when you started doing that scale, I was like, wait a second, this is ugly. I can’t even see the sun. There’s so many things on the power lines. It’s just it’s like I don’t like this. And so we started planning, we started doing, you know, distribution, electrical distribution, all these things and starting to think about how we build infrastructure at scale. So AI, I think, is moving from this cool experiment that started to scale into AI as infrastructure now. We need to start thinking about it as infrastructure and planning it as infrastructure.

 

Ed Loveless (23:20.162)

And that means looking at the problem very differently, which is where government politics come into play. there is some ro place to to regulate and provide some guidance around how to do infrastructure well. But not a I don’t think there’s a case to just stop at all, because I think that’s how you quickly get votes until people realize that the stopping actually hurt them in the process, right? If you say, hey, we’re gonna we’re gonna just block all data centers in this community and then all your neighbors get them instead and they get the benefits. If there are benefits to go with that, you say, look, well, we just shot ourselves in the foot by demanding one thing and not realizing that.

 

you know, there was benefits we could have had with that as well. So it’s it’s finding that balance where there’s the ability to to benefit from this infrastructure, AI’s infrastructure build out, but also to do it well so that everyone kind of benefits together. It’s not benefiting on the backs of someone else. you know, it’s it’s understanding how to how to do that properly. And that’s I think that is where, you know, ultimately government has a bit of a say in that. but we have to be very careful to not over regulate or under regulate so that we’re prudent, we’re careful.

 

but we also allow innovation to to scale very quickly and do the cool stuff that sets us apart globally.

 

Shawn Cordner (24:24.684)

I think we could also learn a lesson from globalization and how lost manufacturing jobs here and kind of became a a service economy without actual planning. And I I’m I’m careful to say that because I I recognize that in economics there’s often you know, unintended consequences when you try to engineer outcomes through regulation and policy. So I’m acknowledging that, but I I

 

think we could the lesson that we could learn is that we started businesses started shipping jobs overseas because labor was cheaper over there and they could save money, but we didn’t have a plan to upskill the people that were losing those jobs and to help them transition their careers and and give them the tools that they needed in order to still make a living as the economy was going through such a massive transition. And

 

now you you hear talk about trying to bring manufacturing back here, but that that ship has largely sailed. And those people still don’t have the careers that they could have had if we were were to you know have a plan to transition them. So I mean you you can agree with that or you could disagree with that, you know, from a political perspective, but just maybe from a pragmatic perspective. Should we be investing in helping people upskill to an AI economy?

 

Ed Loveless (25:49.366)

I I think it’s pretty, right? If if this is what gonna change, do you ignore it and hope that for hope for the best, or do you embrace it and figure out how you can work together in it, right? to your to your point, anytime we overemphasize a single variable and try and optimize for, all the other variables get impacted, right? And this is our opportunity to look at the industry where it’s going and say, Hey, we wanna do this, we wanna do it right, and we want everyone to benefit from it. We don’t want anyone to unfairly pay the cost of

 

and understand that we can all kind of find a balance in there that is, you know, fair for everyone and generally lifts lifts, you know, you know, rising tide lifts all votes. One of those opportunities we can all rise together or we can all drown and one person can rise. And I think that’s where a lot of the pushback is it this perception that, you know, potentially one group is rising on the backs of everyone else. And that’s where you’re getting a lot of the pushback. And there are things we can do to upskill and to look at how do we how do we use this stuff, right? We

 

I don’t think anyone’s realistically saying we can just ignore this, it’s gonna go away. It’s now how do we embrace this and how do we do it carefully? And how do we do it in a way that enables everyone to benefit from it? not just one group of people whoever that is.

 

Shawn Cordner (27:00.278)

Okay, so we’ve established took us twenty seven minutes to get there, but we’ve established now that there is a rational reason to dislike data centers or at least to be concerned about their propagation, right?

 

Ed Loveless (27:15.51)

I I I think there’s a rational reason to. I don’t think it’s it’s solely on one thing, but I think there’s a rational reason to dislike a lot of things that are are new and scary. but it doesn’t mean we need to be afraid of them, right? I think we just need to understand it better. And I think that lack of understanding is a a big part of this. it’s just it’s scary and it has you know we moved we moved everything to the cloud, you know, fifteen years ago. and it just disappeared. Like the rack disappeared and no one you probably

 

The number of people you know that have personally been to a cloud data center is probably very, very low. You know, probably count on one hand if you know anybody that’s physically been in one, right? We no one has seen the cloud in person. We just move stuff there to it. so I think it’s one of those things where we have to, you know, a address the fact that people are now realizing that these nondescript buildings that, you know, have a few cars parked outside and no sign on the front, maybe it’s actually doing something more than we thought. It’s not just a storage warehouse. And we have to open the doors a little bit and explain this is how things work, this is how it works, this is why we do things the way we do it.

 

And here’s we thought about as a community. And how do we how do we find that balance so it’s it’s understood? No one gets upset when there’s a telecom palestol on the side of the street because they’re like, that provides the, you know, the phone service or the internet service or fiber to my house. they’re not going in and looking at as the bad thing because we understand it, because we’ve it’s been there long enough. We you know, we as kids who grew up like these are the power lines, the you know, the phones are on there, the the power’s on there, that’s serves the house. And when there’s a car accident that that line falls to the ground, that’s why the power’s out to our house.

 

We understand that stuff. We got to get there the same place with the AI data center and all the infrastructure that supports it is, hey, this is kind of how it works. This is what that is. You know, it’s not scary. It’s what enables the stuff that we rely on every day to work. and we’re okay with that. And, you know, we’ll get there. Just it’s change. Change is hard and rapid change is very hard.

 

Shawn Cordner (29:01.078)

You you’re right about the fiber thing, except every once in a while something happens that reminds people that it’s a tenuous agreement that we have with the companies providing the infrastructure. Just a week or two ago, I’m in the northeast here, the US and we lost like three airports, three major airports because of a fiber cut. You would think that, you know, from where we’re at today and all that we know about technology and infrastructure that we would not have a single point of failure like that, but here we are.

 

Ed Loveless (29:31.522)

Well, in I think in that case, and I I don’t know the the outcome of it, but I’ll I’ll you know, being in this space and d work with a lot of large profile customers on high availability designs, there’s high availability in design and there’s high availability in practice. And if you’re not testing those backup options, you don’t know it’s broken until you find out the other one’s broken. so there’s a there’s a lot of trusting blueprints and plans and stuff that was built years ago and assuming it still works. And then there’s people that are more prudent to say look

 

we have it, but we don’t trust it. We’re gonna we’re gonna use it all the time. You know, in the SD WAN space, which I spend a lot, you know, last five years heavily in, everything is used, everything’s active, even if it’s just a a backup link. We want to send traffic on that because we want to know it works. Right. So I I think that is a a big piece of it is just, you know, it’s the approach has changed. we we trust things differently. it’s not a trust based on trust, is it’s a trust based on verify verification. and yeah, those things impact us greatly.

 

I I had the pleasure of getting hit by a tornado a couple of years ago. and it was it was small relatively sp speaking, but it meant that our entire subdivision essentially suddenly had no power. and they also turned the gas off. So even people that had backup power generators lost their gas services. So the things that fail, you suddenly become brutally aware that, I actually really depend on that stuff. and I didn’t I depend it in a way I didn’t think I depended on it. And I think AI is going that route too.

 

We think of AI, a lot of people I use ChatGPT I use Gemini, whatever the, you know, that app is. We don’t realize that it’s also underpinning a lot of other stuff we depend on that makes things easier for us to do. Even running this, you know, in this podcast right now, we’re recording an application that uses a lot of AI underneath to help you with your workflows and stuff like that. We think of it as a recording platform, but in reality, there’s a lot of AI underneath that’s helping expediate your ability to to edit and write a podcast. So there’s a lot of things that as a society we’re starting to become more religious.

 

resilient and reliant on that is actually underpinned by this technology that we see as a an out there not part of us thing and we we don’t realize how much we actually start to depend on already until we there’s a big outage. Just like cloud services ten years ago, you know, you have a big AWS there’s your region outage and suddenly everyone realizes all the stuff they love and they love to spend their day using is fully resilient reliant on some portion of that infrastructure. So yeah.

 

Shawn Cordner (31:53.112)

You’re right, but this app, Riverside has yet to make me a good host. So AI’s got a long way to go here. so let’s talk about the difference between the data centers that we were talking about, the cloud data centers versus AI data centers today, because I I don’t think the public understands that there’s a difference. So what is the difference and does it matter in the conversation?

 

Ed Loveless (32:18.476)

in short, it’s density, right? coming from the telecom side, my first time walking into a CO was kind of, you know, a bit mind blowing. I wasn’t sure what to expect. I walked in there, it’s all running minus forty eight volt DC. everything’s giant bus bars, you got huge copper bars on the side of the wall that are running, you know, thousands of amps of DC current. And you realize that, you know, a misplaced wrench and you’ve got yourself an arc waller. it was a very different environment.

 

If you go from there into a more traditional data center that people think of, you know, say five years ago a data center, you’ve got racks full of compute, you know, you’ve got high voltage AC power going into it. You know, maybe it’s two eight volts or or it’s one twenty, what you have in your house. Stuff that seems familiar, but it’s just a lot of it in compute. AI is taking that concept and both of those kind of mashing together and they’re changing, you know, the physics is starting to cause a problem. how much density can you get? It’s now a full rack of GPU. So

 

AI is going from this data center with a bunch of different machines that, you know, one’s the router, one’s the server, then one’s the the hard drive storage array. Now it’s just there’s an entire rack and it’s full of GPUs and all they do is crunch crunch numbers essentially. And those GPUs burn a lot of a lot of energy. So you you go from this, you know, maybe a kilowatt to three to five kilowatts in a rack in a data center as kind of a typical day to potentially a hundred kilowatts plus in Iraq, at which point, you know, the physics involved to be able to power that using

 

commodity electrical rate voltage and stuff like that starts to get really, really heat intensive, really interesting. I was reading about some of the new data centers are moving to like eight hundred volt DC busing, where they’re running high voltages right into the rack to s to skip some of the the losses on conversion. But a hundred kilowatts in a rack is a problem because that hundred kilowatts does a lot of work, but all that work turns into heat.

 

so now you’ve got this high densification that’s going from a tr data center to more like a warehouse or industrial machinery that’s generating huge amounts of heat and compute, and that heat and compute is generating a lot of you know, cooling issues, but also like a lot of electrical demand that’s constant or surges based on demand. you gotta do something with that stuff. You gotta cool it, you gotta get that either to the atmosphere or to some other use case for that stuff as well. So

 

Ed Loveless (34:34.847)

AI is really changing the game a lot because it’s the densification. Specialized warehouses that do a single thing that are extremely resource intensive is what’s changing a lot of the landscape. It’s no longer we can fit a data center on the street because there’s spare electrical capacity. It’s now we need massive amounts of new capacity. And then we can’t let these things fail. So we need huge EPSs. We need generators to keep them running. the downtime is just unaffordable. So you’re now putting in this whole plant that’s got, you know, megawatts,

 

potentially more of power generation capability that could fire up on a moment’s notice and, you know, maybe it’s running diesel or an aftergas, whatever that’s running, that has a whole environmental look to it. That is a huge problem as well. So it’s going from this kind of lower density, lower power density, which is still high compared to a lot of our use cases, to hey, this thing’s burning the power of a foundry. and it’s it’s using a ton of power and it’s generating heat. It’s not generating by products that we can, you know

 

ship and sell to somebody, it’s generating heat. We got to do something with that heat. So the the cooling, the power, that’s what’s driving a lot of visibility, I think, because we’re now exceeding what the grid can handle just out of the box. It’s no longer just another commercial site. So I think that’s probably the biggest picture is just, you know, this is now like an industrial process more than it is a commercial building.

 

Shawn Cordner (35:52.454)

that’s interesting. So is it mainly the training that requires so much infrastructure? Is it the inference? Or is it you know the being local, you know, having compute power local to certain regions? Like what what drives the the usage so high in the density?

 

Ed Loveless (36:15.394)

Yeah. so it depends. it’s a it’s a bit of a nuanced use case and answer. Every every scenario is gonna be different, right? at the end of the day, AI is is software with a power core, right? It has to sit somewhere, it has to do something. And that performance is gonna be tied to the the use case. So training, a lot of training is, you know, int intensive compute, intensive connectivity.

 

and just spending hours or days running models and training them. And then those models get loaded onto an inference location that does the actual, like, hey, you know, Chat GPT, can you tell me this? And it actually answers that question, right? It runs it through the developed model. some of those models are small. and some of them need to be fast, right? They’re when 5G came out and the ability to do network slicing, there’s a lot of focus on can we do compute at the bottom of towers? And there’s quite a few companies out there that are focused heavily on AI and compute at the bottom of cell towers. They might have a rack that’s, you know.

 

four feet square, eight feet high, and that’s all it is. Sitting at the bottom of a cell tower. It’s designed to be super low latency for very specific use case. Maybe it’s doing AI based routing or intelligence for, you know, electric vehicles. that’s a a very low density inference use case where we want low latency. Like latency trumps all so we have to physically put it near the stuff. Physics rules all at end of the day. big data centers, you know, they’re cheaper to build a bunch of in the same area. So they’re going to build them together. A lot of inference and training reds have the same location, right?

 

When you message ChatGPT, the fact it takes 10 milliseconds more, you don’t really care because you’re still waiting for a result, you’re okay with that, right? If you’re flying a drone real time and it needs to make a decision in real time, you need that thing to do it within, you know, microseconds. Or those milliseconds start to really count a lot for that stuff. So I think it really depends on the use cases. I think that is also one of the biggest opportunities we have is to put the things that aren’t critical for latency where they’re not critical. where power is plentiful, where

 

Cooling makes a lot of sense. where communities are gonna be, you know, ecstatic to have new infrastructure built to support them because you know they’re they’re underserved today. And, you know, these new customers and new demands in the region will drive a lot of benefits for them as a community. So there’s opportunities in this to do a lot of good stuff. simply just stacking and you know, double stacking data centers where we put them today isn’t gonna say scale. It’s not gonna solve the the demand problems of low latency inference.

 

Ed Loveless (38:37.234)

or the space and power challenges of training. Right. And that also adds a whole bunch of resiliency problems if you stack it all in the same place. power and and water is hard to to move long distances without a lot of cost and a lot of losses. fiber and connectivity isn’t so hard. we just, you know, regenerate the signal and go as far as you want to go to where you want to go. So it’s very easy to find really good power and water locations and give them really good connectivity. And that’s where we spend a lot of our time in my

 

you know, my day job is we build fiber and infrastructure to those those areas to enable places that had no connectivity to to be successful. And there’s a there’s a lot of community side benefits to that as part of that as well. So

 

Shawn Cordner (39:16.706)

Let’s let’s talk about the power and and energy use because you you know, you’re saying that they can put it in places that have, you know, favor favorable capacity or are favorable places to access power and water and we’ll talk about water in a second, but power specifically. So people are are complaining that energy costs are going up. So let let’s try to get to the bottom of this. Our energy costs I mean I know that they’re

 

going up, but are they actually going up because of AI? And if it is because of AI, is it because of the power usage that they come with? Is it because of the infrastructure costs and the build out for the power, you know, for the grid? or is it something else?

 

Ed Loveless (40:08.856)

think there’s a clear yes or no answer to that. I think there’s so many factors in play right now. Like electrical vehicles, there’s a lot of demand changes in how we operate, how we live, and what we as residential consumers expect today. so I think I don’t think AI is the the big scary monster that’s driving all the change. AI is the newest thing that’s driving a lot of demand. I think those are different. yeah I think it it’s nuanced and it depends where it is. I think there’s

 

There’s some cases where there’s been attempts to kind of build a data center or build d demand on the on the backs of others, right? Where you come in, you you use the last bit of capacity the town has, and you just pay your bill and you just expect the utility to meet to do all it needs to do. I don’t think that’s very equitable or fair to the broader group. and I don’t think that’s how we can scale long term. You know, that’s kind of coming in and just taking the last little bit for yourself and then leaving everyone else with the problem.

 

We can’t you know privatize the benefits of socialize the the cost of doing things like that. but we can share in both. if a if a data center comes into town and says, Look, we’re here to stay, we’re gonna invest in the infrastructure and we’re gonna cover that cost is building it and s and leveling that up, you know, that also could benefit that town by you know, enabling, you know, electrical charging and all these things for for vehicles that also may be a demand growth that’s happening over time as well. So I think the problem is more about how you

 

approach the the development of the problem than it is about specifically AI is generating, right? If if AI is allowed to come in, build a data center on your existing infrastructure and then, you know, the community as a large has to pay for the upgrade, well, yeah, of course your rates are going to go up and it’s caused by that sort of one selfish actor, so to speak. If they come in and say, look, we’re gonna be here for a while, we’re gonna invest and we’re gonna pay our our portion of the cost of this stuff and we’re all gonna, you know, rise together and they’re they’re willing to commit being there for a long term.

 

That changes the economics. It’s no longer a hey, we’re the new guys that are driving your pricing up. It’s just, hey, we’re covering our way, we’re paying our way, and you’re benefiting because we’re improving the infrastructure in the area. so it I think it really depends. It’s very regional, localized on how things are approached. there’s some really good actors that are doing some wonderful things that are are lifting everyone up in the process and everyone’s happy and benefiting. And there’s some that are, you know, attempting to do it the lowest cost on the backs of everyone else. And I think we need to think about the best way to do those things to balance it because

 

Ed Loveless (42:32.074)

One’s sustainable and one’s gonna allow us to embrace AI as as an infrastructure build, and the other is gonna be kind of you know, consuming what’s there and then making it everyone else’s problem, which definitely is gonna get a lot of p pushback. So I wouldn’t say it’s all AI driving it. AI is just the the latest thing driving a lot of the electrification demands that we’re seeing. you know, we first started electric vehicle, well, heat pumps and electric heating kind of make it returning us to be more environmentally friendly.

 

Drove a huge kind of rethink on demand and then electrical vehicles drove another rethink on that. And then AI and industry growth, really. You know, if we’re gonna do reindustrialization, all of these things take power. AI is just the one that’s kind of the hot topic right now. But if you want to start building big factories and bringing stuff back onshore, that’s gonna drive a lot of demand. All of these things are coming together at the same time, which I think is driving a lot of it. So, no, I don’t think AI is the the sole source of it, but I think AI is the definitely the loudest force.

 

you know, voice in the the scariest thing at the table and we tend to to look to that for it. So

 

Shawn Cordner (43:34.594)

Yeah, I think that yes, I agree. I from what I’ve been reading and hearing, infrastructure costs are driving up energy costs more than people realize and maybe more than the supply demand issue that you assume that it is. And it’s because there’s more demand now for the parts that need to go into it and supplies are you know, the supply chain for a lot of different things are have been impacted.

 

over the last five years, let’s say ten years, especially since COVID, right? We had a supply chain shock and supply chains were just starting to get back to normal at the time that AI infrastructure was really starting to take off and get built out. And so the raw materials for these equipment, there’s a supply demand pressure. The equipment itself, once it’s produced, there’s a supply demand pressure because there’s so much growth of in infrastructure right now.

 

And the companies are covering the cost of the infrastructure improvements that are needed to hook up to these data centers and to handle the capacity that they require. And that’s driving costs up pretty significantly. I would I I’ve been hearing more than just the demand of the actual usage of electricity.

 

Ed Loveless (44:55.33)

Yeah. And it it’s interesting you mentioned the supply chain issue. I I think that is the one area that I think has the clearest like AI growth is impacting the cost I pay for, you know, this computer part. If I want to buy myself a new, you know, SSD for my my computer, I’m I’m seeing that cost, right? I went to buy a memory card for my camera the other day and realized that, wow, something’s changed here. And yeah, it’s driven by the fact that someone much bigger than me has gone and bought up all the supply from some of the major major vendors because they have so much demand that

 

who’s left is able to, you know, jack the price up and extrapolate the most value from the market because it’s, you know, the supply is down, the demand is up. so I think that’s the one area I think most people have directly seen that relationship or when just the supply doesn’t exist. so, you know, we’re seeing who’s willing to pay for the last piece essentially and that’s what’s driving the the cost too up. we’re also seeing it, you know, certain raw materials are are short and those things tend to be used in a lot of stuff.

 

And if you’re building some complex electronic device and you have all the parts but one little chip, you can’t ship it. a lot of vendors I deal with, you know, it’s down to one part they just can’t get. And when that part comes in, they can finally ship the gear, but it’s driving up demand. People are willing to pay a premium because I need I need one right now, willing to pay for it. so you’re seeing a lot of kind of that responsive pricing based on demand. so yeah, it’s I think that’s the most clear cut. Hey, this is actually driving direct impacts to me that people are seeing for sure.

 

Shawn Cordner (46:21.27)

Yeah, and in the telecom space, if someone, you know, if an enterprise needed fiber to a a new location and that fiber didn’t exist in the ground, they would pay for that fiber to be laid and brought to their building. But if the service provider that was doing that build out, if they had to build out their backbone because they didn’t have enough capacity in that point of presence to handle the traffic that enterprise needs,

 

then you would be paying for that and that would be something that would, you know, come out of margins. And potentially if there was enough of that, that would drive up costs. Right. So Yep. We have a similar kind of thing here with AI infrastructure, except I don’t know I j I just don’t know. I don’t know how much the actual company that’s building the data center is paying for that build out of infrastructure to their location. And then

 

If it does impact the grid and they need to make improvements to the grid, who’s paying for that? And I want to use that as a a way to talk about community leverage because now that this is a political issue and now that this this is a media issue, I think the dynamic has changed a little bit and the companies building the data centers have a lot less leverage than the communities. and and that has been a shift where the communities had no leverage and the data center companies had all the leverage. So

 

Now I’m hearing about communities saying you can’t build a data center here unless you make infrastructure improvements that help us all. are you feeling that shift?

 

Ed Loveless (47:57.582)

I think on the fiber side, we don’t see that as much because we’ve you know, already be doing that a lot of the time. there’s some stuff in the news recently that we’ve done with some of the large AI vendors in the space doing huge capital investments to build thousands of miles of fiber. And when we build those those fiber routes, we’re not just building it and saying no one else could benefit from this, right? We’re doing those long haul routes and we’re intentionally building a capacity for the communities we pass through. A lot of the time when we have to work with, you know, at the end of the day.

 

Fiber on the ground, you have to work with every single inch of that path and get permission to put it there. and a lot of the time that is, hey, we’re gonna come into a community that’s never had fiber before. It’s middle of nowhere, it’s on a train line or something like that. And we’re gonna intentionally put, you know, a hut there that’s gonna allow us to break out and then that’s gonna enable us to support some local ISP that’s willing to stand up in the space, get a connectivity from us and provide connectivity to that community. there’s a lot of government funding around the world for that kind of initiative. And it’s you know, we’re not

 

We may not be interested in serving the residential user directly, but we’re interested in serving the people that are. and that’s one of the ways that we are able to get get places and get stuff done is because we’re willing to to do the investments and leave capacity behind for other people to benefit as as part of the build out. so yeah, the AI may be driving the build, but everyone along that path is also built benefiting from the increase in connectivity, the increase in resiliency. a lot of the times it’s you know, you don’t see anything different, but

 

That local provider just got resiliency, right? They got a second peering partner. They’re no longer gonna go down when the single provider fails. They now have a secondary provider. There’s a lot of resiliency that gets built when these things get built out that allows us local people to indirectly benefit from that stuff. So, you know, selfishly on the fiber side, we see that, but I think the same thing happens on the on the power provider side, all those things as well. You’re you’re improving the resiliency because these this infrastructure needs to stay online.

 

they’re willing to pay for that resiliency build out, which helps keep the lights on literally at your house. And, you know, it’s no longer where your house can lose power for eight hours and it’s you know, it’s fine. You just had you know, you had to light a candle to read a book, your your refrigeration, your your car couldn’t charge, you couldn’t work from home. Like the impact is much different now. power resiliency demand has gone up drastically. we can’t just assume it’s, you know, what it was in the eighties or the nineties or so on where it’s just kind of a well, it’s fine, I’ll just go to work, you know, kind thing. So

 

Ed Loveless (50:20.27)

I th I think you do see that and I do think some of this kind of pushback is forcing more of that collaboration. You know, community approval is now part of the permanent process, whether you wanna formally declare it or not. and it’s good to get ahead of that. It’s no longer, you know, acceptable to just say look, we’re gonna do something nice, we’re gonna buy some credits from some, you know, foreign entity that’s gonna do something good for the environment. It’s

 

We’re gonna make sure that you as a community are benefiting from this and we’re gonna do something that you will remember we did it for you because we did something to improve the community, right? I think a good analogy too is subdivision development. at least where I am, we have a lot of process around if you want to buy a buy a farm and turn a subdivision, you’re not able to load that cost in anybody but yourself.

 

you have to pay for all the permitting, the survey, you you’re basically walking to the city and say, Look, I have a fully developed plan you need to approve and I am going to do all the infrastructure build out. I’m going to build the roads, the sewers, I’m going to put all the water infrastructure in. I’m going to pay for upgrades because, you know, you guys need to you guys need to put a pumping system in to interconnect the sewer. So I have to pay for that. and when I’m done, I get to sell all those homes and the city gets to inherit a working infrastructure that is signed off, validated, and then they take over ownership in support of that.

 

it’s not built the cost of the local residents. it’s built at the cost of the developer. So the person getting the full benefit is paying the full cost. And the people in the surrounding area aren’t paying that as part of their own taxes. It just becomes part of a larger city. Now that city grows, those things change so that you may feel the the impression of all this growth happened and now my taxes went up may be more related to the fact that growth happened in the area. There’s just more people, more services being provided than it is that I paid directly for this subdivision.

 

So there’s always kind of that perception there, but the intent is to very much if you’re gonna benefit by building this thing, you’re gonna pay for the the benefits and cover those costs directly, which I think is an interesting model, right? As we move into data centers and stuff like that, they become kind of this, you know, this new threat that’s coming in and they have the ability, incentive to do it right, do it well, and to be accepted and embraced by the community, because then they can build more and and scale up. I think there’s there is a natural benefit there. And I think we’re you’re starting to see that shift very quickly.

 

Ed Loveless (52:31.32)

Some of the moratoriums that are in place right now are basically forcing a pause to say, how do we do this well? And it’s you know, you can’t do that for too long because then just like where can I go instead becomes the question. But it’s you know, pause to think and do it well. I think there’s a lot of advantage to that. and then you’re building the infrastructure. The town’s getting better connectivity, potentially better water solutions, a lot of times getting better fiber connectivity. You know, they’re not just just downstream as a residential service, they now have really good infrastructure.

 

Which means that other commercial businesses, other things can stand up in that place that maybe couldn’t be supported previously or couldn’t afford to justify that fiber bill to bring it online. I think there’s a lot of value in that for sure.

 

Shawn Cordner (53:10.712)

It’s a good example because the public perception is that these developments are costing them money and the assumption would be because of the infrastructure improvements that are required, but actually it’s because more services are being provided, but there’s still a a net growth or benefit, you know, to the infrastructure when these things are built. So people are mad because they see an outcome, but they don’t really understand that the outcome isn’t from

 

the thing that they thought it was and they also don’t understand that there’s a net benefit or at least maybe they intellectually understand it, but emotionally they’re not thinking about the net benefit to their community when they see that, you know, their taxes are up a little bit, right? So, I think you have a very similar kind of thing with AI today.

 

Ed Loveless (53:57.996)

Yeah. And and not to say there isn’t some bad actors that have you know made a bad impression. I think there’s there’s a viral videos going around where it’s like, how did they really get away with building that there? Like why did they do that? It makes no sense, right? I think there are a few that that are definitely causing a bad name for the the industry. I think there’s also a lot of great examples of things it’s like, wow, that was really brilliant. there’s, you know, Three Mile Island

 

buying that capacity for a data center. That’s a really interesting approach to it. There’s other places where there’s towns that, you know, the the smelter went out of business. And they’re taking that all that infrastructure and unused power generation saying, hey, we’re going put a data center on this smelting site. You know, it’s probably got some environmental issues and, you know, no one wants to live there. But what a great place to build industry level infrastructure for AI training, for example, right? You can put it there, you got power, you got cooling, all the things you need.

 

already in place. it’s not in anyone’s backyard directly. It’s not going to impact people, but it now kind of lifts up a community that potentially saw a lot of loss over the years as those industries shut down to suddenly have a lot of demand and and importance. I think there’s probably quite a few areas that twenty years ago had no data centers and were kind of, you know, not doing well and now are, you know, data center centers of excellence where there’s tons of infrastructure there and investment.

 

And I’m sure the the local careers and whatnot has completely changed. You know, there’s lot of white colour versus blue colour, like a lot of things have changed. But for the broader community, lifting that up and and brazing that community up by providing employment and so on. not to say that, you know, sometimes they over emphasize the employment numbers during the build phase where there’s a lot of, you know, outside trades coming in to do stuff. Yeah, it generates a lot of revenue for those trades, but the long term, you know, smaller group of people that operate the thing, that that is a group that has employment that didn’t

 

before in that area, which I think is really interesting as well. so it’s good to kind of find that balance and understand the different impacts. I think it’s interesting.

 

Shawn Cordner (55:52.844)

You you say that there’s some bad actors and some examples of those bad actors doing bad things. I think that’s the problem with the political and media environment that we have is people mistake edge case for average case. And that’s not always the case. But I w I’m glad you brought up the power generation issue because independent of cost, we have an issue with power and that is sustainability. Right. So there’s the question of

 

electricity generation and you know increased demand, is that gonna put increased pressure on the environment? are you starting to see that there are more creative ways that people are generating sustainable energy for these data centers? And are they tending to do it independent in the sense that the power generation is going to happen on site, or are they paying for power generation someplace and putting it back onto the grid?

 

Ed Loveless (56:48.174)

I I at the end of the day, it all depends, right? There’s every scenario under the under the sun is being used. I think cost drives a lot of the decisions around this stuff. So I think you’re finding that where they can drive more cost effective options, they’re embracing them. it’s not a hey, we just have to use this option. you know, if you’re in Arizona and you’re doing a data center, you’re probably really interested in solar and some of those generation things, but solar’s only during daytime hours. So you have

 

The the flip side of the problem for the people doing the actual engineering is we can do all this stuff that’s really, really environmentally conscious, but what do we do when it’s when it’s cloudy or it’s nighttime or it’s cold or it’s hot or whatever it is, and we have a we have a peak demand issue. So you end up having to build a bunch of peak demand capabilities that may not be ideal, but you then balance them out with things that are ideal when you’re able to make it work as well. renewable gener generation is variable. That’s that is the biggest challenge with it. nuclear is

 

A really interesting one. I’m an engineer and a big fan of of nuclear and that technology, but I know you go back the last sixty years, that’s been a pretty contentious topic, right? and it’s not because the average performance metrics of those things for health and safety is great. It’s because of the one or two incidents that brought everyone to kind of look at it, like, my goodness, what if this happened to us? Right. There was a lot of fear mongering around that stuff as well. I really encourage there’s a lot of revisiting nuclear and a

 

I re read some news about some potentially new modular nuclear generation that are sort of ultra safe, ultra low power options that could be done. So I think there’s probably a lot of investment and and discovery going on to look at options for that kind of stuff. the three mile is interesting as well. Like there is infrastructure that’s being turned down or you know, it’s you know, scheduled to close and there’s opportunity to put the investment into to refurbish it and bring it back to life for a much longer lifetime. I think those things are all great opportunities.

 

Nuclear is probably the best match for data center demand because constant demand, constant production is a really nice matching on that. and I think the demand for these things, if it’s done well, will drive better investment in those areas, which then provides cleaner, more you know, balanced power in those areas. Obviously, there’s a bit of a race right now. so there’s a bit of a habit to just get it online at whatever cost and having a bunch of multi megawatt diesel generators spewing black smoke out.

 

Ed Loveless (59:08.45)

does a pretty bad public image. and there’s a bit of thought maybe about how do you b better plan ahead on that stuff and avoid creating a a visual public perception issue during the the onboarding of a of a location. the the biggest issue I don’t think building the physical data center is the challenge. Getting infrastructure to it is the bigger challenge. So there’s a lot of how do we actually plan for this stuff in the future, right? We see this happening, we expect growth. How do we start thinking the next three, five years about building

 

infrastructure capabilities for future consumers that we can do the the slow work now and we can do the expensive work later to bring them online faster. that’s that is the one kind of contentious bit is that the speed of execution can’t catch can’t be matched with infrastructure investment. And I think that’s what’s causing the other kind of perception issue is the hey we’ve got to get this online right now and if we got to burn diesel for a month, we’re going to do that. And that definitely has the hey, this thing just, you know, is a smog cloud, that’s all it does, which may not be the

 

the truth long term, but it definitely has a you know, a bad optic for sure.

 

Shawn Cordner (01:00:11.884)

Yeah. I think nuclear is the perfect cautionary tale because it’s a technology that had huge benefits that really got sidelined due to media and political pressure and public outcry that ended up shaping policy. So media elevated incidents that really were not of huge substance to I guess you know higher levels that created a lot of fear, which

 

meant that a lot of p politicians had to step in and, you know, represent the constituencies whether it was founded or unfounded fear. And so you end up having the government put moratoriums on nuclear development, nuclear plant development. And so we missed an opportunity to have a lot of our energy coming from what is ultimately a clean and safe energy source. Whereas a lot of you know European countries haven’t, like France for example, did not.

 

Derailed the same way that we we did, and now most of their power, I believe, comes from nuclear. So i with AI, you’ve got pretty much the same factors happening. Public outcry, media creating fear, politicians, you know, using it as a wedge issue, and speak talk of moratoriums. I I I don’t want to divert too far from the things that we’re talking about, but do you think that the future growth and adoption of AI might be hampered because of that, the same way nuclear was?

 

Ed Loveless (01:01:38.558)

The potential’s absolutely there. the the the political thing’s interesting because it’s it’s the it’s not driven by facts, it’s driven by perception, right? political involvement and making things wedge issues are all about the perception of the people that are gonna cast the votes more than it is about the facts about the situation. And that getting people to understand the facts, I think, is the easiest way to close that gap. moratoriums there’s two ways to look at it, right? Moratorium is let’s just stop this thing. It’s bad.

 

And some people seek a moratorium because they want to stop something. Other people see it as a hey, this is pause. Let’s figure this out. Let’s do it right. and I think if you’re if you’re pausing to give yourself enough room to make a good decision well, and as fast as you can, that’s that’s beneficial. Stopping to do the right thing makes a lot of sense. If you’re pausing because you’re just trying to buy time to get votes and but push kick the can down the road and not address the problem, it’s not gonna go well because when that

 

gets lifted eventually, it’s just gonna be back to where you started, right? But if you’re using that time to figure out, hey, we want to do this, we want to enable you to do this. We want to enable you to do it successfully and equitably for everybody involved, there’s a lot of, there’s a lot of wisdom in that. but I think you’re probably gonna see regionally both of those things happening, one or the other. And the ones that do that moratorium well and come out with really good, well thought out things are going to do very well long term. They’re gonna have a good strategy that’s gonna enable, say, hey,

 

You’re welcome to play our sandbox. Here’s our sandbox rules. Let’s let’s do this well together, versus the ones that say you’re just not welcome here until the next guy says you can come in. I think you’re gonna see long term success. But it’s gonna be a balance between short term gain and long term success. And gotta, you know, there there is there is some level of urgency. because if you’re just gonna say no, it’s gonna it might be three months to three years, we don’t know yet. If I’m a data center provider, I’m probably gonna be looking at a model.

 

my other options, right? I I’m being tasked to go build. the demand is there. I want to I wanna fill it. I’m gonna find somewhere else to take that demand as well. So there’s a there’s a balancing act there of doing the right thing with urgency, versus trying to just, you know, head in the ground and say we’re just gonna ignore this for a couple of years and pause it, which might work until you that pause gets listed lifted and then you’re now back at the starting gate without the benefits of the people that thought about the problem and actually really approached it with some long term w wisdom.

 

Shawn Cordner (01:03:58.518)

Pennsylvania, I think, is an interesting model where it’s a moratorium unless you meet certain conditions. And those conditions are very favorable to the communities that those data centers would be in. One, for example, being you have to somehow put more power onto the grid than what you’re using, right? to to hopefully mitigate some of the cost of rising energy. So I want to talk about water next.

 

People are concerned about data centers and what that means to the water. There’s jokes and memes about, you know, every time you create a AI image of a cat riding a tricycle, you you use, you know, a a lake’s worth of water. is that true? Do these things just rip through water and are they damaging the ecosystem in that way?

 

Ed Loveless (01:04:46.324)

there is truth to it, but I don’t think it’s a universal truth. d can data centers consume Kobi CMS water? Yes, absolutely. But do do data centers have to consume Kobi CMS water? No, not necessarily. And that’s that is a engineering decision that needs to be made, right? Cooling is an efficiency problem. water is a fairly efficient way to cool things, but it doesn’t have to be. You can trade electrical consumption for cooling efficiency if you need to.

 

Right. That little hut down the street that ha serves your house with fiber. It’s probably got a, you know, completely passive cooling equipment that’s has no fans or anything, you a Pelch A or something like that. Super super small and it does some really cool stuff. there’s no watering use there. It’s just taking power to move move heat back and forth. We do that stuff, right? In your home you’ve got a heat maybe a heat pump if you’ve got a new heating system that’s taking a small amount of power to move a large amount of heat in and out.

 

a lot of data centers use that technology. Some of them can move that heat from the inside into water and then they evaporate that water to get sort of the heat loss and to cool it, but there are different ways to do it. It’s it’s really an efficiency problem and a and an electricity consumption problem. So it doesn’t have to be. I think I’m sure there’s some data center out there that throws a lot of water, uses a lot of water, and that’s being used as the kind of epitome of all data centers, and I don’t think that’s very true. you can use closed loops.

 

systems that don’t consume any water at all. the only water going into the building is the employees that need the break rib, you know, and the facilities. So it really depends and that really becomes a regional problem. Some areas, water for cooling is abundant and rich. there’s why wouldn’t you use it? It’s it’s efficient and it’s a much more good use of the environment. Other areas, water is the most precious resource and it should be the last thing you use to co do cooling. I think the answer that fundamentally is is completely localized.

 

And I’m really thinking back to the, you know, where do you put things? If if training centers, you know, if you really think about positioning those, there’s lots of places where you can optimize for those things. and if you really want to get creative and and think about it, there’s op opportunity to collaborate with other, you know, consumers of heat, right? You look at in Europe, there’s centralized geothermal plants that heat the whole city. Is there an avenue to take all that heat from a data center and provide it at ultra low cost for some remote community up north?

 

Ed Loveless (01:07:11.558)

or to to to run greenhouses in the Arctic in the winter or something like that. Something you never you would never pay for based on the energy cost. But if you suddenly have this free source of heat because it enables, you know, training and inference up in Alaska, that might just unlock the ability to enable, you know, fresh fruit and vegetables in an area that typically can’t get that stuff. So there are opportunities, I think, to get really creative. I think the generalization is wrong, but I think there is truth to the generalization in some areas.

 

And thinking through the most optimal way to do it is an engineering problem that needs to be optimized based on availability of of power and water and those things should be balanced based on the regional area. there there’s there’s no one size fits all for data centers, right? that’s why there’s data centers on the ocean and space, all these different things being contemplated because they solve some of these n engineering problems, but they introduce new problems on the process. So it really is hyperlocalized. You gotta think through where where where can this go to optimize these things the most way the best way possible.

 

Are we going to optimize for water? Are we going to optimize for power consumption? you know, are we gonna find an area that’s naturally cool? Bitcoin mining was a really interesting one. You suddenly had a lot of new entrepreneurs that could do bit mining bitcoin mining. Suddenly they’re going way up north, building those greenhouses in the middle of a f farmer’s field and just running fans through it because they could take the low temperature aside to do their cooling. if you take that, do it better, really think about it, there may be opportunities to do some really creative things that optimize for this stuff.

 

not every data center has to be in the middle of your residential neighborhood either, right? thinking through the positioning where it goes, where there’s infrastructure, it it’s a balance of all those factors for sure.

 

Shawn Cordner (01:08:48.738)

Yeah. I the water thing, I I know it’s a potentially r real issue and I know that there’s definitely technology to mitigate that, but if we look at the size of the problem compared to, you know, almond farming or golf course, the example you always hear is, you know, golf courses use way more water to water the lawn than they do than data centers do. So I mean, is the scale of the problem as big as it seems to people?

 

Ed Loveless (01:09:16.875)

I I think it can be and it definitely could become the same scale problem, right? these things are still minor consumers of water compared to other industries, but you know, any time you have an industry of a rapid growth in consumption, it does threaten the balance of things, right? And I think that is that is more relevant, I think, than the actual numbers. It’s really what is this changing? Is it is this replacing another industry? Is one industry going away and this one taking over? Or is this in addition to that stuff? Are we doing a net, you know, increase in water consumption, in which case

 

Yeah, I think we should really look at that stuff. It’s it’s gonna drive a lot of, you know, thought processes. if there’s things that this is driving out and removing, that’s also interesting because there there may be a net balance there to that as well. I’m not saying either way, but it is holistically a a larger conversation. and if water consumption is a big sticking point in an area because it’s not available, then don’t design based on consuming lots of water. You have to rethink your design and deployment or rethink your positioning in that area. I think it’s as simple as that. It’s just

 

being considered a the bigger problem. and, you know, it’s we’re now approaching AI as infrastructure. We need to think about, you know, AI as infrastructure and its impact to everybody involved. And like if you do that, you can do it well in a way that doesn’t cause those issues.

 

Shawn Cordner (01:10:30.508)

You mentioned that these things don’t have to be in populated neighborhoods, but why are they so frequently? And the you hear so much from those communities about the noise and the physical community impact too. how real is that?

 

Ed Loveless (01:10:47.85)

I don’t I don’t have exact numbers to give you how real that is. There are examples. I’ve seen them, you’ve probably seen them online where it’s like, wow, that that is does seem kind of annoying or frustrating, but I would flip it around. How many people live next to induc industrial factories today? and you know, they’re it smells, it’s not nice, it’s loud, they’re used to it. This is just the new use case in the industrial zoning. I think that’s more of a a zoning question is you know, how

 

How close are we letting industrial and residential coexist together? We’ve done it a l a long time when, you know, small towns that had a big factory, everyone lived around. They just put up with the noise because that was their job and it paid the bills, right? Now it’s it’s not their job and it doesn’t pay the bills and it’s annoying. so I think there may there may be some perception changes that are part of that as well. But I think I think it’s really localized and there’s probably a lot of historical context to why those things are there. and why it’s, you know, now it’s being

 

raised up as an issue because there’s the the benefit, the net direct benefit to my family as a source of income is gone. it’s now my neighbor and it’s annoying. and I think that that is part of the, you know, the consideration. And if if we’re building new things, that should be a consideration before we build, not a an after the fact, let’s put a sound wall up kind of thought process.

 

Shawn Cordner (01:12:06.572)

I don’t know empirically if this is the case with data centers, but historically that’s one of the inequities is that those kind of facilities would be built in areas that were of lower socioeconomic stature because land was cheaper in those areas, or it could be the other way around that those things existed and housing was put there because the land was less expensive and

 

Those people typically have less political capital and less actual capital to fight these kind of things. And I think we are, you know, starting to hear those voices. But do you do you know is that something that has been a issue with data center development?

 

Ed Loveless (01:12:49.842)

I I I personally don’t. But I will say that social media has become a great leveler in terms of influence, right? It’s no longer it’s not this is not being fought in the courts, this is being fought in social media, right? And social media tends to feed a lot of the news now. So whether you’re you’re high or low income or in the right location, it doesn’t really matter anymore. If you have a compelling video that seems, you know, I can identify that that if that was me, I would be unhappy with it. It’s gonna go viral.

 

And I think that’s definitely shaping the the balance of power and the whole conversation for sure.

 

Shawn Cordner (01:13:20.93)

Yeah. okay, so you talked about the concentration requirements. Do risks come with that as well?

 

Ed Loveless (01:13:29.872)

a hundred percent. And this is my in my day job, right? We’re trying to help people plan infrastructure. And every couple of months there’s some major cloud outage today. And everyone learns who who they’re reliant on and who their their favorite app is actually reliant on and whether they actually bought resiliency or not. and w we’re gonna that’s not gonna go away. And it’s not because we’re not building well designed infrastructure. It’s just there there are commonalities that will cause issues, right? And you know, look at Texas and the power outage that that happened.

 

couple years ago. That was not you you can mitigate some of that stuff, but it’s not just the power outage. When you have that major snow event, you also have some other things starting to fail, infrastructure, maybe fiber gets hit because of some of those failures. Like there’s all these compounding failures that we we we we hyper focus on the first one. We forget about all the small minor ones that happen as part of it that also can cause the outage. So we can build it back up power generation, can do all the stuff, but if we it just takes one thing to break to take it offline.

 

And ultimately the best resiliency is true field like geological diversity, right? figuring out how do you get into a different weather ecosystem to the point where you’re never getting hit by the same storm both times or you know, the same power outages, the same providers, the same kind of problem sets. And, you know, if we we keep building everything in Northern Virginia, that’s great. It might be very efficient to do that. And there are things that make a lot of sense to do that, but other things, if we’re building that new stuff.

 

It doesn’t have to go there. Let’s build some diversity to that, right? and yeah, it’s gonna be net new investments, but then that’s your opportunity to build it well, build it right. And then, you know, fiber can come into that, enrich that community and build that diversity. And you wanna build that stuff diverse as well, right? Be able to build infrastructure at all levels that’s diverse so that you’re you’re no longer trying to pr prevent the outage, you’re trying to pr protect the resiliency of the service. So

 

It’s okay, this one offline. That’s bad for us internally, but as the end user, you don’t notice it. it’s very different than a, hey, we’re gonna do everything we can to prevent an outage because we only have one. and you’ll see a lot of things are moving towards more resiliency planning. Resiliency in terms of business outcomes, it becomes more important versus actual infrastructure protection, where it’s a that’s fine, it’s down, we don’t care. We just we have other ones that they’re gonna pick it up, and you as a user will never notice it.

 

Ed Loveless (01:15:47.108)

it is a bit of a different thought process. You know, if you’re a small business or a medium business, you have your one data center, you protect that at all costs. So now you’re in the cloud, you’re in your data center, you’ve got diversity. So when you lose that data center, it’s it’s a bad day for you, but it’s not a bad day for your customers. And I think that’s driving a lot of kind of rethink about stuff. And AI should be doing the same thing. Let’s build diversity, diversify where it’s going so that it can, you know, be protected and not taken up by a singular event.

 

Shawn Cordner (01:16:13.122)

Can can this infrastructure be atomized in any way? Meaning, you know, using edge computing or maybe smaller data centers, you mentioned data centers that were attached to cell towers, for example. So, you know, is there a way that we can, you know, break this huge need for compute down into even smaller units than what a typical data center would be and find places that are less obtrus less obtrusive?

 

more accessible to infrastructure today, maybe even leveraging end users’ facilities so that, you know, we have a more diverse, more resilient, but also less expensive and less public you know, not impacting the public as much.

 

Ed Loveless (01:17:00.904)

I mean it’s it’s possible. The the downside to that, the trade off is cost, right? Every time you break apart the unit, you have to build all the stuff that supports that unit. So I think there is a cost balance to be done there, but there’s a lot of opportunity to do localized low latency stuff where low latency is important, right? And that that stuff can be pretty small. there’s a lot of really creative things we’re seeing out there, even some of the customers we work with where they’re going to really small compute in very, very dense environments where

 

You know, who cares if 10% are offline doesn’t matter because we’re just taking stuff sort of ubiquitous densification approach. But if you’re trying to do stuff at scale, scale tends to work together at scale, right? There is a lot of efficiency to be gained with large data centers. But I think you’ll find that my my prediction, at least in the industry, is that over time we’re gonna hyper optimize for those use cases. And you’re gonna find AI kind of sitting everywhere doing different things, optimized for where it’s most and what’s most needed at the the lowest price point with the most efficiency.

 

And ultimately the the least impact where yeah, there might be a small racket computer at the bottom of a cell tower that’s handling something for, you know, electric car, you know, autonomous vehicle navigation, for example. But your chat GPT session is going to, you know, the localized inference thing in your city. And then that’s doing its training at the you know, the middle of nowhere, h highly optimized, half off grid in in training center, for example. I could I could see a a world where those things

 

exist, and everything’s hyper connected with really good fiber connectivity, but it’s very diverse and resilient as a result. and it’s optimized based on those use cases. So we’re not building giant data centers in the middle of city because that’s kind of annoying. But we’re building the stuff that the city demands high performance for in the city, right? if you look at electrical distribution, there’s a lot of major cities have huge substations in, you know, the downtown core, but most people don’t see them because they’re either disguised in some fake buildings or they’ve got really good

 

boarding around them, stuff like that to make them kind of hidden. So we’ve done a pretty good job of saying, hey, this is necessary, but we’re gonna make it feel integrated. We’re not gonna make it feel i intrusive or, you know, kind of a an eyesore. And all that stuff exists exists in telecom and hide in power generation, sewage, pumping, all these things are generally hidden from the public quite well because we’ve gotten good at that. when we first started deploying them, there were eyesores, people didn’t like them. and it’s just a bit of forethought and how to do that well to to hide it. So

 

Ed Loveless (01:19:27.864)

I think you’ll see a bit of all of the above.

 

Shawn Cordner (01:19:30.67)

So I have asked people two weeks ago I asked somebody, is this a bubble? And he said, No, it’s a Ponzi scheme. and the reason that he said that is because you’ve got companies like Nvidia that are you know, building chips, selling chips to these companies like Anthropic or OpenAI, and then they’re giving them money.

 

investing in their business, which they’re turning around to use that money, that capital for more chips. I think that you’ve got a a a big challenge in the enterprise space right now where people are not seeing quite the ROI that they expected. And it doesn’t mean that they won’t see ROI ROI, but I we are starting to see the boardroom and CFOs asking

 

Why are we investing so much money in this and what is the benefit that we’re getting from it? And when do we expect to see the return on that? And I wonder if that plus the community pushback and public opinion being so against AI may eventually cut into growth. Now, I’m not saying that it will contract, but if it doesn’t grow as fast and scale the way that they expect that it that it would, then the investment thesis might be at risk here. So

 

do you feel concerned about the foundation that all of this economic activity is sitting on? Because in the US here, you know, half of twenty twenty five GDP was related to AI investment. Our stock market is largely being propped up by the AI companies. So if the underlying investments and capital that’s flowing because of AI

 

is at risk, then really the whole economy is at risk as well.

 

Ed Loveless (01:21:23.52)

I I mean th there’s there’s definitely risk of change. I I will say when the dot com bubble burst, the internet didn’t go away, right? It changed. There yes, there was areas that had lots of infrastructure overbuild. you know, in the fiber space, there’s a lot of consolidation that happened. That’s kind of how I ended up here, right? There’s a lot of stuff that got gobbled up and consolidated, but that stuff that was built out got used. It just didn’t get used for the thing we thought it was gonna get used for, right? so I I don’t think there’s

 

I think there’s little chance of a huge implosion of suddenly the bubble bursts and all this stuff just goes away and we go back to what we were doing before. is there a high chance that who the players are will change drastically? Absolutely. Right. There’s a lot of, you know, there’s a lot of assumptions being made and a lot of speculation and those things are gonna be wrong at some point. but that just means that someone else that’s better at at projecting or b better at managing their balance sheet is gonna walk in after and may they may get some discounts on some stuff and they’re gonna pick up where they left off and run with that.

 

and use that investment, that capital, that infrastructure for what they’re working on next. So I I think the picture may change, the players may change, but I think the the problem remains and I think the the potential remains. so I don’t yeah, say what you will about all the different structures and the way things are done. It’s there’s a lot of creative stuff going on, there’s a lot of speculation for sure. But what’s getting actually built is real infrastructure, real real capability that has a lot of use cases. We’re just all fighting for who gets to use their use case first.

 

so if some of these go away, someone else is gonna step and say, look, this is great infrastructure for my use case. it’s you know, a different type of AI or different type of compute ex intensive stuff. So I think the opportunity and the growth is gonna be net positive. I just don’t know who’s gonna be holding the bag at the end of the day and who’s running the show. May not be who we think it will be. but I think what’s actually being built is, you know, generally a good good positive thing. And it’ll be interesting to see where that ends up. And, you know, it may be a bit of a bumpy ride.

 

through those changes and, you know, as things come and go and stand up and fail and whatnot. But we’re building real things. It’s not all as much as AI feels like it’s just a bunch of software that’s running stuff, there’s a lot of good infrastructure being built. and there’s a lot of good benefits that come out of that, I think, as part of that. So

 

Shawn Cordner (01:23:40.384)

So let’s say that somehow a average citizen found this podcast and is listening, not a technology leader, not somebody that’s in this space and really understands everything that we just talked about there. What would your message be to that person about AI and data centers that maybe come into their neighborhood?

 

Ed Loveless (01:24:02.83)

I think it’s good to try and understand it better. and to understand where it’s going and what it means for you. I think it’s it’s things are scary until they’re understood. I think understanding is the first thing to just try and understand how it works and what it what the net benefits are. I think they’re you know, bringing concerns is it’s valid to bring concerns up because concerns are how we raise issues that haven’t been thought for by others, right? These companies are heavily focused on the problem they’re trying to solve. They’re not thinking about the problems they might be causing.

 

And, you know, raising those concerns in a constructive manner is generally net positive at the end of the day because they bring visibility to things that should be considered and that changes how people operate and act and allows them to do better next time. I g I think I generally believe people are trying to do to do what’s right and to try and do things well. but they’re often operating with huge blind spots because of the the way they’re they’re attacking problems. So I think understanding and and communication is key on both sides to say here’s our concerns and for the other companies to listen and say here’s

 

what we can do to try to address those things. So I I think, you know, healthy engagement and understanding is definitely a good way to kind of become less scared about it. also just understanding what AI can and can’t do and realizing that, you know, it’s a novel technology. There’s a lot of hype. not all the hype is true, but some of it might be. And trying to understand what’s hype versus, you know, scaremonger mongering and what’s, you know, reality is good.

 

You know, if you’ve you just heard about AI in the news and you somehow stumbled on this conversation and you want to learn more, go try some stuff, go understand it, go understand that it’s good at this and it’s actually really terrible at this and you know, give that kind of balanced perspective that, you know, it’s not as scary as it looks, but it can do some pretty interesting things. and yeah, it’s it’s finding that balance. I think it’s there’s a there’s a lot of conversation around AI and you know, education on kids and and capabilities and stuff like that as well that’s really interesting. And AI is just a set of new tools, right?

 

I have a bunch I have a couple of kids and you know, y you’re you’re in elementary school, the the younger grades, you’re learning to do math, they make you do it by hand. And they make you do it by hand because you need to understand how that works. But when you get to a certain grade, they give you a calculator. You don’t need to learn how to, you know, multiply three numbers and three numbers all the time. AI is kind of the same thing. It’s a it’s a powerful technology. We still have to learn the basics of things. There’s a big difference between outsourcing your work to AI and

 

Ed Loveless (01:26:20.426)

Uplifting or upscaling your ability to work better with AI. smart people can do great things with AI, dumb thing not so smart people trying to use AI to appear smart, appear smart, but they they don’t know if they’re right or right or or wrong. And I think AI gives the ability and the impression of intelligence. It’s not always intelligent, but be able to offload some of your work to AI.

 

is a force multiplier, right? Just like having a calculator, it’s a force multiplier to say, like, I don’t have to spend five minutes doing this huge mouth cup. I’m just gonna type it in and get an answer. That is using AI to to upscale and do a lot more stuff. But outsourcing and saying, AI, write this essay for me when you’re in grade six and you have no idea what you’re talking about, that’s not really doing doing your work at all. You’re not learning anything. You’re just making a machine do the work. So finding that balance is key and understanding that there are things that AI can make better for you and things that AI can make harder for you. if you don’t

 

learn those basics are something to think about as well, right? It’s not doesn’t remove people completely. It allows people to work with it’s like a force multiplier, allows you to work a lot faster, right? Just like calculators, they allow you to do math a lot better, a lot faster. Doesn’t remove the need to understand how math works to start to understand the problem in the first place you’re trying to solve. So

 

Shawn Cordner (01:27:33.134)

It’s like jazz. You you need to learn the fundamentals before you can improvise. Right. Great place to leave it. Ed Loveless, thanks for your time and expertise.

 

Ed Loveless (01:27:42.062)

Great, thanks Shawn for having me, appreciate

 

Shawn Cordner (01:27:44.725)

Yeah.