Latest / Two Talking Fools Podcast / AI, Water and Power: What Every Town Should Ask About Data Centers
Transcript
- 0:00Technically, that's a swear word in Finnish, but not in English.
- 0:04I love it. Thank you. Boy.
- 0:09All right. Welcome, everybody. Welcome back to the Two Talking Fools podcast.
- 0:13I am coming to you live today here in the great state of Connecticut.
- 0:18And as many of you know, I have had a want to understand different topics throughout the world.
- 0:26So the first topic I've decided that I kind of want to take on on the Two Talking
- 0:30Fools podcast today is we're going to talk about data centers.
- 0:33A lot of you have heard different things on different news, Fox, CNN, Aaron Brockovich.
- 0:39I decided to get somebody that is a professional, an expert,
- 0:44if you will, in the field.
- 0:46And today with us, we are so grateful.
- 0:48We have Jack Smith. Jack is an infrastructure project manager with more than
- 0:5320 years of experience helping organizations navigate technology change from
- 0:58help desk to system administration
- 1:00to leading data centers, networking, cybersecurity, and digital transformation
- 1:05projects across multiple industries.
- 1:08Passionate about the human side of technology, he is also the host of the IT.
- 1:13Horror Stories with Jack Smith, where he shares real-world stories about project
- 1:17successes, spectacular failures, and the lessons learned when technology doesn't
- 1:23quite behave as planned.
- 1:24Jack Smith, welcome to the Two Talking Fools podcast. Thank you.
- 1:29Thanks for the introduction. I am not worthy.
- 1:35So what are you saying? You're a fool like us, huh?
- 1:39It seems that, yeah, I started installing Windows 95 in the 90s,
- 1:44and well, here I am today.
- 1:47Those were the good old days. Those were the good old days. it's just zero procedures
- 1:54just no paperwork you just fix it and if smoke came out you pushed the wrong button.
- 2:04So, Jack, you and I, we met, we were talking online.
- 2:08I had specifically reached out to somebody to talk about data centers.
- 2:13Here's, you know, I don't know anything about them. I hear tidbits on,
- 2:19you know, Fox, CNN, this one, that one. And I feel like I want to get to the
- 2:24source. Somebody who really understood it, you reached out to me.
- 2:28I have some experience in them indeed. I have opened them. I've closed them. I've moved them.
- 2:34I've actually had two fires in them. So it's always a good conversation to start
- 2:39when I say, let me tell you about my first data center fire.
- 2:42And then when the nugget hits, wait, what, your first one? Yes,
- 2:45indeed, there's been multiple.
- 2:47So yeah, that's my opener at any summer party these days.
- 2:51Okay, let's start there. Out of curiosity, where was this data center? How did it go down?
- 2:56It was a...
- 2:59Power system failure to brief it at a place that i worked back then this is,
- 3:06i think the early 2000s the uh was a,
- 3:11transportation logistics company so they have they have trucks airplanes trains
- 3:16everything and they have a 364 day a year business and the one day new year's
- 3:24eve that they didn't have a business,
- 3:27the power system said i'm taking you all with me and it went up in smoke we didn't have,
- 3:33a total disaster it didn't burn down but we were offline there for a few days
- 3:39so it was them all in all that was a very successful total disaster do they train you for that.
- 3:48Oh no they don't there was you are you are making things up so you do have some
- 3:54things if you're lucky and you have the budget and people actually give you
- 3:57the money you you set the backup systems so if you have say,
- 4:03one big server you make a second copy next to it so if the first one goes poof
- 4:08second one takes over so we could move over to the backup ones we had budgets
- 4:12for That was taken care of.
- 4:14So luckily it was okay. Also, we had a magical thing called backups and we knew
- 4:20how to actually not just make the backups, but how to get the backups back restored
- 4:26onto something to keep it working.
- 4:28That is what you did. And you have millions of small procedures and it's fake it until you make it.
- 4:38You are completely offline. So if it takes five minutes longer, Uh-huh.
- 4:43Literally like it it becomes a magic word description called it's best effort now
- 4:50we're in best effort mode there is we didn't sign any paperwork
- 4:55we will you will hear this a lot in marketing terms we guarantee you five nines
- 4:59up times six nines six nines up times that is how much percentage per year you
- 5:05want the stuff to be available
- 5:07we didn't sign that so it was best effort. I think we did a good effort there.
- 5:14I don't think we can complain.
- 5:16So, let's start out because, as you and I were talking beforehand,
- 5:21there are people, and I don't understand it either, there are people out that
- 5:25don't understand what a data center is. Now, you just said something.
- 5:29You said in the early 2000s.
- 5:31Are you saying that we've had data centers around for long before now?
- 5:37They have existed as long as computers existed.
- 5:40But remind you that the mainframes, like the big fridges, those were not in
- 5:46the homes. They were in specific air-conditioned rooms because these things,
- 5:51they're hot, they're boiling.
- 5:55They're space heaters, literally. You have to cool them down in specialized
- 5:59rooms. So from the 50s, 60s, 70s on, you would have data centers,
- 6:04just a big concrete box with a big AC system in it.
- 6:09And that is where you would host a computer. And putting a computer on your
- 6:14own desk in the office, that is something from the 90s.
- 6:19And computers exist from way before that. But even then, your desktop computer,
- 6:25in the beginning, you would be trading floppy disks.
- 6:29A floppy disk is a USB stick, but bigger and lower storage.
- 6:34Right. you would need to move those around and you would like to you know oh
- 6:39i have worked on this i would like to share it with my colleague next door and
- 6:42then well it's cool to that you put the cable between those computers and you
- 6:46can transfer it over the cable,
- 6:48and then everybody has their copy of the file and you start making mistakes
- 6:52because one edit chapter a one edit chapter b and it comes together and it doesn't
- 6:57work anymore so you put them somewhere central a server a file server a mail
- 7:01server, a database server.
- 7:03And in the beginning, you put them under a desk.
- 7:06Which is not the place where they belong because desks get hot, they get.
- 7:12Coffee, they get dust, they get cleaning ladies that plug out the plug for the vacuum cleaner.
- 7:18I wish I made this up. This is every cliche is true. So you put them in your
- 7:22small air-conditioned room and that becomes your own data center.
- 7:26And you fill it up, you fill it up, you fill it up.
- 7:29You start with one machine, you start with one cabinet, And in the end,
- 7:34you end up with 20, 30, 40, 50 cabinets. So look at one cabinet is one big freezer in size.
- 7:43That goes quickly. Yes.
- 7:47Go ahead. Not to cut you off, but when they talk about the cloud,
- 7:53when I talk about Gmail cloud,
- 7:57you're i always just assumed that it just literally just went up in the but
- 8:03is are you saying that that information goes to a data center somewhere,
- 8:07yes multiple places actually you can you there's public maps that says because
- 8:13they advertise it with google says oh we have data centers this state that state
- 8:17that country this city that city,
- 8:20they advertise it oh like If you're in, say, New York, and Google says,
- 8:26well, we store your data in Los Angeles, it has to go across the U.S.
- 8:30Every single time you ask an email or something else. That just takes too long. That takes time.
- 8:35So they say, oh, no, no, we have a local in New York, and you can even pick
- 8:39the Bronx or Manhattan. We have both. Pick one.
- 8:42And you can actually, if you are a big customer, so not your,
- 8:47free win-of-the-mill Gmail, but when you're a corporate customer,
- 8:50you can say, I want my data, my email to be in that specific data center,
- 8:56and I want my backup copy to be over there.
- 8:59I want my backup copy to be in New Zealand. You can perfectly do that.
- 9:03Okay. Tell us about the size and the magnitude of these data centers.
- 9:10So, like we were talking, a lot of people right now are talking about the water.
- 9:15What, in general, how much energy do these take, and where do you see them in the world right now?
- 9:25Well, the answer is yes. Yes, they take energy.
- 9:28How much nobody will publicly state how much power they use but you are counting in,
- 9:37the megawatts per data center easy depending on what they do but that is an
- 9:41average you can you can balance this off you know if you have solar on on the
- 9:47roof you can you can during the daytime you can use solar.
- 9:51I have actually no data centers that put a windmill nearby that they would they
- 9:56they like in an industrial estate
- 9:58the industrial estate will put two three windmills in to share power with all
- 10:02the companies that are there including the data center and then maybe there's
- 10:06a factory or some small business they will all plug into that and they make
- 10:09their own little power company,
- 10:11it's there it's it's not that it all comes from say the national grid you have
- 10:16alternatives that you can get your power but yes indeed it all uses power all
- 10:20those machines need to have power all those discs
- 10:24need to turn all those acs need to be cooled the elevators need to run the cameras
- 10:30that secure the place need to have power it is,
- 10:33quite power hungry although,
- 10:36one data center might look a lot and if you don't scale it is globally but if
- 10:41you start looking at how much power say any industrial factory uses,
- 10:47it tends to level out actually, although you always have the exceptions left and right,
- 10:54that happened. So?
- 10:57Yes, it's there, but it is not a huge problem.
- 11:03Okay. Could you compare it to any other type of mill that we would see,
- 11:08like, here in the United States? Like, does it take as much energy as a,
- 11:11as a coal mill or, you know, I'm not well versed in this, but.
- 11:15Yeah, me neither. Let's see.
- 11:17I'm looking at having worked in the, well, I've worked in both mining and breweries,
- 11:26not together, but it's, it's like the two edges of everything.
- 11:30But say that an average brewery will use more power than a data center because
- 11:37at a brewery you're boiling water,
- 11:41you are moving bottles across the production line, you are using big industrial
- 11:45machinery, you will need to power your forklifts if they're electric etc.
- 11:50So yeah that could be a comparison but it all depends on size.
- 11:54You're small breweries, big breweries, Small data centers, big data centers,
- 11:58and then how they powered.
- 11:59Preferably, it's a mix of solar, wind, classical.
- 12:05You have geothermal, if you have that accessible, whatever can give you some
- 12:11heat and that makes you power. Okay.
- 12:15So, when they're looking, when a data center is going to open,
- 12:19what is their ideal conditions? What is the environment they're looking for or gravitate towards?
- 12:27It depends on where you are, but to keep it easy, I'm going to go for the U.S.
- 12:33Because that's where you have lots of space.
- 12:36You need to put in a big, ugly concrete box in the middle of a field.
- 12:44If you're living next to neighbors, neighbors are going to complain.
- 12:47So you need space. You need to put it somewhere. That somewhere,
- 12:51of course, needs its power, as we discussed.
- 12:55So you need to be able to pull in a power line. Of course, you can put your
- 13:00own generators next door. That's not a good plan. Everyone do it.
- 13:05And then also you need to be able to hook it up because it is all cloud and cloud needs internet.
- 13:12So you need to be on a place where you can provide any data connection,
- 13:19any internet connection there.
- 13:21And the bigger the data center, the more data will go in and out.
- 13:25No, it's like a highway parking. If you have slots for 500 trucks,
- 13:30you're going to have to have the on and off ramp to get the 500 trucks in and out easy.
- 13:34Sure. Principle. So you need the data connection to actually provide all the
- 13:41stuff that goes in and out. You can...
- 13:44So you will probably see them being put up on a place where there is data connection.
- 13:51That's fiber optics. Fiber optics. So it's fiber in the local terminology.
- 13:56And ideally, it's something called dark fiber. Dark fiber is a cable that was
- 14:01put in years ago and has never been used.
- 14:04And you can now basically rent it or you can buy it for a while.
- 14:08And maybe you rent 20 miles. Maybe you rent 20 feet. Maybe you rent 2,000 miles.
- 14:16It's up to you. That's a completely separate business that nobody in the consumer world has ever heard of.
- 14:23So, dark fiber. Yes. Why?
- 14:28Okay, you're saying that it's in the United States right now. Oh, no, no, no.
- 14:32Dark fiber is globally, and dark fiber is fiber optic that's down in the streets,
- 14:38down in the fields, that is not used. There is no light going through it, therefore it's dark.
- 14:43Okay. And why haven't we ever explored this?
- 14:48Well, it is being explored, but people say, oh, I estimate that there is going
- 14:54to be a lot of data traffic between New York and Boston.
- 14:59Okay. So you put in cables and cables and cables and cables and cables,
- 15:05and they're all just waiting for a customer.
- 15:07That customer can show up on day one.
- 15:10It can show up after three, four years. You got to see that. How do I compare this?
- 15:14A fiber optic cable is not just one it's not like one power cable it's not one strand so,
- 15:21a fiber optic cable is actually a bundle of 16 32 64 your powers of two go up
- 15:27really quickly and you can just
- 15:29rent one of those 32 cables in there well pairs one up one in and one out,
- 15:36and then your dark fiber becomes active fiber if you start using it so it is see that you're.
- 15:44You're selling nuts and bolts in your shop and you're making a large stockpile
- 15:49of some type of nuts and bolts that you think will sell tremendously within
- 15:54the next five years and you stock up,
- 15:56and then you can sell from your stock instead of having to,
- 15:59buy new stock or dig a new cable because you have to dig up streets and highways
- 16:05and that's expensive. Got it.
- 16:08It's dark fiber. Okay. How interesting. Damn.
- 16:15All right. How many, what does a data center tell people? What does it look
- 16:20like inside a data center? I have an idea in my head. I want to know if AI is
- 16:24fooling me or is this a truth?
- 16:25Is it just like rows of computers just, or machines?
- 16:30In the actual data room, yes, but that's after three, four security doors. Okay.
- 16:38When you come in, it's like a regular business lounge.
- 16:42There's a guard, there's a desk, there's some chairs, usually when it's nice.
- 16:47I say, hey, I am here for customer Y, because quite often these are shared.
- 16:52You will have one data center, can have multiple rooms, and rooms can have different
- 16:58tenants. They can have different owners. They can have different people that use them.
- 17:02And some can then let them sub-rent them to others. It goes down.
- 17:08You can actually rent a data center, a room, a cage, an aisle, a rack, or part of a rack.
- 17:15So a room is a room. A cage is a size of a room that they put literally metal
- 17:21grating around it with your private door. And that's it.
- 17:24You can have an aisle which is what the ai tells you you're along a hallway
- 17:28of server cabinets you can have a rack which is like one fridge door and you
- 17:33can have part of a rack that they will then secure off,
- 17:35so okay so you can put in one machine you can put in 1000 machine they will
- 17:40adapt it for you but that's,
- 17:42only when you're in there so when you come into the data center you say hi i
- 17:47am for customer a this is me you will have to pre-notify that you're coming
- 17:52so that They can prepare, yada, yada.
- 17:55They will let you in. You go through some security door. They will give you an access badge.
- 18:00Usually, it's one of those turnstiles.
- 18:04If you're SOL, they are tiny, and they actually weigh you so that you don't
- 18:11weigh more when you leave than when you enter, so that you haven't taken anything with you. Really?
- 18:17Oh, yeah. Yeah, it's really funny when you're over 150 pounds and you get stuck
- 18:22in those for half an hour.
- 18:24That was a thing.
- 18:26And then you have access to your room, to your cabinets and whatever.
- 18:32And then indeed you will have the AI pictures where you have the rows and rows and rows of machines.
- 18:39But quite often when it's done properly, it's more shielded off for heat management.
- 18:44So you will have a warm aisle and a cold aisle so that in one the AC blows in
- 18:51and in the other one it sucks the hot air out. so that you have else it mixes
- 18:55up and it's less efficient and it just costs more.
- 18:59You can drive that very far. But in general, your AI image of rows of computers
- 19:04and cabinets is more or less correct.
- 19:07How many employees for a whole data center is there?
- 19:11Actually, guys running the place. That's just the security team.
- 19:15That is three, four guys in a fishbowl. That's it.
- 19:19And I asked that because that was one of the things as I'm learning more about
- 19:23this that people say, these data centers, they take up all this space,
- 19:28but they only render a few jobs. Correct.
- 19:32They do render a job during construction and during installation when everybody
- 19:36is there to set things up and pull the cables in and pull the power in and move all the systems in.
- 19:41But once that's up and running, you can basically close the door and throw away
- 19:46the key and nobody needs to be in there.
- 19:49I've run around in data centers for weeks and weeks during you're doing installations.
- 19:55I haven't seen anybody. I was all alone in a three, four, five hundred square
- 20:02meter, sorry, European sizes, the room here on my own. I haven't seen anybody.
- 20:09I'm just fascinated. Now...
- 20:12So we've had it for a while because they hold, you know, and save information for us.
- 20:20But now there's all this hubbub about them. I have to assume it's because of
- 20:24AI and how much that, how we're going to need these, I would assume,
- 20:29more data centers because of AI. Do I have that right?
- 20:33Initially, yes. So currently, well, everybody is jumping onto AI and AI needs
- 20:38computers to run on and you need to put them somewhere.
- 20:41So you put them in a big gray box out in the field that's what you do that's your ai data center now,
- 20:48everybody is my personal opinion overdoing it a tiny bit so everybody is now,
- 20:55installing data centers everywhere and i am afraid that we're going to end up
- 20:59with a lot of white elephants so basically,
- 21:02unused because there is something that's called moore's law that for me still applies,
- 21:11explain that so Moore's law is Moore was one of the founders of Intel the chip
- 21:16company and he found out and it's quite often misexplained as every few years.
- 21:25The power of computer doubles so they will do double the same work as say five
- 21:32years ago that's often a misquil but that's how it's explained more correct is that,
- 21:39you can put twice a mount, the same of computing power, so being able to calculate your AI chips,
- 21:46in the same spot for the same money.
- 21:49So you can double your horsepower and,
- 21:53for the same space that you need. Meaning that currently we are first,
- 21:58second generation of AI chips.
- 22:01Our people at, our friends at NVIDIA have made very good market from it.
- 22:07They're making lots of money. I wish them all the best.
- 22:10However, any second generation of AI chips will be smaller, cooler, more efficient.
- 22:18And instead of, say, 4,000 AI chips per cabinet, suddenly you can put in 8,000 AI chips per cabinet.
- 22:27So you will only need half the space.
- 22:30And in the end, you will need a few data centers.
- 22:34We've seen this in the 90s and the 2000s when also everybody was putting in
- 22:40data centers because everybody was doing websites and the .com and you were
- 22:44putting databases on there and everywhere.
- 22:48And back then, computers were huge. They were big.
- 22:53They were literally fridges.
- 23:00I've put database servers into these racks, which were two feet high, literally.
- 23:06One machine, and then we have two. So one backup, one actually in use.
- 23:12And then two years later, the same computing power, the same storage was a few inches high.
- 23:20And that's just going to repeat, not tomorrow, not next week, but in the future.
- 23:26Mind you, a data center needs a few years to build and get operational.
- 23:30I'm afraid that we'll have a lot of empty data centers soon.
- 23:34This is my personal opinion. I can be very wrong. I am in this one,
- 23:38not a specialist, but I would not invest in building too many data centers today,
- 23:44because the computing power, the AI chips, they are going to get smaller,
- 23:48faster, cheaper, instead of,
- 23:51worldwide 500 data centers, you will only need 250 in the end.
- 23:56Right. Well, I mean, I was a child, a product of the 80s. So I remember having
- 24:01the disks, but then it just got smaller and smaller and smaller and smaller.
- 24:06It's almost like when the Olympics, it's all the white elephants we have just
- 24:10from the Olympics, every different location. Now we have all these coliseums
- 24:13all over the place not being used. And that's a problem. That's a problem.
- 24:18But it's funny because this is the first time that I'm hearing your opinion
- 24:22on it and that makes a lot of sense to me and that should be a concern.
- 24:26I want to talk a little bit about the big, what everybody's correlating with
- 24:32data centers, which is water usage.
- 24:34And now, but what else is scary about them besides this water usage,
- 24:40which I want you to clarify. Well, first of all, it depends where you're getting the water from.
- 24:48Are you pumping it out of the ground?
- 24:54That's water for your cities For your farmers For your ground level water That's not coming back,
- 25:04It's raining fewer and fewer The water table is lowering And if you use that,
- 25:11It's one time use You use it and you throw it away,
- 25:15That's problematic in mid-term to long-term If you put it by a river, like a power plant,
- 25:23you take the water from the river, you use it to cool your systems. And of course...
- 25:29You will then dump the water again into the river. It'll be a bit warmer, but it can take it.
- 25:36So it depends where you get the water from. Also, how you use your cooling.
- 25:41If you have a dry atmosphere, the desert.
- 25:47You spray water to cool down. Now, if you go into your garden and you live in
- 25:54a flyover state, nothing against flyover states,
- 25:57you spray the garden hose you have a mist of water, you walk through it cools
- 26:02you off, exactly that principle,
- 26:04if you're in say New England where it's very moist in the air you can't do that
- 26:10because that's not how it works you need dry air and then you can cool systems
- 26:16down and if you then use groundwater,
- 26:20one time use water for that yes that is not ideal you would need to plan to
- 26:26put your data center by a river, of course, by a river, which is where the people
- 26:30live, and the people don't want a gray box in their backyard.
- 26:33Wow. And that, you know what, now that you say that, I believe that's what Aaron
- 26:37Brockovich is now talking about, is one town,
- 26:40there might be more, but one town in particular, people had a loss of water
- 26:45pressure in their homes, and come to find out, this one particular data center
- 26:51was hooked directly into the town's water,
- 26:55source, and they had no idea they were even paying for it.
- 27:00Yeah, it is.
- 27:02If local government, local politicians basically bend over to have the data center there,
- 27:09Because as you said, they promise all the jobs, which in the end don't come.
- 27:13Right. They was, oh, we put in a data center real, we'll put in 300 jobs.
- 27:18Well, during construction, sure. But once it's there, you know,
- 27:20it goes from 300 to maybe 30. And that's it.
- 27:23As in, if you go to a town with a data center, ask around on Main Street,
- 27:28do you know anybody that works there? No, nobody does.
- 27:33So there it is. It is, you know, where does all the things come from?
- 27:37Again then cooling an alternative to water is,
- 27:43using classic refrigeration chemicals which then of course you need power to
- 27:48refrigerate them but that power that that can come from solar can come from
- 27:52wind can come from nuclear wherever you have a power mix that you that you get that thing from,
- 27:58that is to decide wherever you you get it from but yeah so,
- 28:04cooling is a thing again for now because new generation chips new generation
- 28:10technology will not run as hot as the current one it will run cooler and therefore
- 28:16will use fewer cooling and this is just,
- 28:19a continuation of morse law now morse law initially was for chips and cpus but
- 28:25in my brain it scales up to computer numbers and the numbers it's,
- 28:31it's still valid for that for me,
- 28:34The future can prove me wrong, and then I will accept it. But for now, that's how I see it.
- 28:40So in your mind, there's not a fear that AI is requiring so much,
- 28:46so many data centers that it's going to overtake where there's going to be farmland
- 28:50in Maine at some point and,
- 28:52that where they figure out a way to make even New England work because you need
- 28:56so many of these around. You don't foresee that.
- 29:00Sadly, the pendulum always swings too far in the beginning.
- 29:05So it swings through the extremities and has to come back. But I'm afraid we're
- 29:10not at the extremity yet.
- 29:12So it's going to get worse before it's going to get better. And for that,
- 29:16you really need to hope that all the new AI tech will become available soon.
- 29:24I know that there is a company in the Netherlands called ASML.
- 29:29ASML makes machines that make computers.
- 29:35Chips and it is they are working on new technology that will exactly allow denser
- 29:43computer chips lower power usage,
- 29:46lower smaller size everything is being worked on and then of course this is
- 29:52what we know about there is this,
- 29:56big country called china where we have no idea what they're doing,
- 30:00but guaranteed they're working on the same.
- 30:04What's the scariest part about, or not the scariest part, but what do,
- 30:08what should we all be concerned about when it comes to exactly what we're talking
- 30:13about right now? China and AI data centers. Let's touch on that.
- 30:18It's, if it's, if, if it's one thing, it's AI has many problems of its own for the record.
- 30:26You need to, you need to learn the things. AI is stupid.
- 30:30It doesn't know an apple from a carrot unless you tell it the difference between.
- 30:35And where that...
- 30:38Source material comes from that ai uses that is very very questionable that's
- 30:44an entirely topic on its own i am no specialist in that so i'm not going there
- 30:48but that is that is one thing
- 30:49secondly is what's it's being used for there is one,
- 30:54local say popular scientist slash comedian just like you know ai is,
- 31:02a very very good intern,
- 31:05you give it the mission it does a lot of work in a,
- 31:09short amount of time, but you need to double check everything it does.
- 31:15And that last step currently is not really happening.
- 31:19For the podcast, actually, I've looked at, I saw the ads come by on Facebook
- 31:26says, we guarantee you number one Google hits in three months. Like, I am intrigued.
- 31:33So I said, like, how do you do this?
- 31:35And oh yeah, we use AI to write you an article every day and then we use the
- 31:40correct buzzwords that will get you picked up in chat gpt and on google so that
- 31:44you are number one so when when somebody enters,
- 31:48the search searches for chickens then you will be the number one in chickens
- 31:52in every google hit or chat gpt hit that you do and for this we will do everything
- 31:57ourselves and here's our team and i looked at that,
- 32:00at that team from for this company and there's like 25 30 people that they're
- 32:05okay small business fine but then I found out there's not a single developer
- 32:09not a single programmer there nobody is writing actual programs oh no, we use AI,
- 32:15yeah, but you're not making it you're just buying something off the shelf you
- 32:19don't know what's in there and off you go,
- 32:22that's the current danger right now people do not know what they're doing with
- 32:26it it's all going untested out in the market and.
- 32:31I'm I'm afraid we're really, really running against a very, very hard brick wall.
- 32:38I want to touch on what you just said there because I come, I'm a web designer by trade.
- 32:43So, you know, the whole idea is when you're building an informational website
- 32:48is that you are an expert in your field.
- 32:51So you want to make sure that when somebody types in the short or long tail
- 32:55keyword, that your website comes up as the professional. Fair.
- 33:00But here's the problem with that. So now we used to have, you know,
- 33:04content writers. We used to pay people, good writers, people that honed in their
- 33:08craft, now everybody's spitting everything into chat and GPT. The problem is, is that.
- 33:16Where the hell is chat? Is chat getting all of their information from, like Google does?
- 33:23Like they're just putting spiders out and it's coming back and they're figuring
- 33:27out who they feel is the best professional? Because that's dangerous.
- 33:31Because you can, you know, you'll be familiar with the robots.txt file if you're a web. Yeah, yes.
- 33:40Hey, dear Google, Bing web crawler, do this or don't do this.
- 33:45But that's not mandatory and you can ignore it. And many AI crawlers just ignore
- 33:51it and just go through everything and they just take it.
- 33:55Every single public picture that you put out on your Facebook,
- 34:00on your Instagram, unprotected, it is being scraped somewhere.
- 34:04It is being used to build,
- 34:07an image of something wow that that currently is what i see the danger is is
- 34:13is where the info comes from and what and the uncontrolled way we are currently,
- 34:21doing that there was i was on a bit of a bit of a side tangent here on on ai
- 34:25i'm gonna come back to data center soon but for just for the fun of it me and
- 34:30a buddy said like let's do something in ai and we said okay we're,
- 34:35gonna make a website that will compare the average water usage,
- 34:42in Western Europe and then we.
- 34:47Mine that data and then we sell the data set never planning that this would
- 34:52be something but hey he had some tokens left,
- 34:57in his cloud code system and whatever just toppy toppy toppy the thing said
- 35:02yes sir and after 15 minutes the entire system came out. We had no idea how it worked.
- 35:08But it did what we asked it to do.
- 35:11Was it correct? Well, we did some spot checks. It was mostly correct.
- 35:15Was it because it's, oh, yeah, yeah, I have. When an AI starts talking to you
- 35:20in the I-form situation,
- 35:23it tickles my back,
- 35:28oh i have already made a connection with stripe so you can already accept payments,
- 35:33how who how how whose account is it going who's getting a cut like you just did it like,
- 35:42can we pull the handbrake it's you know well,
- 35:47i see it all the time ai is taking chat gpt free version everybody's doing it
- 35:52every now every event flyer is at every so much is ai you're better off just
- 35:59posting regular photos it's ridiculous it's getting,
- 36:03it's shitty i'm so tired of it.
- 36:09Guilty as charged myself if i need to if i need to like i've you know
- 36:15i live in a place where people speak four languages like i get an i get an email
- 36:21in french and like zips and i just copy paste into into chat gps i've gotten this email from,
- 36:29this guy please please tell him friendly to get lost and he just types the email
- 36:33in perfectly friendly first to copy paste and go,
- 36:37guilty as charged absolutely,
- 36:39You said it best. I love exactly what you said. AI can be your best personal assistant.
- 36:45Oh, my God. It's so true. It can literally take, well, you know, like our podcast.
- 36:51If I need to quickly write a blog about what's been going on and I don't have
- 36:55any, I'll give you the crux of it because you don't know that.
- 36:58But then she makes it beautiful.
- 37:01I'm referencing her as a girl. It's just amazing what it can do. So guilty.
- 37:07Guilty. But then all that stuff, all those examples, all that knowledge,
- 37:11that needs to be stored somewhere.
- 37:15It is storing. Every single text ever written, like think Wikipedia times Wikipedia,
- 37:24stored somewhere on a spinning disk that this system then grabs the information from.
- 37:32Like if you say, chat GPT, hey, make me an event poster with three balloons,
- 37:39and a clown driving on a unicycle,
- 37:43it will have stored 5 million images of a unicycle somewhere to come to an average.
- 37:48And those 5 million unicycle pictures have to be stored somewhere.
- 37:53And that's our full circle back to the data center. You have to store this somewhere.
- 38:00You have to keep that data, all those examples, you have to keep it somewhere physically.
- 38:06And of course, that's one AI, but there's a multitude of them,
- 38:09because of course they're not sharing,
- 38:12because that's their commercial advantage, because one says,
- 38:16oh, I have the best collection of unicycles, therefore I am the king of unicycles.
- 38:22So you're saying like ChatGPT is not going to share their information with another
- 38:27AI program? Absolutely not.
- 38:29How many, who are the top three programs right now? You always hear about Chat. Who else?
- 38:37Oh, people that develop use Cloud, say AOD, Cloud, whatever in language when I say it.
- 38:45That's used a lot. Otherwise, I use ChatGPT because that's what I'm used to.
- 38:52I don't know the others I know that there's a lot of things going on About our
- 38:58friend Mr. Musk's Grok system Okay I don't know anyone personally That uses it,
- 39:06Or maybe they do when they just don't admit it. Because, you know, as I said.
- 39:14Everybody uses AI. People that say that they don't are lying. That's basically it.
- 39:21Sorry. It is like, maybe not for your final product, but if I,
- 39:27for the podcast, have to make a t-shirt, I feel like, hey, give me five t-shirt ideas. Right.
- 39:34Can I please pick two? And then I will make those designs myself.
- 39:38But like, you know, hey, we are now doing an episode on a data center fire.
- 39:43Make me three t-shirt designs that are funny, witty, and a bit tongue-in-cheek, and it just comes out.
- 39:49But all those examples of every t-shirt ever made, every t-shirt on Amazon that
- 39:55handles about fire and IT,
- 39:59and sarcasm, it has to be stored somewhere. 100%. And that all goes into those
- 40:05big, great concrete boxes.
- 40:06And there are actually multiple lawsuits where one AI company is suing the other
- 40:12because they have stolen their data and they have crawled their entire database.
- 40:16Those lawsuits are out there. Wow.
- 40:21Wow. That is so... Wow. Okay. So, Jack, let me ask you this.
- 40:30If you lived here in the United States and you found out that a data center
- 40:33was coming very close to you, you've heard the questions that I've kind of told you about.
- 40:39And I know you know all that, but the questions I hear people asking and raising
- 40:44curiosity about, what are people not asking? What are the questions that people
- 40:49should be asking about data centers that they're not? Is there anything?
- 40:55Who is your customer? Well, let's just say for, like, let's just say...
- 40:59Sorry, that would be my question. I see, I see, I see. Who are you going to...
- 41:03Like, who is building it?
- 41:05Because it will just be new data center, LLC, a shell company. They're smart enough.
- 41:10But, like, who's going to come in there?
- 41:14Is it going to be just for AI? Or is it going to be for, say,
- 41:20the Coca-Colas, the Fords, the GMs,
- 41:24the Walmarts will they also take space in there if it's only AI you risk of
- 41:32getting a very big empty box within the next decade,
- 41:37Because it will be no longer of any use because technology will have caught
- 41:42up and they are no longer needed or they are not economically viable anymore because,
- 41:52the cost has gone down. They're not making much more money.
- 41:57By the way, not a single AI company has made a single dollar ever.
- 42:01That's a completely different topic. Really? Zero. Like, it costs,
- 42:06like, no, because data centers cost billions and billions and billions and billions and billions.
- 42:12And do you really think that your $20 a month chat GPT subscription is going to help? No.
- 42:19So then what's the bigger picture, I guess, is the question.
- 42:22Well, currently, it's, well, the bigger picture of what? The data centers?
- 42:28Well, I guess if they're not, they're getting an empty one.
- 42:33But no, I guess is the value what people put into ChatGPT every day?
- 42:40Is that what the nugget is? Are they selling information?
- 42:47Everything that you put in there becomes property of ChatGPT and they can do
- 42:51with it whatever you want.
- 42:54How will it be beneficial or make them money in the future if they're only charging
- 42:59us 20 and it's costing? I'm not really sure they know.
- 43:03I'm not it is it is it's so for,
- 43:06people with great beards you had something called the dot-com boom in the late
- 43:1190s where people say oh we'll do something and name it dot-com and people oh
- 43:17yeah dot-com but we'll put money in it,
- 43:19and whatever and people would at those companies you would hey what's your business plan,
- 43:26the business plan would say oh we get investment money, angel money, investment money,
- 43:33angel money, we do a tour, we go public with an IPO, and then we'll find out.
- 43:41That's what they're doing out there. They're just gathering information to become
- 43:46the biggest one and hopefully the only one left.
- 43:50And then... Wouldn't that be a monopoly? And then put a value and monetize the data that they have.
- 43:57But they're not there anytime soon.
- 44:00Also, statistics differ, but the amount of investment in AI and data centers
- 44:07today is thousandfold of what the dot-com bubble was.
- 44:13Huh. It's hard for me to even wrap my head around it. It totally makes sense
- 44:17when you put it that way. The numbers are mind-boggling.
- 44:21But when you put it that way, it's like, wait a minute, what is the bigger picture here?
- 44:27What is going to be, what's coming down the road, like that dot-com era that
- 44:30was unbelievable? I was around for it.
- 44:32I loved it. It was awesome. That was great. But what is going to happen with this?
- 44:37We're all just we are dumping the most private information into this and we're
- 44:43falling in love with it i know people feel like chat is their friend,
- 44:48i know multiple colleagues that
- 44:52are actually having day-to-day discussions with chat gpt on their phone,
- 44:56and and then they will they will come so when i'm in a meeting with a customer
- 45:02and i say like hey this and this and that and that The customer really replies
- 45:06to me, yeah, but ChatGPT tells me that.
- 45:12How yeah and then i'm wasting so much time debunking super google because that's what it is like no,
- 45:23right and then right and then the question is like but again where is chat pulling
- 45:28its information from everywhere every every single thing ever posted every every
- 45:34public newsletter, texts,
- 45:38everything will be taken in.
- 45:41Companies just sell. There's a good example on that. Pokemon Go.
- 45:48Yeah. So,
- 45:51Where is the value of Pokemon Go?
- 45:56The ads? I don't know. No, if only. No.
- 46:01The company, so Pokemon Go, the company, Myantic was sold to some investment fund.
- 46:08And now what happens is, you know, you have the Pokestops.
- 46:12Pokestops have pictures with a precise geolocation. This fountain is there in this location.
- 46:19And that information has been sold, like,
- 46:23that picture database, that map database has been sold to multiple companies,
- 46:28to make an even better, say, Google Street View, Apple Street View, and map makers.
- 46:34Because 5,000 people have confirmed that this spot is there.
- 46:39This puts a value on the data and on that picture. So this storefront, this fountain, this...
- 46:46It's accurate. whatever is accurate this is correct accurate data and and and
- 46:52they they've made multiples that what the of of your you know,
- 46:58a five bucks for five thousand pokemons to get more pokeballs nobody cares about
- 47:03that it's all in the data that that you put in there it's it's it it always comes back to,
- 47:10if the product is free you're the product what you're doing is being sold.
- 47:19I never thought about that that is unbelievable you're welcome holy shit yes but it kept us,
- 47:29it really makes you that I'm just thinking I'm like thinking of all the free
- 47:33things that I do what the hell are they taking from me.
- 47:39Who's the, who, are they like, is there, is there ones that are like more corrupt than the other or more?
- 47:44Nah, they're all, they're all. They're all bad. They're all like,
- 47:47you know, on, on, on level from one to Mr. Burns, they're all in Springfield.
- 47:53Every single one of them.
- 47:54Oh. Like, you know, on a scale from one to yes. Yes. Yes.
- 48:01Does it scare you as somebody that's kind of, you're not kind of in this realm?
- 48:05Well, scare you? No, because, well, being a bit in the industry,
- 48:10I am careful with what I put out there and what I upload, because I know that
- 48:14whatever I put on there is no longer, you know, my property.
- 48:18It is that one thing that I know you have to get back to data centers,
- 48:24but people that publicly upload pictures of their children on Facebook,
- 48:28on Instagram, stop doing that.
- 48:32Really? Because those pictures, if you look at the license agreement,
- 48:39they can do with it whatever they want.
- 48:41They can sell it. They can use it to train their AI models. They can do everything with it.
- 48:46You are literally signing away. Like the prom pictures, the school pictures,
- 48:52the baby pictures. Yeah, you know.
- 48:57They don't even pretend that they're giving you privacy. No, you have zero privacy.
- 49:02Do those photos end up on the dark web not directly,
- 49:10I'm one of those idiots that has never been on the dark web I have not either
- 49:14I don't care, I do not care about it I can find all my strange fetishes on the normal internet,
- 49:21you know,
- 49:23I'm okay with that. Oh my god.
- 49:29That's great. No, no. So I honestly do not know.
- 49:34But yes, there are indeed people that will scrape pictures of teenagers and
- 49:41put them on private repositories,
- 49:45dark web or just interweb, and then go for that. Yes, absolutely.
- 49:52It is it is when you and i now we're going extremely off topic here but it's
- 49:57all good it's all the same but do not put,
- 50:00pictures of your 10 year old daughter in a swimsuit on the internet do not upload
- 50:05that on instagram because,
- 50:08people will crank one off on it on it somewhere literally don't do that,
- 50:13well we know it we know that the world's evil the world's got evil in it so
- 50:16what makes you think they're just going to browse over that photo and not the other ones,
- 50:20swipe swipe swipe oh nice one um.
- 50:26No it's it's it's,
- 50:30i actually want to go on a further i'm going to ask you one quite further tangent make my okay,
- 50:36my name is kate i am going to be 42 years old in a couple weeks here And I do
- 50:43not understand when people say mining Bitcoin,
- 50:49Yes What does that mean?
- 50:52Because now I just know that you're going to be able to What does that mean?
- 50:57Tell me, Jack Please help me Okay, so you know what Bitcoin is?
- 51:02I know it's a coin, but it's not a physical coin It's a virtual coin It's a
- 51:07virtual currency Okay And each Bitcoin is technically, it's a very long, very large number.
- 51:18And because of all mathematical formula of which we are missing multiple PhDs
- 51:24to comprehend, the number of Bitcoins that can be created and are valid is limited.
- 51:32That creates scarcity. It's supply and demand. Who created the Bitcoin?
- 51:38Well, that is a million dollar question.
- 51:40Nobody knows. Nobody knows. There are some journalists that claim they have
- 51:45found a person that created Bitcoin.
- 51:47Other people say that this is not the person, but that is the million Bitcoin
- 51:52question, who created Bitcoin?
- 51:57So where do you go if you want to see how many Bitcoins are available?
- 52:03There is a trace of every Bitcoin ever found.
- 52:09So you basically come back to the very long number.
- 52:13You have to take a number and ask, is this a valid Bitcoin number?
- 52:17No next one no next one no next one in the beginning the numbers were short
- 52:23it was like really tack tack tack,
- 52:25you would have one one bitcoin a day one bitcoin a week one bitcoin a month
- 52:31now is one bitcoin a year that you are mining because you are just checking
- 52:36numbers is this a valid bitcoin that,
- 52:39corresponds to the mathematical formula that defines this a bitcoin That's it.
- 52:44So when they have found one now with Bitcoins being $10,000, it's party time.
- 52:53So that's what mining for Bitcoin is.
- 52:56And then there is a theoretical limit on how many Bitcoins can be made with
- 53:01the mathematical model.
- 53:03And there is an estimate. I don't know.
- 53:07I am not a crypto bro. I don't ever want to be one.
- 53:11There is a guess on how many Bitcoins there are, and you can find this through,
- 53:16a public history called a blockchain, which is like a public database that says,
- 53:21oh, this Bitcoin has been there and there and there and there and there, etc.
- 53:25But then there used to be a period where everybody wanted to use blockchain
- 53:30in a professional world because it lists where information has been.
- 53:35Nobody has found any use for it and that quickly died out, tanked random deity.
- 53:44But like I'm not gonna go I'm not mining for this Bitcoin right,
- 53:49somebody else is, right yes, so that's also,
- 53:55they found out that a good way of mining Bitcoin is graphics cards,
- 54:00which is why suddenly all the big expensive graphics cards rent on Obtanium
- 54:05because everybody was buying them up to mine Bitcoins.
- 54:09And then you would have Bitcoin farms where you would have a row of PCs with
- 54:15graphics cards and they would all check these numbers to see if they found a
- 54:19valid Bitcoin according to the mathematical formula.
- 54:23And then that went away because the number of Bitcoins that they found went down.
- 54:28The power usage went up and it didn't balance out anymore because they were
- 54:33spending $10,000 per month on their power bill and they were only mining.
- 54:39$8,000 worth of Bitcoin, and then they will stop.
- 54:43Holy shit. So we're finding less and less Bitcoin?
- 54:49Yes, because we are getting close to the theoretical maximum level of Bitcoins that can be created.
- 54:54And therefore, it takes longer to find a valid number. So look at it that you're in a gold mine.
- 55:02And when you start digging in the mine at the entrance, you find nuggets all
- 55:06the time. Here's a nugget. There's a nugget. And then it's, oh, this is easy.
- 55:11Nuggets for everybody. Make it rain. Let's be easy. And now that the mine is
- 55:16slowly getting deeper and deeper, you have to dig more left and right and up
- 55:21and down. And maybe every now and then you still find a nugget.
- 55:24But the rate at which you are finding nuggets is going down spectacularly.
- 55:29And it's not like the average Joe is ever going to find one.
- 55:32Like, what's the likelihood of me mining and finding a Bitcoin?
- 55:37You might want to start playing the lottery.
- 55:41So get a bunch of scratch cards yeah,
- 55:45don't don't make this one of my get rich it is technically i can i can take
- 55:51my spare laptop put a bitcoin miner up there,
- 55:54and if it's by pure luck i have a magic number and it says you yes i have found
- 55:59a valid bitcoin and here now is something,
- 56:04a row of numbers which somebody will pay ten thousands of dollars for Well, you've won the lottery.
- 56:10Okay. Okay.
- 56:15That one there, I think, is probably going to be another podcast.
- 56:21Yeah, my number.
- 56:25I think what we'll do is people will sympathize with me for a few weeks after
- 56:28listening to this. They'll send in, and they'll be like, okay,
- 56:32we get it. We don't understand it either. I'll be like, okay,
- 56:34we'll have Jack back on. We'll start from the beginning.
- 56:37I don't know, Jack. I mean, as far as the data centers, I think we really touched
- 56:44on it. I think after listening to this, it makes a lot more sense.
- 56:48And as I said, all the stuff that we've just talked about in the last 30 minutes
- 56:53and went off topic, like all this data needs to be somewhere. Right.
- 57:01And that is, you know, you can increase the density so you can have hard disks with more capacity.
- 57:10Right. You can have 8 terabyte hard disks or 16 or 32 and you can increase the capacity.
- 57:18Sadly, the rate at which data is being created at the moment is still higher,
- 57:24than we can add places to put it.
- 57:28Wow. But that soon will, like you said, as things get smaller,
- 57:32more space, that shall change. Yeah, it's a curve that we are still adding,
- 57:39adding, adding, adding, adding data.
- 57:42At a certain point, that is going to level off because we cannot create that much data.
- 57:47Unless, of course, you're going to start putting in the data that all the AI uses and creates.
- 57:53So if you have all the AI making the news articles and all the blog posts,
- 57:58that's all going to go back in there. And then you will have AI reading AI to
- 58:03write. A hundred percent. Right.
- 58:06That's like, okay, what are we doing here?
- 58:11Yeah, because then where's the truth? The truth gets dissolved.
- 58:14It gets dissolved and compacted with AI reading AI reading AI.
- 58:20We're already seeing that because I believe it's Google. They now have a strategy
- 58:24that's like, you know, if you give in your search term what you want to look
- 58:27for, say minus AI, ignore all the AI stuff.
- 58:31Wow, that's good to know. That's coming and is actually here because people
- 58:36said like, we don't know because, you know, AI says something, but.
- 58:43It's 50-50% chance if it's correct, yes or no.
- 58:46So specifically if you go into politics or sports betting.
- 58:57The craziness is through the roof. And then we haven't even mentioned AI hallucinating
- 59:04where it just goes itself off on a complete tangent, which is completely wrong,
- 59:08but it'll make you believe that it is correct.
- 59:10And if you are then absolutely believing what this AI model is telling you,
- 59:16Because, well, it's AI, and it wouldn't lie to me. Right.
- 59:22That's tricky.
- 59:24That's a different discussion.
- 59:26Because then who keeps AI in line? Well, the question is always who watches the watchers.
- 59:33Who watches the watchers. And I don't think currently anybody is. And,
- 59:40everybody is.
- 59:43Many people are still maybe a bit blindsided by all the promises which are being made.
- 59:52And slowly they're discovering that not all these promises will be able to be kept,
- 59:59but that is a slow process and there are still more believers coming in than
- 1:00:06disbelievers going out and then there's the realists on the side of the theater like me and you,
- 1:00:14like just screaming into paper bags laughing.
- 1:00:21Oh my my my my this is a lot to take in it really is um it is it is a lot,
- 1:00:29it it's it's a lot of tech jargon and lots of things of how,
- 1:00:34this world because as you said no the cnn's and the foxes and the whatever the
- 1:00:40world oh ai is doing this but then there's the the entire like,
- 1:00:46the the entire funnel behind it of all the things and the resources that it's using,
- 1:00:53so both you know in logistics or in finance and also in natural resources your
- 1:01:00your power your water etc,
- 1:01:03sure that is something you know that you forget of course now people,
- 1:01:08folks will say like oh you know every time you do an ai search you are using
- 1:01:13this many kilowatts of power and you are you are using five gallons of water.
- 1:01:19If you ask ChatGPT something, you will use five gallons of water.
- 1:01:24No, it depends what your question is.
- 1:01:28AI as such is in the Google Translate. That's an AI system.
- 1:01:33It's like, hey, here's a text in English, translate to Spanish. That's an AI model.
- 1:01:40Will that use five gallons of water? I don't think so.
- 1:01:43But maybe a deeper, maybe a deeper search, something that requires a little bit more.
- 1:01:48So when you say to your AI system, hey, make me a mobile application and a database
- 1:01:55and a website and a payment thing that will allow me to, you know, index the water prices,
- 1:02:02in Western Europe, like the wonderful idea that we had to play around with.
- 1:02:06It's like yeah sure that will because that thing is calculating and doing and,
- 1:02:12stuff in the background like crazy for half an hour in a data center somewhere,
- 1:02:18looking at five million five billion gazillion programming examples that it
- 1:02:24pulls from somewhere because it does like those things don't know how to write
- 1:02:27a single line of code they can't they have,
- 1:02:31fucktons of examples that they throw together and oh if this doesn't work well
- 1:02:35let's try this and this oh this works is the result what we want yeah sure we
- 1:02:40keep it but you know let's do five other ones and well,
- 1:02:44you give the result a value when the best value wins and then you keep this together but then.
- 1:02:50In the end you you use so many resources for something that we did for shits
- 1:02:54and giggles and just threw away and that of course is you know we will have
- 1:02:59used a lot of power and a lot of water as some resources.
- 1:03:02Is it gallons and gallons and gallons? No, but it is something that you need
- 1:03:05to take into account and will be used in the background.
- 1:03:11You know, our little blue dot marble still has a finite amount of resources
- 1:03:17that we are currently, well, not being very efficient with. Right, right.
- 1:03:25So yeah so would i be happy with a data center in my backyard i said depends who's going to use it,
- 1:03:31if it if it would be for the umpteenth ai startup i would be very critical to
- 1:03:36be honest because you know what's their business plan what are they going to
- 1:03:40do what are they they're not going to do what is their,
- 1:03:44management plan for,
- 1:03:46you know heat power water that's that's the thing you know in in,
- 1:03:52any location that has a few hundred square miles of desert, say,
- 1:03:57hey, we'll put something there.
- 1:04:01We promise you 300 jobs. By the way, Mr. Local Official, you have a contracting
- 1:04:08company. We will need a supplier for something.
- 1:04:14That is more of a thing of who watches the watchers because the people that
- 1:04:19approve those things and give the building permits.
- 1:04:23They will have to be held accountable. The question is by who versus what and how and when.
- 1:04:30That currently is not yet happening enough.
- 1:04:35It's starting, but right now, so locally here, I have, well,
- 1:04:41there's Google, Facebook. You'll have data centers about 90 minutes from here.
- 1:04:46Short but as i said they are they are there they are responsible,
- 1:04:52ish enough now they are also next to a river in an old industrial estate that was dying so they
- 1:04:58redid the old industry was all old steel mills and mining and industry which
- 1:05:04has gone away here in this in this in this part of the continent
- 1:05:08so sure you you put the old industrial grounds to a new use,
- 1:05:12perfectly fine You are not chasing away people.
- 1:05:15You are not overly abusing city water.
- 1:05:20You will actually use less power than the steel mill before you. Right, right.
- 1:05:25That's that's all that's all good and and if you replace industry on a one-to-one
- 1:05:31scale that's fine but yeah if you're just gonna say oh we're gonna go into,
- 1:05:37i'm making something up completely we're gonna go to south dakota or wyoming
- 1:05:41and we're gonna put in 10 data centers there in the middle of the desert and
- 1:05:45we're gonna pump up cooling water yeah.
- 1:05:49That's like okay so who does the checks and balances who signs off on that and
- 1:05:52what do they have to win in the end who watches the watch because it's one secondly
- 1:05:56is follow the money right right.
- 1:06:00You know i feel like i feel like in a year not even a year i bet you in six
- 1:06:05months to a year there's going to be more and more developments coming our way or more talk of this,
- 1:06:10and i'd love to obviously i i hope we keep in contact and i would love to follow
- 1:06:16up on any developments on this and,
- 1:06:21as i said like on all things I've said, I hope I'm wrong.
- 1:06:25I literally have proved me wrong. This is how things are happening and going ahead.
- 1:06:31On some things, I hope I'm very wrong.
- 1:06:35Jack, where can, after this article, not this article, this podcast,
- 1:06:39a lot of people are going to want to know how to hear more about you.
- 1:06:43Where can people go to listen more from you?
- 1:06:46Best would be the podcast, IT Horror Stories with Jack Smith.
- 1:06:49So as you said, 25 experience in IT and data center fires and things not working
- 1:06:56when they're supposed to. So there is the podcast with that name.
- 1:07:00Website, ithorrorstories.eu. EU for Europe, because that's where I am based. Love it.
- 1:07:07We are currently in our second season. We are putting out episodes about every two weeks.
- 1:07:14Good for you. And we have currently, let's say, doing a lot about shadow IT,
- 1:07:20which is like unofficial IT that just happened without having me in control.
- 1:07:25So we have laptops running warehouses.
- 1:07:28We have sales departments putting in their own systems that then break and things go poof.
- 1:07:35We have an episode on getting PTSD from the French train station,
- 1:07:41Jane-Elef to go to Paris so many times to put up a 5G antenna on a roof with
- 1:07:47French people that wasn't experienced.
- 1:07:50Wow. So there. ithorrorstories.eu is the URL and everything.
- 1:07:57It's, we're on the Spotify's, the Apple's, the YouTube's, everything that sources
- 1:08:04from there. So any platform you use should be able to find us.
- 1:08:10Excellent. Excellent. Jack, thank you for your time. You gave me so much and just such a...
- 1:08:17I hope not too much because I have... No, no.
- 1:08:21No, no, you did. I always listen to my podcast afterwards because you know when
- 1:08:26you're recording, you kind of have to be,
- 1:08:28your mind isn't always, and I always love listening to it afterwards,
- 1:08:32and then I get a lot more, even with just the talking, but I'd like for us to
- 1:08:36follow up with one another.
- 1:08:37I'm going to send you a link afterwards with this, but I'll have,
- 1:08:42I'm sure, other things, and you're my expert now, so.
- 1:08:46Jack, thank you very much. It was really an honor to have you.
- 1:08:49Thank you so much. Pleasure. If you need something, let me know.
- 1:08:53Yeah. Thank you. We'll be in touch. All right. Have a good one.
- 1:08:56Cheers. You too. Take care.