A (sort of) Clip Show

Marc: Okay. All right. So it'll be kind of loose here, but I think it'll be fine. We're pros.

Renee: We should be after three seasons. Wait, I'll put the timer on.

Marc: Oh, you got a timer? Okay. I'm going in four, three... Hey, Renee, you ready?

Renee: I am so afraid. I am.

Marc: So this is going to be a little bit different than what we've done in the past. We usually have a script of some sort, and then we go off of it anyways. But this time, we're not going to have a script. And so I think it was last week. We were chatting after. What are we going to do next week? I don't know. What are we going to do?

Renee: And then I mentioned— Everybody doesn't know that we don't have a schedule here. Like when we're done with one, we're like, which one do we want to do next week?

Marc: Yeah. So I mentioned Bolero because Bolero, we did the bowling episode and I forgot to do one of the things in the bowling episode around private equity and how private equity is now starting to ruin bowling, by putting strings on bowling pins. And I thought, oh, well, yeah. Why shouldn't we? We should go and, you know, update other things.

Renee: Yes. Yes. The world changes and we would be remiss if we didn't, you know, agree that that's true, right?

Marc: You were going to say something about strings on pins, weren't you?

Renee: They drive me crazy. How is that a thing? How is that a thing? How is it a pin? How? How did they not, like, just drag the pin and then it just drags more pins? Like, I don't. Oh, my God.

Marc: Really strings really yeah.

Renee: Don't they move when you pull them up how do they stand steep to wait for them to quit swinging.

Marc: No it's it's kind of cool actually because they you know you the ball hits them and they go flying around just like normal but because there's string attached to them there's they don't they don't bounce around as much so the cascade i think is not as good but whatever, and so when they're on the when the flopped over, Then what they do is they pull the string up and then they, and then in an ideal world, the, the pins flow into almost like, like, like when you, when you put rack the balls on a, a pool, you know, pool table.

Renee: Oh, okay. So they kind of fit really tight.

Marc: Yeah, they fit really tight. And then it, then it kind of goes back down and drops it back to the floor and then sets it and then, and then lifts back up again. So they're, you know, they're, they're standing. But oh and the thing i mentioned to you last week the reason they're doing this is because the expense of the pin setting machines is really high, and you know we talked about pin boys and and all of that and and pin setters the pin setter machines allowed bowling alleys to scale because there was you know this machine that then allowed them to you know drop 40 lanes or whatever and in this scenario, it's, hey, I want to run whatever my 40 lanes, but those pin setter machines are really expensive. So let's build a less expensive machine. And the less expensive machine happens to have strings attached to the pins. And then, some- You know.

Renee: All I'm saying is like the pro bowler tour is suffering and the pro bowlers are thinking they're not going to make any money anymore. They don't make a ton of money anyway.

Marc: Of course not.

Renee: And they have to drive to all of their competitions because you can't take 30 bowling balls on a plane. And so I feel like I just hope it dies soon so they don't have to live with this. I mean, it's so much of how they play, like pros, is cascading falls. Like when it knocks something out in the corner, they're that good at it.

Marc: Right?

Renee: Like for them to sit there and be like, seriously, strings? Like, seriously, this is what we do now? Like, we're the pro string guys?

Marc: Like, what the hell? I'm sure they don't, I haven't, I don't know, when I did look into this, when we recorded, I just didn't have, I didn't add it to the script. I don't think, from what I read, that the Pro Tour does not run on lanes with strings.

Renee: Oh, good. They don't pull out the duck bowling either, so they don't put the gutter ball guards up.

Marc: The guardrails, yeah.

Renee: Yeah, because I'd need that. I would need that for sure.

Marc: But it is interesting. So anyways, that's the conceit here is that there's things that change, and we like to update stuff, and this is our chance to update on a couple of topics. So we picked three topics, which are very interesting because three things that we talked about in the episodes have all come true in the last year since we've talked about these topics. So and I'd be remiss to not to, you know, mention we love The Simpsons. We love the clip show. Every sitcom has them.

Renee: 138th is clip show spectacular.

Marc: Exactly.

Renee: 138th episode spectacular was a clip show. Yeah.

Marc: There you go. So this is sort of a clip show, but not exactly a clip show. And we'll talk about three topics.

Renee: Mostly because we're not animated.

Marc: That's why. Well, there you go.

Renee: That's one of my favorite lines is when Homer's talking to the woman who does the itchy and scratchy voice.

Marc: Oh, yeah.

Renee: And he says, are we live? And she's like, we don't do live cartoons anymore there, Homer. It's really hard on the artist's wrists.

Marc: Yeah. Yeah. Oh, my gosh. Oh, my gosh. Okay. So we're going to talk three topics here. And I'll give some information and data. But we'll talk a little bit about like where we left it and then and then we'll just go back and forth and and we could talk about the topic so way back, renee do you remember way back in season one episode three, your own hyperlocal meteorologist in your pocket okay all right so yes there you go so what we talked about there's kind of a linear progression for dial and weather to you know weather.com, IOT sensors, mesh networks, what I called raid for weather and personal use cases, you know, sort of trivial next to this, you know, chain of agricultural data and public safety and all these really big use cases. So that's kind of where we left it. Smart cities, sensors, lots of data and all this. But here's the thing that Renee said that, you know, and it's got the truth in it. All right, Renee, your exact words.

Renee: I can't wait to hear. This is so scary. I hate being quoted, but go ahead. I talk so much, I don't even remember what comes out of my mouth. That's why this is scary.

Marc: But go ahead. There you go. Yeah. One day, somebody will just say, I just don't trust NOAA data anymore. And then what?

Renee: I was right. I don't trust NOAA data anymore. I was right. I'm the one that doesn't trust the NOAA data anymore. Because you know what? I got to check five apps. All of them say something different. All of them, by a lot, too. Like, you're going to get hit by a hurricane. You're not going to get hit by a hurricane. You're going to have some winds. You're not going to have any winds. Long Beach is going to get flooded. Well, that one you got right. Like, I don't, no point is any, and then, and then, because the ocean sensors got dodged the other day, we took out the ocean sensors. Like, I don't know why we needed to get rid of them. They were doing nothing. They were sensing. Like, we didn't have to do anything with that. Yet, that's how we tell if winds are coming on shore. That's how we tell if they're going offshore. That's how we tell. Like, there is so much, like, I am frustrated beyond belief that something like NOAA, something like NOAA is being handicapped to the point where you think AccuWeather does a better job. But let me just remind everybody, guess where AccuWeather's data comes from?

Marc: That's right. That's right. Yeah, so let me run through.

Renee: So AccuWeather is trying to privatize weather data.

Marc: There you go. Let me walk you through some data here. And this was happening when we started this podcast. But in February of 2025, 880 NOAA employees were fired. Two months later, over 1,000 deferred resignations and buyouts were approved. So, you know, 1,800, almost 1,900 people in the first few months of 2025.

Renee: And these are our smartest meteorologists and scientists. Right. These are the smartest people we have in this at a time where glaciers are just kind of falling apart up high above the mountains and then causing havoc when the water makes it down, you know, the hill. Like, if you don't think Alaska is going to have a lot of that or the Rocky Mountains or, I don't know, the Sierras, like, we're all headed for this, right? Like, and if we had people watching the mountains or melt, right, or we had sensors that were paying attention or we were doing, you know, we were going out in the field and figuring it all out, like, that would make you feel like that's good. But we don't do that anymore. We don't do any of that anymore. We're lucky. We are lucky we get any kind of forecast at all. As a matter of fact, I'm starting to think about maybe the Farmers Almond Act had it right. Keep the government out of my weather. At this point, I almost feel like a libertarian, dude. I almost feel like that. I almost feel like you don't belong anywhere near this stuff because I can't trust you with it. Right? The idea that you didn't take the opportunity to hire more data scientists to go work at NOAA. Like you didn't take the opportunity to spend more money in NOAA knowing that climate change is coming. Although the president just said the other day that it's a hoax again. Again, after a call. It's all a hoax. So I don't know what to do.

Marc: So one of the things we talked about but we didn't talk about in detail was the fact that every day, No, a launch is weather balloons. So weather balloon launches halted or reduced at 11 locations. Morning launches across the lower 48 states in the U.S. are cut by about half now from a baseline of twice daily at 100 sites. So it's a pretty significant change in how many sensors. And if you go back and listen to that episode, the thing that we really highlighted was that, the number of sensors, just the presence of lots and lots of sensors, increases the accuracy of meteorological data.

Renee: All the way down to your ring camera.

Marc: That's right. Or hyperlocal. Right. If you reduce the number of sensors that are getting launched every single day, and these are really sensitive, you know, things that NOAA launches, then, you know, that's model degradation. So, yeah, fun.

Renee: I'm upset about NOAA. I want everybody to come back. I hope they'll come back when there's something to come back to. And... I don't know. You're looking at somebody who flunked meteorology twice. The second time I failed it. I would evacuate places that the hurricane was never headed to. Like, I don't, I didn't know how to, wind direction confused me. Listen, like, I finally went to the professor and I'm like, listen, you're not gating your profession from me. Like, will you fail me? Because you think there's no chance this chick could ever be a meteorologist. No kidding. I don't want to predict the weather. I just want to read it on TV.

Marc: That's right.

Renee: Can we just cut me a break here? Just cut me a break.

Marc: You were taking meteorology just to be a weather forecaster on Channel 7.

Renee: No, not forecaster. Presenter.

Marc: Oh, presenter. Yeah.

Renee: I just wanted to work with the green screen. That's all I wanted to do. Yeah. We need people far better trained than me. And I'm afraid people like me are the ones running it right now, which is not a good thing. Yeah.

Marc: Yeah. So one interesting thing, though, is that NOAA, in this kind of period of time, NOAA puts its own AI models into operation, though, in January 2026. So that's relatively recent. That's great, right? AI-based AI models for weather. But here's the catch, right? The models are trained and initialized from the, you know, from the data and that observation network is being cut. So the balloons are what the models learn from and what it needs, you know, every morning to start. And that, you know, sort of changes, you know, changes things. We said that resilience would come from lots of cheap, redundant sensors. And what's happening is redundancy is being stripped from, you know, that NOAA layer, while everything leans harder on the model, but, you know, the model can't exist without the data. So, like, think of it, this is where my head was kind of, like, wrapped around, like, How can you forecast something that never happens, right? So if the data has to be generated every single day, The models rely on the data, but the data isn't coming in. So the models are doing what, you know?

Renee: Okay. So we'll go back to what that means in AI. Because what that means in AI is, is that my model's now drifting.

Marc: Yes, exactly.

Renee: Right? Like I'm in model drift now. Like I'm using old data. It's speaking to me like it's current and that I can rely on it. And then I'm going to go use that in other platforms. and other people who pay for my data. Like, that's borderline crazy. That's borderline crazy, right? Because garbage in, garbage out. And you're right. If you're going to cut back on, like, we did them twice a day and we do them once a day. We did them in 100 places. We're doing them 50. Like that.

Marc: Yeah.

Renee: Okay. Okay. So you have partial data that you're going to make full decisions on. That's what you're telling me. So you have partial data, you're going to make full decisions on, and you're hoping they're going to be accurate. I don't care how accurate your model is. By the day after tomorrow, it's going to be out of sync. And by two weeks from now, it's not going to make any sense. And three months from now, it's going to be still telling you that it's going to be 28 degrees out, except that it's summer.

Marc: I think this is the part that, like, that drift, like, it's not detectable at first, right? But at some point, the drift is obvious.

Renee: And then it's really too late because you probably spent months or years, right? Like just using that data, making all kinds of crazy decisions from it. And yeah, yeah, yeah. That's what you're in for. Yeah.

Marc: And then now you're a quarter off, you know?

Renee: If you're lucky. Wasn't that a thing in NIST, though? In NIST, it was like in the security guidelines, SP800-53. Yeah. So in that standard, one of the things they talk, it's not 53, it's the risk standard, 130 something. But they talk about the locksity. the velocity of the risk, right? So, How long will it take me to detect it? Well, it might take me three years. It took Amazon four years to figure out they hadn't hired a woman because they never hired women. And they trained the HR platform on all their own resumes and stuff and their own hiring data. And so when the algorithm took over, it didn't hire women either. It took them four years to figure that out. Someone in HR was like, hey, when was the last time you saw a woman in here for a job? And they're like, oh, I don't know. Why? We should look at that. Like, that's when they figured it out. Like, that's what you're in for, you guys. Like, you do that with weather data, go ahead. But now you're just a data babysitter. You're not a scientist anymore. You're data babysitters.

Marc: So congratulations. So, okay, so here's a question. All right. Does private hyperlocal sensor data actually backfill a shrinking public network? Or does it mostly resell what the public network already produced?

Renee: So how useful is hyperlocal data like that's like that like if i want to know what the weather's like in my neighborhood i can step outside like like in my mind like that's it i think for me it's like hyperlocal i don't give a crap that's like that's literally right now it's like right now right like i'm more concerned about the predictive stuff like what's it going to be like while i'm in london next week or what's it going to be like well yeah i'm going to be in next week what's gonna be like in vegas should i take a coat like it's high desert like like so, yeah so no i would trade i don't think that's a backfill for anything i don't think it even if you used all that data and said i'm gonna do predictive analytics, why would you want that much data to do it from like that seems crazy to me.

Marc: I i think i i'm a little bit i'm a little bit skeptical of. Equating oceanographic currents and global, you know, ocean temperatures with, what the temperatures are on my street. Like, like, yes, they're related. And yes, the data is interesting and useful when you put this together, but the temperature of the, or the, the way that the currents work in the ocean feeds the entire global weather system. And that's different than, you know, it's a one degree difference between the top of my street and the bottom of my street. And I don't think you can rely on, like, I could just see it now, right? NOAA drops their operations, they spend less, they have less predictive capability, and they rely on models. And then what do they do instead? They let AccuWeather build their own set of sensors or AccuWeather buys data from Amazon Ring and that, which they already do. But they then deploy their own sensors. Now, you've put a gate around information that impacts farming. It impacts commerce. It impacts, you know, so many different things. I don't know. I just have a – I kind of like that the government is, you know, has all that sensor data.

Renee: I do too. I actually – I am too. Because then we can share it with other governments. We can share it with companies. we can share it with. It's the sharing piece of that I think Noah was famous for in the world. But it's clear to me that Americans think that there's nothing other than us on this planet, right? Because I don't think they get how integrated we all are. But Noah shares data with everybody. Every airport gets that data. Everybody. And so to give that to, accuweather and for accuweather now to be responsible and we're picking.

Marc: On accuweather because it's a name everybody knows not that they're like actually doing you know.

Renee: If any anybody who is sourcing weather data right anybody there you go yeah and and and hey i have an app that that tells me what the weather is in swear words it's a really crappy day renee like It's all swear words. It's like, F this, crap that. It's hilarious. It's like, another crappy day in the desert. It's going to be 150 degrees, don't go outside. This is how little I care about it now, that I got an app that insults me while it's telling me what the weather is. I think at some point, this idea that a private company would have that data, and now we're going to rely on, them to secure that data, to manage that data, to properly secure it. And then on top of that, use it for what? Flood alerts? Amber alerts? Think of what the government does with this stuff. I'm going to get a text from AccuWeather. Well, what if I'm not a subscriber? Do I still get it? I don't know. But now I might have to subscribe to know that there's a flood coming down the road.

Marc: That's the scary part, right? Like, that's very dystopian, but I can see that happening.

Renee: Yep.

Marc: Right?

Renee: Well, come on. VNW wants you to buy a subscription for the heating package in your car. What makes you think they're not going to do this?

Marc: I know. I know. All right. Okay, last question here, and then we'll move on. Right. And this one is more philosophical, but, like, if forecast quality degrades, what's the detection mechanism? You know, like, how do you detect that? And then who's to be believed then?

Renee: Yeah, this is always my problem. So you would say to Noah, how are you securing your data? Because I've worked in security for all those years, right? At Forrester, I was like risk management was in the security team. So you were talking to CISOs all day long and twice on Sundays. And so I remember I was like, how important? They're like, well, Renee, it's public data. I know. Like, so how many countries rely on this? Like 60 some. Like, okay. Okay. So what if I don't steal your data? What if I just change it? Yeah yeah what if i just change everything by 15 degrees everywhere like, would that make any difference like like like now it's completely unreliable maybe i do by 30 degrees so it's unreliable like like more obviously but i mean you only get one of those and then maybe i'll trust you again do you do that twice and i'm not going to rely on you for anything i'm not like airports will quit relying on you and other countries will quit relying on you because your data can't be trusted does that mean they have to put up their own weather balloons i guess But I don't know what knowing what the weather is like over the UK is going to do for us. Like, I think Americans also forget the landmass that we're dealing with here. Like, it is a very big country, right? Like, it's just too big to behave like this. Like, this is stupid. Weather's no joke. It's no joke. And weather's getting more and more dangerous. And like, being able to predict that stuff, tell people ahead of time they need to get out of the way, you know, having the right information to restore services, like all of that, crazy important. I can't believe this is what we did. And you know what? But I refuse, I refuse to lose faith in the bureaucracy. I refuse.

Marc: I refuse.

Renee: It'll be back. Don't you worry. Paperwork always wins.

Marc: Paperwork always wins. I, I... Well, did you see in Japan, Nagoya, huge floods, most downpour rain ever recorded in the city? Just today or yesterday or day before.

Renee: El Nino, baby.

Marc: Crazy, crazy. And that's obvious. There's a flood, you can feel it. But did you have enough information to predict that? And, I think that the answer is probably still yes, for the most part. But I think it's the insidious change, just enough difference, like one degree difference in those predictions in the sensor data. If the sensors don't have enough coverage, let's say, for, I don't know, here in Kent, it's farmland, right? Well, what if the temperature drifts just a couple of degrees at the wrong time and nobody knows it? Then the crops are gone. People don't have food.

Renee: And you can't figure out why because the weather's good and temperatures are good.

Marc: Exactly.

Renee: Why did we have frost last night?

Marc: Exactly.

Renee: None of this was supposed to happen. I would have covered it if I knew that.

Marc: Yeah. Yeah, just a couple of degrees, just a little bit of difference there. And I think that's the kind of insidious mechanism right there. Like, if the quality degrades, what's the detection mechanism, right? The detection mechanism is bad stuff happens, and it's too late.

Renee: So funny you should say that. So you saw the flood in Nepal, right?

Marc: Yes.

Renee: Yes. Oh my gosh. Yeah, the landslide. And then it took that piece of the glacier with it. On the way down the water, the ice turned into water just because of the friction. They said it was traveling so fast that, yeah, it just turned into concrete, like liquid concrete. It was just crazy. But there were sensors all down that river. It's really funny, though. When you see the data for that, you see like, oh, there's something coming. and bam, sensor's gone. Like, it's just gone. They're like, we didn't know what happened. We're like, did it? Like, why did it? And like, dude, it was buried under 20 feet of water. That's what happened to it. Like, it was buried under the water. So like, so I think like, yes, yes, yes, yes. Like, people, and they said the cities where they could relay the, in that valley in Nepal, where they could relay the information, like, there was a sensor up river, that sent stuff down river. And then another person picked that up. It was like a chain. And then throw it forward until it got all the way to the end. But it never got to China. So all those people at the border, where you saw it come in and take out that building, that giant, like it just came from behind it, and just like it was terrifying. It was like the worst movie. Like I couldn't believe it was real. And I thought to myself, oh my God, they didn't know. Why didn't they know? Well, none of those censors told them.

Marc: Yeah, yeah.

Renee: Crazy. So, you know, here's what I'll say. It's getting more dangerous. It is a national security threat. I don't understand why we are defunding it. It's dangerous. It's stupid. And we need to stop doing that because this is stuff that causes billions and tens of billions of dollars in damages. And we won't even know it's coming.

Marc: Yeah. Like, come on.

Renee: And I don't want any of us to forget that this is the same president who saw the path of a hurricane and then took a Sharpie and made it bigger. Yeah.

Marc: Let's try not to get too political, but yes, that's true. That happened. But, you take the politics out for a second. Because the funding for these programs is not huge. We're not talking big numbers, right? So if you get this much money to spend and you get just tons and tons of benefit out of it and protection, like risk mitigation, why mess with this? Anyways, all right. Anything else you want to talk about on weather?

Renee: No, except that it's like 115 degrees here again today.

Marc: Is it still hot out there? Oh, my God, yeah. Yeah, September is still hot.

Renee: So here's what they said, and I've checked every year. And you can go check the weather. Like, literally go check the weather for Palm Springs, California. And you'll see, like, they tell you, like, after September 11th, the weather drops down to double digits again like that can't be true they're like no it's that specific like like after 9 11 like there's no more 100 degree days and guess what after 9 11 there's no more so we're going down to 98 97, like but even in the in the winter time here like you'll get snow and it'll be 84 here yeah, yeah that's why i live here well.

Marc: Yeah i it was still here in in in non-freedom units it was like 24 25 even a couple of days this year this week so it's still warm today was the first kind of cool day in a while so.

Renee: Wait you guys like and poor the uk i mean you guys get to figure this out because you don't have air conditioning you don't have oh my god in france france is arguing about it like we shouldn't have air conditioners i mean so you want people dying in their homes like go call texas and ask them what happens when that crap happens It's like, no, you need to do this.

Marc: Yeah. Like, that's a topic for another time, right? But I'm big on water security, and this country just does not, they do not treat water, like, appropriately. Anyways, all right, speaking of water, let's talk about data centers. That was the other topic. That was the second topic. Now, we've covered data centers twice before. The first one was our very first episode, which was our second episode, but our very first published episode. The AI gigawatt hunger and then that was all about electricity and then in our second season opener, season two episode one the AI thirst trap all about water usage for the industrial uses of water and and we talked about AI as one of those our data centers as one of those use cases so, so that's so if you want to go back and listen to those our very first episode, forgive me because the audio is not great on that one but it's still good it's still a good episode and then the first episode of season 2 all about the industrial uses of water so, We talked about compute demand always outruns power planning, but this time this is kind of a step change in generation. Efficiency improves at the margins while consumption is planned to double. And we talked about the, we picked the grid as kind of the deciding force around. Either the growth enabler or the growth slower. I said, you know, adding renewables and cooling innovation would help. We both, you know, thought that there would be an AI winter, which means, you know, no more data center development around 2030 if renewables in the grid didn't scale. In the other, the episode about water, we talked about water as a four-pole industrial input. Every historic scale-up crisis was solved by redesigning the flows, not by using less water. So that I think is really important. We talked about a lot of the different industrial use cases there. Where we landed on in that episode was there wasn't a single fix. Then the real constraint is political will. And we'll come to political will in a second. And this, you said this exactly, until public policy matches up with the aspirations of AI, we're not going to be able to do this. So those were your exact words.

Renee: Okay. So proof point here is Texas just closed, they have a moratorium on data center buildouts. Now, Texas was building out the most data centers anywhere on the planet. Like they were hundreds of them. Hyperscale data centers were being built out. And then they realized, oh, hang on, the grid. Like when they actually added up, you know what, here's, I'll give them credit for this. I'm gonna give the state of Texas some credit here. Like not only for being like number one in wind and number two in soil, in solar, you know, in the country. Like you guys think California's hugging trees. Nope, it's Texas. Texas is hugging all the trees. So I'll give them credit for that. And I also give them credit for, you know, actually sitting down and saying, we approved all these data centers. If we add up all of the energy it's going to pull from the grid, it's going to pull more than we create, which means all of the people of Texas will have no energy. And we can't do that because, you know. Pay taxes. And so we can't do that. And Texas has nothing to lean back on because they're not connected to the national grid. So they don't even have the rest of us to fall back on if anything weird happened, right? So if you want people in Texas to have air conditioning and Claude, you're going to have to build out the twice as big of a grid. Oh, and by the way, that supply chain to get things to build out the grid, well, where do you think we get all that from? Well, we get it from China. Well, they're a little mad at us right now. So what used to take a year to get now takes five. So if you're going to build out a data center, you're waiting five years on parts just for they'd be able to put it into the grid because they're going to have to update the grid to get it there, right? So like I stand by that still until you want to give the Department of Energy money to help all the, you know, to help the country, you know, figure out what was once called the greatest accomplishment of the 20th century. If you go to the IEEE, it's the engineering people that do electrical engineering. You go and you look and they say, you know, we're going to rank the top 15, you know, best inventions of the 20th century. Like it's the grid. Grid's number one. It always has been. It always will be. They think it's a miracle. It's a miracle. We have electricity that's reasonably, you know, priced. It used to be anyway. And available 24 hours a day. We don't have blackouts. We don't have brownouts. We don't have any of that. It's an amazing thing. But if you're going to throw 700 data centers, hyperscale data centers into that mess that do megawatts of crap. No. There's a physical limitation to what we can do before you start, Making the people so mad that they'll start doing the crazy stuff they're doing, like going to city council meetings and screaming at people and going to, you know, the housing meetings or the zoning meetings and screaming at those people. I mean, we are like I haven't seen yelling like this since Obamacare in 2008. Remember how they used to like put on their tricorder hats and go run off and like scream at people that there's there's death. What were they death panels? Like Obama wants to have a death panel that kills you and you're not even sick. Like, it was some crazy crap, right? But now we're doing it with data centers. So now it's all like, they're going to poison me. They're going to poison my kids. We're all going to go deaf, and we won't be able to drink the water. Except most of that is true. Most of that is true. So I don't know. And you're talking to someone who used to run data. I love a good data center. I love a good data center. But this hyperscale crap, no. Like, wherever that footprint is, that's about to become a super site. A super fun site, right? So if you want to have 700 super fun sites, go ahead. But that's what we're about to do. Like we're about to ruin the environment over, over what, over what?

Marc: Yeah. So, okay.

Renee: You're making ludicrous deepfakes. That's what we're doing.

Marc: It's not. Well, I mean, not to discount, not to discount the like impact, the environmental impact of your, your Bigfoot AI slot. But there's definitely a lot in the enterprise space that is, that is happening. That's not, you know, not just. Whatever, chat GPT, tell me how to get to Paris and back. There's a lot of real enterprise use cases. I'm not saying right or wrong. It's just not all slop.

Renee: Okay. So I'll take you up on that. And I also think that those enterprise use cases don't take a ton of compute. They don't take large language models. They don't take frontier models for sure. Most of the crap that we're doing in the enterprise is limited to the enterprise. It's limited to corporations. They follow the generally accepted counting practices. They have certain regulatory things they have to manage. They're pretty much all the same. I know they like to think that they're different, but they're not. They're all the same, right? And I would argue that when it comes to that particular use case, corporate, It's limited. It's small. They'll probably do it themselves. I don't think they will, but they probably could, right?

Marc: Yeah, okay.

Renee: And it's probably something that Amazon... If AWS is your cloud provider in the enterprise, why can't they host your GPU? They can.

Marc: They can.

Renee: Why doesn't Azure do it? They can. Why doesn't Google do it? They can. Well, then we don't need all this stuff. Like, what do we need? Like, don't blame it. Don't... I would go back to, like, the argument that, like, well, this is what business wants. Like, here's what kills me, right? The trillion-dollar valuations of these companies. and this idea that they're going to make so much money. If they could just get a trillion dollars, they could finish this thing. But then you go look at the Fortune 500 and go look at their gross revenue for one year. It's about a trillion dollars. What makes you think every company on earth is going to give you all their money so that you can make, I don't know what, AGI possible? Like, no, no one's going to do this. No one's going to do this. If you want to do this, then you get a DARPA program going. You go back into the lab. You let the government pay for it. They'll take it from you before it gets out. And then we'll all be back where we started. And it won't matter because Watson has been doing this for 25 years. All right.

Marc: Okay. All right. Breathe. Breathe, Renee. Okay. So the great rant. I love your rants.

Renee: You should see me on a panel.

Marc: But the, you know, so we both, and a year ago, we both talked about... The public policy side of this. And from an energy and water consumption perspective, there needed to be public policy. And we thought, you know, there'd be more on the national side than the kind of local side. But in fact, it's flit. And the other way around, yeah. And especially in the States, we'll talk, I've got some facts and stuff around Europe and some of the other markets, But in the states in particular, at the local and county, municipal, city, county levels, that's where the political will has started to pop in. And I think that's really interesting. So I'm going to run some numbers here. So largest single campus that we talked about, you know, what we thought would be 1.8 gigawatts scaling towards 10 gigawatts. And there was a site in Wyoming that was sort of unnamed. And that's pretty close to what has happened. Right now, we've got XAI Colossus, which is about one gigawatts. I'm targeting two. And that's 2.6 million square feet, about just under 800,000 GPUs, and it's across the Tennessee, Mississippi line. So that's running right now. There's obviously plans for much larger installations. The, what is it, O'Leary's, the Stratos or Wonder Valley, which was targeting about nine gigawatt installation, which is, it's scaled back to seven and a half. Now, whether or not that actually happens is a different question, but we were right on the money there, like kind of looking at, you know, where, where the numbers were a year ago, where they're going. And those are, those are massive numbers by the way, right? One thing to note on the, on the XAI Colossus, that's a natural gas plant that got built for, for that, for that thing. Yeah. And, you know, so a year ago, even a year ago, we were talking about this, this 1.8 gigawatt unnamed site in Wyoming, which I don't even know if it's come online or not, was kind of treated as an outlier. But in 2026 alone, five different sites of that size, a similar size, have already come online.

Renee: Have you ever seen that graphic where they overlay Colossus over Manhattan and it takes up the whole island?

Marc: Yeah, yeah, yeah.

Renee: Right? Like it's as big as Manhattan?

Marc: It's huge. Yeah. Stargate, which is another project, is announced a nine gigawatt by 2029. A third of a gigawatt live already in Abilene, Texas, six further U.S. sites building. Okay, this is really interesting. Global consumption. What we said last year, 460 terawatt hours a year, but the IEA is projecting greater than double that, 1,000 terawatt hours before 2030. And that trajectory is intact.

Renee: Right? We don't have enough energy, right? And as long as we keep canceling the windmills, because... And as long as we keep canceling the solar, we can't get enough electricity into that grid to support any of this. Like, we just can't do it. We can't do it. And it's not going to be like, oh, well, in order to have AI, we can't have electricity between 7 and 10 p.m. Like, no, no. None of this is acceptable.

Marc: No. So we thought grid plus renewables or we would have an AI winter. Well, what's actually happened is, and since we talked about gas meters last week, it's kind of interesting. Guess what's powering this stuff? Natural gas. XAI built its own 1.2 gigawatt gas plant in Mississippi. And at least three of seven Stargate sites are using on-site gas to skip the interconnection queues. So if you want to get hooked up and connected faster, you build a natural gas plant on your own. Can you imagine like, hey, I'm a hyperscaler and I'm building data centers and compute. Oh, and I'm an energy company now.

Renee: I guess that's what I don't get. Like you're an energy company now. You have one customer. It's you.

Marc: Yeah.

Renee: Because I'm assuming it's not going to happen any other way, right? Or maybe you have a bunch of customers. But if you do, like maybe Anthropics in there, maybe OpenAI is in there.

Marc: That's definitely what's happening. Yeah.

Renee: Right? And so like, and you're a utility? Like, okay, well, now you have to sit under the California Public Utility Commission. Like, okay. So now they get to decide what you get to charge everybody. They get to put, stuff on that. They say that you have to keep all your billing data. They say, like, do you have any idea what it takes to be a utility?

Marc: Yeah, yeah.

Renee: Like hundreds of people just to keep the compliance crap straight.

Marc: Yeah, yeah. Keep that thought because it does come in with one of the other countries here. So where were we right? Absolutely, we were right. The grid was the binding constraint. You know that was like we're totally right, Okay. But where we were wrong, we thought it probably wasn't going to scale as fast as it was. So we were like, oh, yeah, renewables, or else things will slow down. But in fact, the industry decided not to wait for the grid to scale, and it's building its own. It left the grid completely.

Renee: So I wonder what that means to air pollution, though.

Marc: Yes. Yeah, definitely.

Renee: You're trading one for the other, right? So you're trading water pollution for air pollution.

Marc: So so ireland ireland now requires like xa what xai did they built their own gas plant ireland has now if you want to build data centers in ireland which will come to a very scary statistic on ireland in a second, but ireland requires you to build your own so to get a grid correct and get a grid connection in ireland a new data center has to install, on-site generation or storage covering that full demand, and export it back to the grid when the country needs it. And, you know, yeah. So I thought that was interesting. Yeah.

Renee: I like that idea. You know, if you drove down to Culver City to the Sony lot, And you're coming down, I don't even remember what that was, I don't know, Culber Boulevard, maybe? And you're coming down the road and you'll see, like, right as the, like, the studio takes up like four blocks, right?

Marc: Like, it's huge.

Renee: You've been to Sony, right? Like, it's huge. And so, but on one corner, it's a power plant.

Marc: Right. Yeah. Yeah.

Renee: They needed their own power. They, and they have one. Yeah. Right next to me is La Quinta, Rancho La Quinta Country Club. If you go to the corner of their lot, they have a power plant. So, like, you know what I mean? Like, so if you do have, if you are going to pull that much electricity and you're pretty much not going to get it from this grid or you're going to take an inordinate amount of it, then you build your own. And you take it from there, you pull it from the grid when you need to, and then you'll pay for whatever that is. But, like, yeah. So, yeah, of course. Like, you should have to do that. Oh, and by the way, it should have to be a certain percentage of renewables. It can be a certain percentage of something else, but it cannot be gas and it cannot be methane, you know, outflow.

Marc: It can't.

Renee: It can't be. Done.

Marc: Yeah. Well, I just think it's cool that, yeah, if you're going to build, you should build for your capacity, right? Because the grid can't handle it. The grid's not for you.

Renee: That's the right thing. The grid's not for you. The grid was never built for you. You don't get to have it. It's not yours to ruin. Like, the world isn't yours to ruin. You know, it's ours, and we don't want this. And so I don't know why. And, you know, I'm a huge fan of AI, too. Like, I love this crap. But at the end of the day, like, yeah, not in my backyard. They want to build a hyperscale data center down in Coachella. And we're just, we're all mad. We're all mad. It's next to a school. And then the housing around it is all like mobile homes. Like, if you don't think those walls are going to rattle from that thing, like, no, those kids can't learn in that environment. No, no, no. Like, we're down there screaming at people every day. And it's not that I don't love AI. I do. Not that I don't love data centers. I do. But that crap in neighborhoods? No, no. Oh, and that's where they want to build it all. They want to build it all in poor neighborhoods because those people don't complain. And then, you know, next thing you know, no one can live there.

Marc: Yeah, exactly. So what I thought when I looked around to see, because the states, it's easy to see the news in the states because it's quite loud around, you know, this project, this mega project, you know, people protesting, local councils, whatever. And I thought, well, that's, you know, that's interesting. But what's happening outside of the U.S. In this space? So, yeah, in the States, people are complaining, not complaining, but like protesting and proving against land use, the noise, who pays for the substation, and it's being fought at that local level.

Renee: It's getting to the point where like if their local government, like their city council agreed to this and didn't have hearings because they weren't doing that. They weren't having hearings. They were saying, yeah, and they're signing it. And all of a sudden, like, construction started and people are like, what the hell is going on here? And they're starting to strap themselves to bulldozers and stuff. And so those people are taking the very next opportunity to vote out that entire, city council. They all ended up fired.

Marc: That's what happened in Utah with O'Leary's thing.

Renee: Oh my gosh. Utah's unbelievable. Can they just tell you? They get out and protest about data centers every single weekend. 20,000 people go out and do that.

Marc: Free politicians lost their seats over that one, which caused them to scale back their plans significantly. In Europe, it's all about regulators and share, right? Like what percentage of the grid are you allowed to chew up? So that's a regulatory question in Europe or most of Europe right now, which is really interesting. And then in Latin America and Asia, some parts of Asia, it's a kind of procedural thing right now. So the fight is landing around projects being reclassified to avoid environmental assessments and over consultations that don't happen. So you're seeing those tactics where politically things get sort of subverted and moved down a different channel in the bureaucracy, which I thought was interesting. Lots of opposition, you know, obviously in lots of places, and I won't go into all of that. But, okay, here's the scary statistic on Ireland. Data centers in Ireland were 22% of metered national electricity in 2024.

Renee: Oh. Yeah.

Marc: Oh, that's a lot. Can you believe that? That is nuts. More than a fifth of the entire grid in Ireland was used by data centers.

Renee: They, okay, they did go through this thing where they were like, we want to be the place you put your data center.

Marc: Yep, I know. Yeah.

Renee: Okay, again, as someone who ran data centers, like there's a data center and then there's a hyperscaler data center, right? Like those aren't the same things, right? Like they're just not the same things. So, yeah. Oh, my goodness.

Marc: And in fact, Dublin put a connection ban. So that was a grid connections, a ban in 2021. 2021, it was lifted in December of 2025, so just recently, but conditional on full on-site generation or storage plus export back to the grid. So that's where that regulatory shift has changed in Ireland, which I think is good.

Renee: Think how smart that is though. Okay, you can come in here, you can do that, but you're going to build your own grid and then you're going to give it to me when you're not using it so that my grid gets bigger and I can do more things than I want to do, right? Like, yeah. That's not dumb. They're going to expand their grid on the back of Elon Musk. You go Ireland. The motherland is doing it.

Marc: There you go.

Renee: I'm a Murphy. It's the motherland. I'm a Murphy. I don't get to claim anything else.

Marc: So a couple of things on the UK, because that's where I live. And I want to make sure that. My beautiful orchard isn't zapped by data centers here only about two percent of great britain electricity in 2024 went to data centers so that's okay two percent versus 22 all right that sounds good right, But grid connection applications for data centers went from 41 gigawatts in November of 2024 to 125 six months later, by June of 2025, against a national peak of roughly about 45 gigawatts. Okay, so like crazy, crazy increase there. So nearly three times everything the country uses from one industry is what's in the pipeline. And only about 20 gigawatts is at the final investment decision. Like this is the thing with these data centers. They are meeting so much opposition. There's so many of these projects that get shut down. So, you know, okay, some of them make it, some of them don't, which is quite interesting. But here in the UK, it's sort of like opposite to what the rest of Europe is. Data centers are brought into nationally significant infrastructure regime in the beginning of January this year. And that allows, this is the part I don't like, it allows the central government to bypass local councils. So basically, if the British government said, nope, we're going to put a data center here, local councils can't do anything about it.

Renee: Dude, eminent domain is just built right into the language. Like they're not even playing games.

Marc: I know. I don't like it at all. That's a rough one. It's a rough one. But, you know, yeah, anyways, but this one, this one sets me off. It really ticks me off. 84% of the proposed UK data centers sit in water stressed areas.

Renee: So that's the one that kills me. That's the one that kills me. They're building them in Vegas. They're building them in Arizona. They're building them here. Like why? Because we have a lot of desert space. Like that doesn't make any sense to me.

Marc: Doesn't make any sense. Yeah, so let's talk about water. Two-thirds of data centers built or in development since 2022 in the states are in water-stressed areas. Southern Arizona, Colorado River Basin, Texas... Two-thirds of new projects in the States. Google's South Carolina permit to draw 1.5 million gallons a day in South Carolina. That was their permit. Obviously, it drew some formal conservation objections, but come on. That's a lot. But some of the larger facilities can consume up to 5 million gallons a day, which, you know, that's a town of 10 to 50,000 people.

Renee: I was going to say, that's a lot of people.

Marc: It's a lot of people. And just so people, like, go back and listen to the episode because I saw this just nonsense from Sam Altman. Well, it takes, you know, a chat GPT query is less than growing a thousand almonds or whatever stupid thing that, you know, he said. And I'm like, okay, dude.

Renee: I can't eat a query.

Marc: Exactly. I can't eat a query. You're not baking in the water used by all of those systems and people and processes that actually made that query happen in the first place. Like, it's such a dumb thing to say. But, like, okay, the industrial, when we talked about the industrial uses of water, we talked about waste, we talked about cooling, we talked about chemical, we talked about residential, we talked about all these different things. And the thing that's different here is that, These data centers that are in water-stressed areas, they're hooking up to the residential grids.

Renee: 57% of data center cooling water is drinking water. They're hooking it up to the residential. They're taking your clean water. They're pushing it through servers. They're dumping it out into some kind of crazy hole because it's runoff now. It's runoff. Yeah. So here's what was happening. Meta had a data center that runoff was heading into the water source. Right that was going to the filtration facilities where they take water from the river and do that yeah they're finding they have a really hard to cure bacteria they can't get rid of it when they traced it back to its source guess where it was coming from those servers, right i mean like you don't know what you're you didn't do an ecological study you don't know so now they told them you can't dump it you better dig a well and stick it in there you i don't know what you're gonna do chuck it back out of here we don't care you cannot dump it they shut it off you can't dump it back in the river and we don't know what you're gonna do with it now right and so yeah i mean, this is why nobody wants them, right? I mean, just the other day, there's a thing on YouTube. It's a channel called More Perfect Union. And so it's like the liberal answer to craziness. And so one of the things they talked about.

Marc: They brought, they bring- Where's Renee on the political spectrum?

Renee: Right. They take AOC, the congresswoman, they take her, she's a progressive. She's not even a liberal. She's a progressive. They take her to Georgia. And they said, we want you to meet some some Republicans, Trump voters. We want you to be some Trump voters. And she's like, okay, because, you know, and she's like, why are we going? And they're like, where do you see their drinking water? Because they just built a data center across the road. So she goes there and this woman hands her a jar and it looks like oil. She turned it over and it was just brown. And it looked like oil. And it was disgusting. And so it's like a documentary thing. It's like an hour long. You watch it and you're like, oh, that was interesting. And then I'm watching her on TV at a hearing and she pulls out that jar and she said, she says, This is what's happening. Wait, I got another one. And she pulled out another one. And she's like, oh, hey, another lady gave me one. And she pulled out another one. She's like, this, this is one neighborhood in Georgia that's across the street. And the only difference is before they had clean water. Now they don't. And the only difference is your data center. Like, yes, thank you. Finally. Right. But like, that's what happens. Like if like if if it ruins your water, if it doesn't ruin your water, it's definitely going to be noisy. And I keep reminding people that, like, these are big servers. These are, they generate a lot of, one of the things they're generating, though, is infrasound. You can't hear it. You can feel it. It makes you nauseous, right? Yeah. It'll make you sick. It'll give you migraines. It'll make your stomach hurt. Like, it's weird. So, like, that's what people are up against when they live across the street from one in a mobile home. Like, you're going to be in your mobile home with migraine.

Marc: Right? I know people probably think I'm crazy sometimes, but I can hear electrical appliances all the time, like all the time. Fluorescent lights, I can hear them, like power adapters.

Renee: You're sensitive to EMFs, that's all. Maybe you're psyched. Oh, I bet you could go ghost hunting.

Marc: I bet they'd talk to you. This thing here, my Apple chargers, I can hear them when they're plugged in. So, you know, and if you do that at mass scale, like, that would just drive me nuts. It would drive me nuts. I could walk into certain buildings and I can hear things. I can hear the 60 hertz cycle, you know.

Renee: Yeah, right?

Marc: Yeah, the thing that, like, people haven't... Because I feel like there's a lot of misinformation around particularly the water side of data centers. And like Sam Altman, for example, it's just spewing nonsense about his queries taking less than it takes to make almonds or whatever.

Renee: You know what? Because he couldn't make the real argument. The real argument is power plants use 12 times more water than a data center cooling system. Like, right? So he can't even say that, right? Like, we don't use that much electricity. You just asking to do it costs electricity. We wouldn't have to be doing even that if you weren't taking it, right? Like, so, yeah, I just...

Marc: But I think...

Renee: And water's running out, too.

Marc: Yeah, the misinformation side of it, that's what makes me mad. And, you know, people say, well, they're servers. What water are you talking about? It's like, no, we're... Like, yes, there are systems that literally pump water through servers. Yes, that happens.

Renee: Yes, yes, yeah.

Marc: That happens.

Renee: And they're closed loop or they're open loop.

Marc: Yeah.

Renee: So it just depends.

Marc: Right? Closed loop, open loop. But there's also massive, massive electricity generation going on. And you can't just, like, do you know how electricity is generated? Come on now. Like, you burn stuff. It gets hot. What are they burning? Water. It turns to steam.

Renee: Water, coal.

Marc: Coal.

Renee: Because we don't have any of that here.

Marc: Yes, right? Even coal generation. Coal heats water. Water turns to steam. Like people, it's not just like, it's a lot more going on. And every time major industry in history, because this is what that whole episode was about when we talked about this. Saw mills and different chemical plants and all of that stuff, they were not hooking up to your drinking water. They were hooking up to the Columbia River. They were hooking up to the ocean. They're hooking up to different water sources that you're not drinking from. And it's a different scenario. And I just wish that the conversation around water and electricity and usage for data centers was more genuine instead of just, feels like it's just memes and propaganda and nonsense.

Renee: I think it's hard to talk about it and be fair without talking about the politics because they turned it into something really political, right? But at the same time, they're gutting the Clean Water Act. They're gutting the Clean Air Act. There's only two reasons. There's only one reason you'd do that, right? One, you don't like regs. I get it. Whatever. It doesn't matter to me. But you're going to kill the Clean Water Act while we have data centers dumping poisoned water into water sources, drinking water sources, right? And we're going to get rid of the Clean Air Act while Elon Musk is putting tons of methane-burning, backup generators places. Like, we are, until this country's, you know, policy is aligned with what AI needs in a very meaningful way, not we got rid of all the renewables because we're going all in on coal, because that's what we did. Like, I don't, like, we need to stop this. If you want that, you can't have them both. You cannot have them both. Hey, America. Hey, world planet of Earth. You can't have it both. You can't have all the water and all the data centers. You can't have it. So which one do you really, really want? Do you want the thing that's going to take away your job and possibly kill us all? Or would you like water?

Marc: I don't know if it's that binary, but I'm thankful that in Europe there's at least...

Renee: You're still signed into the Paris Climate Accords. At least there's that.

Marc: Yes, yes. We're still signed into the Paris Climate Accords. Anyways. All right. Anything else we want to talk about?

Renee: This is making me mad, and I'm getting an ingestion.

Marc: This is making me mad, too. So we'll go to the even more infuriating topic in a second here. But, yeah, please go check those episodes out. Great episodes. A lot of fun. I really like the one where we talk about the industrial uses of water because it wasn't just a data center thing, but it really helps you kind of frame, like, what we use water for, which is kind of fun. All right. Your favorite topic, Renee, caption.

Renee: I can't imagine.

Marc: All right. Go on. Hulk mad. Yeah. All right. So we called that episode Stolen Cognition, season two, episode nine. And I stand by that, Stolen Cognition. Okay. Boy. So I did go back and do some more research on this one. And I have, like, many times lamented this sort of fact that I get hit with the stupid grid on CAPTCHA a lot. And not only do I get hit with it, but I get hit with it and I can't get past it. I'm sure it's because I'm stupid, but it just feels like I'm being punished. And guess what? I am being punished. Okay. So why? Because for several reasons. And let me rant here for a moment. One, I use browsers in privacy mode, like private mode. Okay. I use Firefox, okay? When I'm at home, I'm on Starlink, you know, which I don't love, but I have to do it because I don't have much, you know, alternative here. What does that mean? I have an address that is not directly tied to a geography. You know, so, oh, I use Apple's private iCloud proxy. What do you get when you put all that together? You get a profile that Google trips on and says, you know what? You look less like a person because I can't get as much information out of your browser, out of your location, out of your privacy settings. So therefore, I'm going to put more barriers up to you, and I'm going to take more of your time. Let me tell you. It makes me furious, furious that this is the case. Like, when did Google become the gatekeepers of access to the freaking internet, just because I don't want to share my information, my version number, my location, like cookies.

Renee: Because nothing's free.

Marc: I block all third-party cookies, and that gets me punished. How messed up is that?

Renee: Nothing's free the internet is not free the internet costs you all your personal data, that's what it costs to use the internet so it started out as hey let's all keep in touch with each other across the universities so that we make sure that, our you know studies are coordinated that we went from that to hey here's this great new thing where we can buy stuff and talk to each other and be in chat rooms to hey you want in, who the hell are you yeah and tell me everything open that road baby show me like yeah i don't yeah, yeah and it's why they're the biggest ad engine in the world right like that's why they're they are who they are and they're they're evaluated the way they're evaluated so yeah, yeah so it is it is crap though yeah i'm still mad about the fact that we were doing all of that machine like training all those machines right like and we didn't know it like i'm still mad about that i'm still mad about it.

Marc: Yeah yeah little reminder for For people that didn't listen to the episode, okay. Captcha, there's a long history around it. But literally, you were training Google's Waymo. Every time you're looking at the lights or the fire hydrants or the buses or the motorcycles, which aren't really motorcycles, they're freaking mopeds. They're not motorcycles. Because they're training the model in India, you know, because that's where they want to put Waymo. Anyways, you know, all of that.

Renee: Every time you were doing that, you were doing training data for automated driving, of which you will get nothing from it. You're not going to get paid for what you just did. You don't have stock in that company. And it was done in a way that was not completely above board.

Marc: Yeah. Not transparent. Okay. So- Yeah, we talked about all those sorts of use cases. The New York Times, OCR, books, you trained a model to read books so that they can then digitize the books to then train more models. Yeah, okay. You see, we should have picked happy topics.

Renee: I'm mad. That makes me so mad.

Marc: We should have picked happy topics here. Why did we pick this one?

Renee: We'll do one show that's like, happy time. happy one one where we did the right things and nobody died, we'll do one episode where everything goes fine yeah nobody dies renee doesn't get angry mark doesn't get angry we'll find one we'll find.

Marc: One we'll have to do something okay, So let's talk about Pokemon Go, because the thing that we were talking about, you know, all of this, you know, the use of Captcha, which is really sort of a more underlying thing around using your data, your thinking, your capabilities to then build that into a big data set for something else. And we didn't talk about Pokemon Go. And this feels like, we talked about sort of what could happen and we predicted that appropriately, but what we didn't talk was about the use case that was going on right around that time. So Pokemon Go. Now, you might think that Pokemon Go is a game and you go and you find the Pokemon and all of that. But there are two things that are going on in that scenario. One, GPS information. And two, augmented reality information. So you're literally building a three-dimensional model scanning the space around you as you're using this. Okay, so what happens to Pokemon Go? Oh, go ahead.

Renee: Wait, can we talk about this? So Pokemon Go was this app where you opened it up and you would scan it around and then eventually it would show you a Pokemon. And you'd be like, oh my God, there he is. Oh, it's this one. I collected him. Okay, none of that was the point. As a matter of fact, they probably didn't hand you that Pokemon to tell you until they were done with the data that you collected, right? They wanted you to scan for at least three seconds so they could get the whole thing and get it rendered right. And then they were like, all right, we're done with you. Go ahead, throw her the Pokemon. Oh, Pokemon.

Marc: Yeah, absolutely. And they set up these gardens and things where you could go find them. But you knew that there was one close by, so you open the app and you open the camera to look for it. And you get closer and closer and closer. You're scanning the space until you get exactly where that thing is.

Renee: So I think, because you know, I'm a master criminal. Like when you do risk management long enough, because you got to try to think of everything that could possibly happen, which is impossible in risk management, but you try to get there, right? When I think about this from a risk perspective, and I think to myself, like, well, wait a minute, they have a database. They know what they have. They know what they have. So, actually, they can plant Hokimons, places that they don't have data for yet, and people will go get it for them.

Marc: That's right. Exactly. That's exactly what happened. That's exactly what happened. Okay. So this company, Niantic- No.

Renee: Trading cards don't do that.

Marc: Trading cards don't do that.

Renee: They don't do that. Just going to put that out there. Trading cards don't do this.

Marc: Trading cards don't do that. All right. So Niantic, which is this company, sold its entire gaming division, including Pokemon Go, to a company called Scopely. All right. So Scopely is a subsidiary of Saudi Arabia's public investment fund through another group called the Savvy Games Group. And they did that in March of 2025 for a reported $3.5 billion. Is that, yeah, it's not a small number. All right, it kept the data. So it was an asset. It moved with the business. Niantic Spatial was a spinoff company, trained a large geospatial model on 30 billion player-captured images of urban environments. The model localizes within a few centimeters from the data that it has, and its intention is to power robots and AI systems, players were running and this is exactly what happened players were running these AR mapping tasks from 2020 and powered up Pokestops from 2021, so that was when they would drop them in very specific locations and then you had to team up to complete the tasks and what is the data used for, right okay well first the first question renee if you had pokemon go do you think that it would be a fair assumption, to say that your data would be used for purposes that you did not consent to.

Renee: I would assume if you asked for my consent and you were explicit about what you were asking for, then of course this data is going to be used for stuff that is not my consent. Like, of course it is. This is why we can't have nice things. Did you learn anything from the last three seasons? We can't have nice things because no one respects anybody else's privacy. So here we are. Here we know. No, of course they're going to use it for bad stuff. And of course I didn't. I wouldn't. I wouldn't say you can use this to round up immigrants or whatever the hell else you're going to do with it. Of course I wouldn't let you do that. And I have no control over it when you do, because America has no privacy laws.

Marc: So what does this data get used for? Like, there's not a smoking gun here, but it's very likely that this data fed AI systems that were being developed for military drones.

Renee: So now they can shoot you within a centimeter of being right.

Marc: Yeah.

Renee: They can take an autonomous drone Send it right to your house Wait till it sees you, And not accidentally kill your wife. Like, that's how it happened, right? Like, that kind of crap is really just beyond the pale. And the fact that you, you know, that the government's the only one buying it, and they're only buying it for DoD, is really just, it just tells you the level of toxicity that geospatial data is. Like, isn't that crazy? Like, geospatial data is toxic.

Marc: It's, and, you know, we didn't- My apologies.

Renee: Esri. I love Esri. I don't want anyone to think that Esri is a toxic company. Just want to put that out there esri the science of where like yeah wonderful people yeah very kind and and i really do like them but i do think there's potential to weaponize this stuff in really egregious ways.

Marc: Well and and you know, what what do you even say about meta and meta glasses like those devices are like like an always-on pokemon go you know and, it's getting the same information and you're giving that information to facebook like.

Renee: On top of that here's what's crazy so it's like for in my mind it's even more than that and maybe the glasses don't do it today but they will soon or and i might by soon i mean like by next year probably but um, probably that your retina is as unique to you as your fingerprint yeah no two retinas are the same right this is what they can do they can scan eyeballs right yeah um but your eye movements are unique to you, what you look at first, whether you look left or right, like, you know, where you look at on the screen, right? They're going to start recording that too, right? Because it's going to tell them something. They're going to figure out something from that. Maybe all women look up to the left hand corner. So now the ads for women's stuff on Facebook will be up in the left hand corner. Like, that's what they do with this crap, right? So like, in my mind, like, yes, you're right. Those things are horrible. They're horrible. They're horrible. On one hand, I don't care if you have it and it doesn't have a camera. If you want to see crap in front of your face telling you, you know, what the movie times are, go ahead. I don't care. But you recording anything? I mean, even the ads are ridiculous. Meta, take a picture. Like, seriously? Like, you couldn't pick up your phone for two seconds? Like, the fact that we could pick up a phone and take a picture. You know what it was like back in the day to take a picture? Like, you had to get a camera. You had to get it out of storage. You had to breathe the house.

Marc: I mean, the fact that in some countries, it's a it's a law required law that when a picture is taken that a sound is played you know so because you know the people creeping like yes, yes messed up and now it's like i don't the where's the camera the camera is, i don't i don't have it in my hands it's literally sitting on my face and that camera you know it takes a picture and.

Renee: And that's the thing. I can't even see what it's doing. I just know it's attached to my eyeballs. I can't see what it's doing. I don't have a, I don't have a, it's all stupid. I hate it. They didn't get away with it the last time. They'll probably get away with it this time.

Marc: Of course they will.

Renee: I don't know if they're going to get away with the camera.

Marc: They already have it. It's already there.

Renee: Yeah, but they're getting a lot of, they're getting a lot of pushback on that. So Meta actually goes to court. So Meta goes to court because a bunch of teenagers sued them because they felt like they were being, it. The Meta knew that what they were creating was addictive and it was causing them, you know, you know, mental health problems. Yeah, of course. You know, I always think I'm being managed by an algorithm. So, yeah, I get it. Anyway, they win. $581 million, something crazy. I think at the end, Meta settles because they don't want to pay a trillion dollars. That's what they'd have to pay, right? So they settle with these kids for 500 and some million dollars. But all the attorneys walk into court the first day, they're all wearing meta glasses, every single one of them. And the judge looked at them and said to the bailiff, get the trash can. And he got the trash can and she said, throw them out. I told you, don't you bring that into my courthouse. Don't you do that. You do it again and I'm sanctioning all of you. And they put their glasses in that trash can and they took them away. Yeah, they thought and he thought that that would that would fly. But he could do that. Look what Elon gets away with. Like Zuckerberg's going to get away with.

Marc: He got he got busted for, you know, you know, securities fraud. Anyway, that's off topic with some other time when we talk about securities. But yeah, look, CAPTCHA is, you know, the premise of the episode was stolen cognition, right? Using your data, your information, your toil, right? For actions that you didn't consent to.

Renee: And one day you're going to buy it back from them. That's what makes me mad. I'm going to buy it back from you? Like, no, no, no. This is all dumb.

Marc: Yeah, yeah. No, it's not fun. And, you know, the irony around that whole stolen cognition piece is that, you know, when you're running through the checkpoints and the gates of flipping the, you know, doing little puzzles or trying to find the objects or whatever, you're going to buy it back from them. The humans are failing more than the bots, and the bots pass the test every single time.

Renee: It was like 99% of the time, right? And then we only get it right half the time.

Marc: It was some stupid number.

Renee: If I get it wrong, you should let me in. That proves I'm not a bot. I got it wrong.

Marc: I know. I know.

Renee: How's that not the metric?

Marc: Yeah.

Renee: Oh, here's a dummy. Let her in.

Marc: And so the proof of humanity is not that you can or can't do the task. The proof of humanity now is that you've accumulated a long surveillance record. That's that's the proof of humanity now. Like I don't have I don't carry a long surveillance record because of the browser I use and the IP address I have and the no third party cookies and whatever. Like that means I don't have a long surveillance record. And so therefore I'm challenged more. How messed up is that? You know, so.

Renee: That is kind of jive because all it means is that I don't run into it that often anymore, which just means they know enough about me. They don't even know anything else.

Marc: Yeah. Yeah. Yeah. Yeah. All right. Anyways.

Renee: California passed a law. Wait, just one more time before we go. California passed. If you are listening to my voice and you live in the state of California, you are permitted now to go to a website on the state's website. There's a section there for be forgotten. Like I request to be forgotten.

Marc: I want to be forgotten. Yeah.

Renee: Yeah. And you put in that request. It takes about 40 days, they said. But in that 40 days, you will be wiped from every broker on earth. And because you made the request in the state of California. So, yeah, I do it like once a quarter, maybe. I make the request. And I don't do it that often, but I get emails every once in a while saying, your request has been, you know, fulfilled. So, yeah. But, I mean, it's California. It's California.

Marc: Yeah.

Renee: If the whole country did it, yeah, there'd be no Google. There'd be no Facebook. The whole ad business would just fall off the bottom.

Marc: Right?

Renee: They'd have to get back to making actual ads.

Marc: Yeah, I was thinking about this the other day, that digital marketing and advertising, which is close to a trillion dollars, in size every year, almost a trillion in spend every year. Like, it's ruined. It's ruined the internet. It's ruined the experience.

Renee: Yeah, you know what? It has. And it's ruined our brains, right? Like this idea that, you know, an ad should take, You know, four seconds, right? Like, it's just ruined our brains. But, you know, I'll go back. I wrote a whole paper about how I thought that AI companies don't understand love. They just don't understand it. They don't understand, you know, community. They don't understand, you know, big cinematic, you know, representations of things. And I go back to 1984, the ad from Apple.

Marc: Yeah.

Renee: Right? Where the girl runs into the arena and takes that big hammer and just throws it into the screen and blows it all up. And everybody realizes, oh, my God, we're awake. Right? We're awake. Like, we're no longer. And I say in the paper, I say, like, well, what Apple was saying to you was we're not big brother. IBM and Microsoft are big brother. It's not us. We're liberating you. We liberated you by Mac. Right? And I'm thinking to myself, like, that was the last great technology ad. It was the first great one and the last great one. Maybe the two guys were the cool guy was, you know, Microsoft and the dorky guy was IBM. That was pretty good, too. But like, we've never.

Marc: Also an Apple.

Renee: Also an Apple one. Like, we've lost that ability to tell those stories that way. And that's really a shame. Advertising used to be fun and it used to be cool and it used to be cinematic and had themes and it's all gone. I feel bad for your kid's generation who just didn't grow up with any of that.

Marc: I know. Because it was great. I have this guy that I follow on LinkedIn, and he's a good guy, but he's obsessed with digital marketing and advertising channel expansion. And he is just a fount of information, of the numbers and the data. I don't know how this guy's brain works. So he posted something the other day about, automatic billers. And wouldn't it be great if automatic billers actually opened up their capability to put digital marketing and advertising in that channel? And I'm like, what the, what? What the hell? No, there's a reason why I have automatic payments on my gas bill. You know, I don't want to hear from the gas company. I don't want to hear from the gas company exactly and his angle was that actually there's a there's an opportunity there and there's you know there's a there's an opportunity around agents and agentic advertising and all sorts of stuff which i'm still dubious about uh overall but you know, and i and i and i just i told him i'm like yeah no that's like there's a reason that people do this and it's to not be inundated with all of this stuff so if my gas and i would probably.

Renee: Go back to the idea that the public utility commission won't let you.

Marc: Yeah yeah well there's that too there's regulators i mean do we know that there's.

Renee: Regulators like i don't.

Marc: I mean if if my electric company all of a sudden like started to you know try to proposition me every time i paid the bill send you those coupon books like yeah from.

Renee: Your gas company like.

Marc: No i don't need your help go away like that's that's the dystopia that digital marketing and advertising has created. And this dystopia, you know, coming back to the capture stuff, like that's it was intended for a good thing. But then it was perverted. It was subverted. It was transformed into something else. And it's just, it's corrupted.

Renee: And the joke was on us.

Marc: Yeah, the joke's on us.

Renee: And the joke was on us. And that really makes me mad. We're not dumb. We're not dumb, right? That just really, but then again, I'll go back to, this is what this podcast taught me was, this is why we can't have nice things. And everything cool will come out of the Department of Defense. You're just going to have to deal with it, right? Like, I have GPS thanks to the Department of Defense. Like, it's terrible, but it also keeps me from getting lost. So, yeah, and that's how they normalize DOD. That's how they normalize technology, too, right? Give it a consumer, like, way of looking at it. Like, just like the peaceful use of the atom. Let's figure out, let's go paint plates in radioactive lead paint. Okay. We use the radioactive stuff. Let's stick it on food.

Marc: Yeah.

Renee: All right.

Marc: But, you know, that's a whole other topic sometimes. I'm like military research turns into, you know, usually it's like a one for 10, you know, one, one dollar spent in military research turns into 10 in the commercial, you know, consumer space in the non-military space. That's great.

Renee: And I dig a drone. I just don't think they should be used.

Marc: Yeah. Yeah.

Renee: Yeah. Okay. So at any rate, I, so yeah, that's what I'll say. I'll say that kept, you know, we all start out with the road to hell is paved with good intention.

Marc: There you go. William Blake.

Renee: We'll just leave it there.

Marc: Yeah. Anything else you want to talk about on CAPTCHA or any of the other topics? All right.

Renee: I'm just mad that I taught auto drive how to drive. I'm mad. Still mad.

Marc: So there you go. If you're angry, if you're upset, if you believe that we can't learn from the past, this is the podcast for you. Thanks for listening to the Nostalgic Nerds podcast. We'll be back next week, hopefully with something less. Yeah, I didn't want this to turn into this like kind of...

Renee: You know what? Here's what I'm going to say. Like when we do the podcast, it's scripted. Like you and I, like it's scripted. We write that. We write it and we know what we're going to say. You know, there might be sidebars once in a while, but we don't do them that often. If you want to know who we are, this is your podcast.

Marc: This is it.

Renee: Want to get to know us? This is your podcast, right? So yeah, that's what I'll say.

Marc: Yeah. So thanks for listening, everybody. And we'll have something new.

Renee: We'll be back next week with something fun. Yeah. All right. Thanks, everybody. Thanks, Mark.