AI Power Play Military Pressure Energy Mandates and Sovereignty | TSG Ep. 1028

Hello everybody, and welcome to the latest edition of the Text Drawing Gang. I'm Mike Biard, and we are joined today with some of our usual regulars. John Schwartz, how are you?

Good to see you again as my friend. I'm good. I hope you're surviving the snow.

I am. You know, hats off to the guys in New York and the sanitation departments been cleaning just about everything up, and we might even hit 40 degrees today and melt the rest of it, so we'll see what happens. We also have Fred Wilma.

Fred, how you doing? Doing great, bro. Good to see you.

Good to see you. And finally, Gina Rosenthal joins us once again. Gina, always a pleasure to see you as well.

Good morning. Good to see you T All right, so the hits just keep coming from Washington, dc I mean, it's amazing. Every morning I wake up and there's something new and different that these folks are up to.

The latest involves Inro, and I guess philanthropic is now saying that they're gonna ease off on some of their commitments to safety for ai, because, well, uh, they were having some issues with the Pentagon about, well, who can actually supply the Pentagon, because, well, I think the Pentagon is trying to say, we're the military and safety's not exactly our primary concern. John, what's going on here? Yeah.

So in a sense, you're right, Mike. So philanthropic in a sense, is softening their stance on, um, privacy or a ai ai safety amid this pressure campaign by the Defense Department. So, for a little context, um, a uh, philanthropic has been doing work with the military.

They're the one AI company that's really thrived in that area. But they've been confronted by Hsif, the Secretary of Defense, who's given them an ultimatum. It went straight to the CEO, and basically, um, they want a softening of their stance, but if they don't soften it, they risk losing a $200 million Pentagon contract.

Now, this is happening as OpenAI and Google and others are starting to escalate their efforts. So Anthropic, one of the co-founders, Jared Kaplan, gave an interview with Time Magazine and he told them that they're no longer gonna adhere to their original constraints because they're concerned about competitive harm to themselves because their rivals are blazing ahead. 'cause they're going a little bit taking an easier path than anthropic.

So philanthropic finds itself in this kind of weird situation. And in a sense, they did, um, officially put out a new responsible scaling policy, which brought up all these questions of whether they were caving in to the federal government. There a spokeswoman, of course, said that it was totally unrelated, but it let looks suspicious.

And, um, it's raising a lot of questions about them because philanthropic wants us all to think of them as the apple of their industry. But in a sense, they seem to be retreating a little bit, or maybe they're trying to kind of split hairs and try to appease both the government and their loyal base. It's gonna make for an interesting dynamic and an argument and a, a debate, and it's really cast a light on anthropic and its original mission statement, and now how it's being interpreted or how it's being, uh, looked upon by the, the general public.

Mm-hmm. Now, friend, help me out here. 'cause there seem to be making a a a comparison here that doesn't really hold for me.

So what they're saying is, I can't innovate without safety. And what exactly do I, is there a trade off there? Or can I have my K two?

Because I don't see why one has to be at the expense of the other. I mean, innovation can still drive forward, and safety is safety, but so if somebody kinda, uh, I don't know, trying to cover a track here, I think the challenge they're running into, and yes, I agree that safety should be, like all things we talk about in every occurrence of safety should be done at the time of origin and the place you're building and all the great things we go with it. It's how you don't bolt it on, you build it in.

But, uh, I can understand the pressure, uh, the application, the pressure, but there's two main reasons why the Pentagon wants to be able to remove some of these safeguards. And I think there are clear and distinct reasons why, uh, anthropic is concerned about it. Uh, AI operated weapon systems and, uh, mass surveillance.

And these two things, typically, if you took out, uh, the AI weapons systems conversation, the mass surveillance thing is, uh, is a problem, uh, of, or a topic of discussion for any, uh, company that's in the, in the, uh, commercial space, uh, different if you're public sector and, and, and the things that go along in this case or the federal government and $200 million of contracts, uh, you know, it's a fair amount of money. But at the same time, when you look at the market cap and what these guys are doing, it's not, it's not at all. And so are you willing to bend on what you know, to already be a console or the reality of the, uh, the efficacy for decisioning, uh, with what you have today in fear of that someone else will do it?

Or is this a little bit like, you know, star Wars analogy, rogue one, you know, somebody's gonna build a death star, somebody's gonna do it. You might as well do it and put a two meter thermal exhaust board in there. I don't know.

But the question is, right, the timeline for this type of execution, not just the mandate, there's a million things we could talk about on policy and whether not the federal government should dictate, uh, you know, uh, private industry about, uh, its use for, uh, you know, mechanized warfare, automated warfare, uh, in addition to, you know, enabling warfighters with mass surveillance in order to conduct, you know, uh, kinetic and non-kinetic operations. But I think the bigger concern is, you know, what, what John said, you know, if these guys believe they're, or they're branding themselves to, to position as being the Apple, right, of the industry of ai, this is clearly not going to help them on the case of privacy or credibility with, uh, users that have access to these things. And here's a really important thing that just came into beta.

So, quad Code now has remote control sessions. So anthropic now has, could have right, access to every single device that Quad operates on. It democratizes its agent models down to your laptop with access to everything, then has a persisted connection to your devices connected across all of your devices conducting whatever operations.

And the authentication requirement for this is just cloud ai. So I just want you to think about that for a second. When you put in context the magnitude of what some of these decisions are, and then cross, you know, connect that with what it looks like to utilize that kind of app, uh, apparatus for me, other than the holistic good.

Mm-hmm. You, you know what, one thing I wonder about is what was the leverage the government had to get them to loosen up? And, you know, if I think about what Claude is, Claude is an LLM, right?

Like, massive, massive. But, um, what maybe, maybe the government just refused to open up access to all of their data sets that are com, you know, things that have never been ified. And given Just, Gina, just let me interject one thing, because I think that what they're basically doing with the government is doing, I think, is now that they have other options and other suitors, they're using that as a way to kind of leverage anthropic.

And they're e essentially they're saying, we're gonna blacklist blacklist here, put you on a blacklist if you don't cooperate with this. So, you know, I'm sorry, just No, but that's, That's another, I'm sorry. One, one step further than that, which is a little shaky in my mind.

They also sent out some letters to every contractor for the United States government, every one of them. And they asked them, what models are you using? Because what they were basically about to tell all these people is that you can't use philanthropic.

So, and suddenly then this article from shows up in time where the anthropic guys are backing off because it's not just losing the United States government, it's all these contractors that supply mm-hmm. I see that too. And I wanna believe that Anthropic has this, um, this really good intentions behind them.

But there's lots of us that remember that they trained on the worst sections of the internet, and they trained on stolen library books and all the rest of it. So like, they didn't start from a really great place in some of our minds. But there has to be, I think there has to be, besides just the, the potential that, that's a big lever, right?

The potential government contracts and the contractors and the rest of it. What if they didn't get access to every, all of the data sets that the government has, and they realized that everybody else, all the other big models were gonna get that access. And, you know, maybe to what Fred was saying, maybe they feel like they've gotta be in there as a watch person.

I'm gonna hope that, but you know, at this point, I don't even know anymore. What are you gonna do? Yeah.

I mean, in fairness to philanthropic, I mean, they're trying to get ahead of this issue, or maybe they're reacting. I, I, it depends on your interpretation, but at least they, they put out a post where they're trying to be as transparent as possible. But, you know, it all comes back to, and I think Fred mentioned this, it all comes back to Apple.

And for Anthropic, this could have been their Apple moment. Remember a decade ago, San Bernardino and the Justice Department was really putting the squeeze on Apple to give it personal information about around the shooters and the information on their iPhones. And Apple stood up to them and set a precedent and actually doubled down and enhanced its reputation as a privacy and security focused company.

So now we're 10 years later and Anthropic is in the same situation, and it seems like they just back down to a certain degree. If not, is it significantly, Fred, is there another way to look at this? Will some companies, because at a concern for safety, consider a different AI model because that one that they perceive to be a little more safe than either anthropic or, or open ai or any of the leading players at the moment who all seem to be trying to drive as much revenue as possible?

It's a good question, but I think this doesn't change, uh, the opinion on anthropic, and here's why. This is a hundred percent transparent, both on the risk, and they've also published a number of safety and security papers to tell you exactly what models are capable of, right? So they've established a benchmark of, of at least transparency, if not truth, right?

In those cases. And I think what's illustrative here, there are, right, so perplexity just, you know, launched a basically an agent of agents that allows you to do almost the same thing, right? With, you know, quad code agents, the battle's on the desktop.

And I think if the worry is that you remove sort of the access of all of these things, devices, uh, you know, contractors, even some probably lucrative contracts that are supply chain, uh, oriented, you do lose a little bit the arms race in a, in a figurative sense. And if you are contributing to the arms rates in a literal sense, there might be a trade off, right? I mean, I think there's another important thing that we're, we are not sort of talking about.

So, other AI organizations may be led by, you know, uh, com communist states abroad are actively participating in, you know, distillation attacks against the latest model of this and the latest model of that, you know, there could be some value for an anthropic and for others to get beneath the shield, right? And in essence, uh, to help them, you know, save some of the, you know, critical intellectual property that they're, that they're trying to build on, if their hope is to build something Apple-like in the future. So, uh, I don't, I don't wanna give them too much credit, but, you know, there's certainly an awful lot of opportunity there for utilizing some of the, the vehicles you're being afforded by the federal government to better, you know, the overall sort of circus of soul that we're starting to see.

Show up here, John. Does AI safety become a political topic? Am I gonna see this Yes.

Campaign? Yes. Yeah.

I mean, we've talked about this before, but I, i, given the, what's going on with the pacs, the, uh, references that there's just a more of illumination, um, as we get closer to the midterms, yes, AI safety is becoming more of an issue. There's a lot of grandstanding. I'm gonna, I'm, I will grant you that, but people are more attuned to this than they have been, say, a couple of months ago.

And it underscores this, this, this argument, this, this, this story. Actually, by the way, I was just watching Yemen now, and this was one of the biggest stories they have in their rotation. They've got a lot of stuff to choose from today, um, given all the depositions in state of the union reaction, et cetera.

And this was one of their major topics. So I think it does resonate with people. Yeah.

There is one thing that I, I wanna add before we close this thing out. Uh, you know, AI targeting systems have been used for a while. This isn't brand new.

Uh, the force leverage, I think is an interesting part here. And then what economies of scale of weaponry are we talking about, right? If you're talking about targeting for, you know, specific types of guns, okay, if you're talking about control of a nuclear arsenal, different story.

And I think, so the, the everything is a little bit of a gray line once you concede the first one. And I think the question mark is how do we help establish and set boundaries and policy before this runs away? Uh, like the pancake man, you know?

Yeah. Well, I think what we've established here is that, um, if you're an AI uncle Sam wants you, whether you like it or not, we'll move on to this topic again. Um, well, let's continue with our conversation here about the federal government because, well, during the State of the Union address, uh, president Trump, almost as an aside, said that he had told all these AI providers that they needed to generate their own sources of energy.

Now, uh, we've seen some investments in that space. Some folks are building some private grids, but it seems like that might take a little bit of a while before we see that. But Fred, are we gonna see all these AI providers get off the public grids?

Is that the ultimate goal? And how long might that take? Well, it's, uh, certainly an, it, it's an interesting contrast.

If you just said you wanted to enable basically the full surveillance of all Americans, and now you want the companies to pay for it, but you're going to use it, it's an interesting place to sit. So if, if Chris Wright said that basically the discussions with all of the Ai, ai developing leaders, uh, 6, 650 to 700 data centers dropped on the grade over the next 10 years, you know, and, and looking at what the costs are for the average person in the current climate, in the current economy of, you know, an 8% increase today, uh, the great question is whether or not the affordability problem is gonna be something passed on the consumers. Uh, I have my doubts.

I mean, probably, uh, probably, of course. And so, you know, now if the requirements are in order for you to have, have government contracts and to continue to build in the space race here, that is you fund the, uh, the grid part of this problem space. You know, there's an awful lot of things that we think about with respect to, you know, the different types of taxes that are not probably gonna be applied states that have the, the wherewithal and the bearing to do that.

I know Gina has talked about this in, uh, in hometown, right? From her perspective of what that looks like. And the real question is, uh, if all of these other, uh, normal methodologies of managing sort of a growth, whether it's the national grid or it's a privatized grid by, you know, Microsoft or whomever for their environments is governance.

And so, you know, there's no effect of, you know, privatized groups, uh, by, by some of the governance that normally the Fed would be held accountable for. So little bit of a stroke of genius to say that, right? Uh, if you're going to force, you know, organizations to build their own infrastructure, you've delegated the accountability, but, uh, it may be an awful lot worse for the consumer.

Yeah. I think it's more than cost too, right? You can just look at the example of Musk in X AI in, is it Tennessee?

It's in Tennessee, right? Um, they pick the poorest part of town to go into, and there wasn't the infrastructure, uh, for power. So they've got generators run running, which are not, that are running on, um, it's not natural.

I'm not sure exactly because I'm kind of gas, but they emit a ton of NOx, noxious, toxic, toxic, you can say those at the same time, noxious toxic smoke into the air, into those areas that already, you know, had been living with that and got it under control with the help of the government. Um, and, and so that's what a privatized, um, power supply is going to look like. And it's usually where they're building them, at least here in Texas, is out in the boonies where, um, the water supply is impacted.

First, people don't have drinkable water, and if they do resort to their own private type of energy to, to power the data centers, it's gonna be most likely these generators that are going to pollute the air. So it's really not a great situation. Well, it's a good thing that climate change is fake news because, you know, all that stuff, I don't have to worry about it.

Right? Woo, good job. There Is something also interesting here that I think, uh, if the assertion is that this is going to help keep up with the global race, um, you know, when you add an additional tax right, to the already, you know, beleaguered sort of level of investment required, right?

You're talking about 10 billion on a hundred billion dollars. You're talking about the, the magnitude of those things. Um, I'm not sure how you can expect, you know, organizations like the Frontier Labs to, Hey, hang on, let me continue to push the race here from an AI perspective, not plateau, because now I have to split that cost on energy, build out for every piece of infrastructure that I build, and somehow maintain the purpose of having a, in the future, a publicly traded company, not completely under the control of the federal government.

Well, that's an interesting point, and I will bring this over to John, but my, help me walk through this line of reasoning. So most of these AI companies are already underwater, right? They're getting massive amounts of investment, are subsidizing the usage of their platforms at incredible rates.

So, you know, that 20 bucks a month hardly pays for anything that you get from your average end customer, and they're hoping that enterprises will help bail them out. But If I gotta pay for the energy, now I added to Fritz's point that cost to their p and l and it makes it less likely they're gonna be profitable. And at what point maybe does the United States government decide, well, if we want AI and we wanna be competitive, this becomes some sort of nationalized service.

John, how crazy might I be? I think you're not crazy, Mike. Um, I, I think there is actually a lot, a lot to that.

Um, you know, there's actually, I think it's in Tennessee. I wanna, I'm pretty, yeah, it is Marsha Blackburn, who was, who is a senator, I think still, technically she's running for governor, but she's also introduced this bipartisan bill about limit limiting data center growth, et cetera. And I think the companies especially, I mean, I saw this number, I think it was a, yesterday, it was from, uh, uh, Jensen, I'm gonna read it up.

Tech companies are spending $700 billion this year on AI data centers. And Jensen says it's nowhere near the peak. So these companies are looking at alternatives, like they're looking at nuclear, uh, they're thinking about the p and l.

Um, God, I think if I'm an open area, I'm really looking closely at the p and l. If I, if I'm an philanthropic, you know, I'm, I'm influenced by what's happening with the government. It's, um, they are attunely and they're nervous about what, where this is going.

And it's, it's weird, it kind of a's not here today, but kind of like I was thinking about Alan and how, what he may have said, I don't wanna put words in his mouth on his, not around, but, um, it, we've kind of been here before in terms of fighting over resources that equate to power. Like I think of the 1930s in LA and the fight over water rights, the water wars, and I'm thinking about electricity and I'm thinking about companies who say they'll pay for it to play nice with the government. But, you know, there's a lot of going back and forth, Mike.

I mean, there, the, the narrative on this is gonna change so often over the, over the coming months is, is we get closer to pressure for these companies to show some sort of roadmap to profitability. But this is hanging over all of their heads right now, right? It's a massive anvil over their heads.

The internet was that way too, right? Because I can remember being from a small town, we were lucky to live next to big Air Force bases, so we were covered. But to get those trunks down to, you know, from the big cities in the south, like Birmingham and Atlanta down to the Gulf Coast, that was really expensive.

And the big trunk was near Panama City that came down. So getting the, getting the access to internet when it was all wired was a big deal. And the government did have to jump in and help because we were digitizing all the information.

We wanted to require everyone to use the internet, and the government had to do something. So I, I kind of see it as the same as that. Right.

And to Fred's point, I'll toss this back to Fred, but you mentioned China. So we see that they are making massive investments and not just AI and, you know, some of which they may be borrowing, uh, illicitly, but I don't know if distilling code from somebody who stole code in the first place to train stuff constitutes a crime, but we'll see. Um, but the thing that comes to mind is they do have massive access to energy resources, right?

And that's being done with the aid of the Chinese government. So at some point, you know, the Chinese model force the issue around the world because, well, if we wanna compete, we gotta figure out how to provide energy to these AI resources and maybe the services themselves in a way that's cost effective for everybody. 'cause right now it doesn't seem like it's gonna happen that way.

Yeah, I completely agree. I think, uh, as soon as the government has decided that all, um, may I, companies have to by proxy become power companies, I don't think any, uh, I don't think Anthropics set out to, you know, participate as a power company, uh, across, you know, the National Grid. The challenge is definitely going to be, you know, when does this become a bubble that's, uh, a burden they can't bear?

And I think critically, uh, I get the rationale behind sort of the midterm elections and sort of softening a little bit on the, you know, governmental impacts. But fundamentally, it's a flawed concept in my mind. And in addition, I'm sorry.

No, It's okay. Were you talk So can I Yeah, that, see what you said, Fred? The, and sorry Gina, I'll, I'll pass it to you really quickly.

But there's, there's this federal data that showed that residential rates are increasing by 8% in 2025. And we're in this era, and it's gonna go higher. It's in this era of affordability.

So this plays part into the narrative about costs. And I think there's a real resentment among the general populace for these AI companies. They're overperforming, their, their stocks of their values are skyrocketing.

And yet there's this fear that they're, they're being coddled by the federal governments. You know, it's like the tech bros and big brother are, are in, are in bed together. And, and it's, it's, it kind of escalates and builds upon this, this whole tension that is, is building about AI and the haves and the have nots.

It's like, it's like the internet, right? With the, the digital divide. I mean, in a sense, we have the same type of thing on an economic scale.

It's definitely the same thing. And they're definitely in bed together, which is where I was gonna come in and say, I think there's another angle to this. I, the government's doing this, it's not to help out the little guy, especially this government, the government's putting this mandate out.

It's to help the big guy, I'm sure just become, 'cause of all the complaining he's done down here, that Musk has really gotten upset about having to comply to air quality and blah, blah, blah, blah, blah, blah in Tennessee. And the same thing is happening here. So the, if you watch some of the, it's not just in Texas, but if you watch some of the protests against the data centers, people are getting more organized and realizing what it's gonna cost from a health and safety perspective.

So I, I feel like the government's doing this to protect the big data centers and protect what they're building and say, you have to be private now and then they can't touch you. You're all good to go. I think that's, I think it's much more sinister than, oh, this is great.

They're trying to think of how to figure out the grit. No, they ain't, they're trying to protect the big guy. I'm gonna, I'm gonna end this conversation here, but I'm just gonna start a rumor.

You know, John, what do you say? Open ar It's acquired by ge 'cause they need power. It's a, it's a symbiotic relationship.

Oh, wait, can I just mention one quick thing? So there's a commercial real estate project, uh, uh, research paper that came out and before they always talked about location, location, location. Now there's this idea that it's more about location, resilience and reliability.

Yeah, For sure, For sure. We, we live in crazy and interesting times, but let's just keep this subject going because, well, it turns out that Trump and the government is also has something to say about digital sovereignty. If you have not been following this conversation, it kind of accelerated a little bit with this tension over Greenland.

And before that, the terrorists themselves, but a lot of countries around the world are encouraging their, uh, companies within them to build sovereign clouds. These are clouds where the entire IT stack exists within the borders of that country and is managed by the folks who live in that country. Now, we've seen some US companies respond to that.

I think AWS has a service along with Google that's entirely hosted in Europe. Um, and we're probably gonna see this phenomenon elsewhere. But now the United States government apparently has told all its ambassadors to have a word with all those leaders of the various countries about not isolating US companies from those clouds, which kind of defeats the purpose from their minds.

Gina. I mean, this is like a tangled web. We're weaving here.

But if you're just a, you know, a poor old IT person in the middle of this, what, what are you making about all this? Oh, it's a mess. Uh, so it does well, so I'll start by saying, so Marco Rubio is the one that is the Secretary of State said, Hey, you gotta go and talk to all the people, especially the GDPR people in the, in, in, in Europe, and tell 'em this is not good for ai, it's not good for world peace or whatever they're gonna tell 'em.

That will be interesting to see. Um, so you've gotta think about why at the end, the data, the data sets that they have, the data sets they expect to get, and what they wanna do with the data. This administration in particular, um, because we've worked on this for, geez, Fred, 15 years, 20 years, it's been a long time, um, because they've all, we've always had satellite data centers in Europe 'cause of the distance and, and all of that.

The cloud kind of changed that and made it easier to get things over oceans into different continents. But, um, as people started to be able to take data and manipulate it, um, they, their laws came up in Europe that said, Hey, we are, we have a really strict privacy laws and we're not gonna send our data to be, um, to be, to be computed anymore in the United States because we don't trust you guys. Because there's no rules against, you don't have the strong rules we have.

And so that's how they're protecting their, their, their, their citizens. Mm-hmm. Um, it's not new to anybody who's been in privacy, who anybody's been in data centers, backup anything to understand what GDPR is and the protections that Europe has.

What's new is saying that we can't do AI unless we have the entire world's data. Why not? So, like, that's the thing.

So this is more of a str i i if they're, if if the government's getting the strategery around, you know, how they're going to control of the AI models, what step down means that they can just take all of the data from the entire world and process it and do what they need to do with it, perhaps put it into palantir's huge databases to help ICE do things. So it's, it's the whole, you know, back from, you know, after nine 11 when they did that power grab and they did, uh, wanted to know what all the information was, put all the information together, um, which is for security is an important thing, and we need to know that for our national security. But that doesn't, to me, give us the rights to do everything else.

Um, one thing I'll mention is I wrote a post, um, on Facebook about, uh, not on Facebook, I'm sorry, on LinkedIn, Facebook about, no, no, Facebook, um, about, um, sovereign constellations. So not only are we worried now about the data centers that are on Graham, governments are spending billions of dollars. It's projected that we, the United States is gonna pay, you know, up, we're already by 2030, we're gonna spend about a hundred billion dollars to get our sovereign constellation over the, you know, over the earth.

Um, collecting data, protecting intelligence, and it'll be positioning, timing, navigation, intelligence, surveillance, reconnaissance. Those are the kind of things, but other China's doing it. Europe's government is doing it so that they have their own sovereign cloud, that that data can't be taken by anybody else like our country or China.

So, and even we, XITE Labs came to Tech Field Day and they've got a contract for their networking equipment to go into all of, um, uh, of starlink V three. So it's really cool to me, from an infrastructure per perspective, like that's through edge. The edge of space is really edge.

So that's also coming into play. So what all is it that Rubio thinks he needs access to, and why does he think, I mean, the sovereign clouds are called that because we are a sovereign nation. We get to make our not own laws about who's a citizen, who's not a citizen, all of that, because we are a sovereign nation just like everybody else is.

So why should we be telling everybody else how to run their nations? I'm not sure that AI is at the root cause of this conversation. To me, the, I I think it's, it's part of the issue, but underneath it all is this US Cloud Act, right?

And the Europeans are basically, you put this cloud act together that says, and you can turn off any service that has US equipment in it. So basically, you know, after looking at what Trump was up to, the Europeans concluded, wow, these guys are pretty serious about this kind of thing. And so they're gonna try to take all that stuff out of the equation.

What people haven't reconciled with is that, well, that makes running it a lot more expensive, right? I gotta build my own stack, I gotta put my own stuff in there. And it kind of defeats the whole purpose of the cloud and many of these software as a service applications that we built.

'cause now they all have to suck up the fact that they need to deploy additional copies of software and run it in various countries. And I guess they're trying to make a claim that says, well, you know, this is bad for US business, but ultimately, Fred, I'll toss this your way. Why don't we just modify the US Cloud Act a little bit and everybody goes home to their neutral corners and calls it even.

Uh, that's a great question. I, I think this is what I would consider to be significant government overreach. Uh, and it looks, uh, it's not just obvious, it looks desperate.

Uh mm-hmm. We don't need, uh, Marco Rubio wandering the earth talking about what impacts this particularly has, uh, you know, as Gina said, there's a, there's a whole space set of, of things in here that's far above what the EU has access to. We've long done a number of things to help, uh, foment the discord with, you know, public cloud providers in the us.

But here's what I would say, um, the EU has also always had a stance that privacy and sovereignty belong to the countries within the eu. And it's not just the eu, right? You know, Singapore is pretty rigorous privacy policies.

A lot of other places in Asia do too, China does as well. And that every, every CISO on the planet knows the privacy policy in almost every country because of that, including the transit privacy requirements for moving data across states in the eu, not just over to the us. And the fundamentals here are this, uh, the EU passed an AI privacy bill, uh, the EU AI Act last year, right?

It's implemented. And, uh, in a lot of cases that creates, you know, some issues for US companies. The best arbiter for US companies in this case are the US companies.

So, because that's not federal policy, I, I really do take some issue with that. Microsoft is gonna build 287 data centers in the eu. They know how to negotiate out how to do the right things.

And look, that policy exists for a great reason because governmental over stuff is something that the eu, uh, we have no privacy in the us but they've long mandated that we maintain that in the eu. And I think that is a completely justifiable position in humanity. And, you know, companies are, if they're willing to do the business, then they've gotta pay the tax.

It's a cost of doing business. Well, let me tell you how Microsoft and SAP are working this situation out. 'cause it's pretty clever.

Um, so SAP has a whole separate business unit that's essentially a IT service provider. It's a managed service. And they have a contract with Microsoft that says, in the event that there is any disruption of service, we will shift the ownership of the Microsoft software that's being installed in Europe over to the paper that belongs to this IT service provider.

So Microsoft continues to deliver the service, but if is there any interruption whatsoever, this little, uh, IT service provider owned by SAP is prepared to step in and continue to provide these services in Germany and France, and I don't know where else. Um, and I think what you're gonna see is, you know, a lot of these deals are gonna get structured similarly to that, where it's gonna be basically, you know, in the event that there's an issue, you can have your sovereignty, but in the meantime, you can still take advantage of the lower cost services that are delivered directly through Microsoft. So Gina, can we just like, you know, put up some clever paper into a contract and, you know, work our way around these somewhat well-intentioned laws, but maybe not so well Bottled?

Are they well intentioned? I'm not sure. So, but, but no, I think, yeah, that's what every, that's what tech has always done.

They write contracts around it and find the right people to do it. I'll put this question back out to you though. How is this something different than TikTok?

TikTok? I'm not on TikTok for privacy reasons, but mm-hmm. The whole TikTok deal, even though it got weird, the president has the authority to say, Nope, they can't work, they can't be in the country because of, uh, you know, privacy issues.

So we're gonna take our sovereign protection seriously, and we're not gonna allow any traffic into the United States from this particular service. So how is this any different? I don't think it is much different.

John. Oh, you know, I was gonna say this, like, this show kind of makes me think of government overreach to, to the extent where this document that Rubio is, is sending out to the di the diplomats, it almost reads like it was written by big tech. You know, they, one of the thing that just stood out to me, and it's just kind of a little bit borderline appalling, is they described the GDPR as unnecessarily burdensome.

And I mean, it's, these countries in Europe created this to protect privacy and national security. They're very keenly aware of that for a lot of reasons going back, you know, historic reasons. And the US is basically arguing just the opposite in this, this whole case.

And, and again, it just underscores. And Gina said how the big tech and, and big government are, it's almost synonymous to, to me now, um, I can't separate the two. And it's, it's be, it's almost become part of this constituency or part of, part of a shadow government or a, the department of big tech.

You know, the more I think about it, the more I look about it from every angle and, um, someone's gonna pay a price. I like that Department of tech. Maybe that's a whole new cabinet position.

We'll see where that goes. Let's go be in charge. So I, I can't remember, I've done product marketing for 10 years.

I can't remember a product we've had that we didn't have a solution for GDPR, so we could sell it in Europe. Like we all knew what it was. I mean, this is going back since before it was formally enacted and then strengthened.

So this is just, you know, I don't get it. It's just like a little horse training here. I can see like everybody's sitting around the table and they go, you modify your little US cloud acting, we'll modify our little GDPR laws and maybe we'll come to you in the middle.

Hundred percent politics. I think the, and then the interesting, uh, parts, the placards that are hung in offices are gonna be the details in this. And we talk about things like biometric categorization, uh, in Europe and how we think about identity.

And there's lots of really interesting and important things. And just to cast the focal point from earlier in the first segment, we're talking about agents on laptops, right? We're, it's not just data centers, it's not just the notion of my phone.

It's everything everywhere. And, you know, again, if, you know, philanthropic is going to enable, you know, remote sessions, uh, you know, where, where those remote sessions, I travel to Europe, right? And my subject to, you know, privacy laws, uh, related to AI, laws related to my laptop and anybody else that I interact with, and the adjacencies of all of that.

Plus there's a backdraft here on how do you remove, uh, information used for model training, model analysis, you know, small models, private models, frontier labs, models, whatever. Uh, and the lineage associated with it. It's a very, very tangled web and I think it's really going to be difficult to regulate this.

Uh, so this is probably the first real volley of this red line that we'll have to turn into some sort of demilitarized zone of what we're willing to accept. And I agree it's gonna be a lot of horse trading, Mike. All right.

Well, folks, hey, mark Rubio, if you're listening, the Techron Gang is prepared to go on a diplomatic mission and sort this out. Let us know where you want us to go and we'll take care of the whole, the details. Hey, I wanna thank everybody for sharing their knowledge and insights.

As always, folks, good to see the gang members. Please stay tuned for the, um, second edition of the techron TV lineup coming up right behind us. You can, as always, is fascinating and awesome conversations that you should definitely not wanna miss.

And we'll see you all again tomorrow.