Anthropic Files Its Lawsuit. Can it Win? - DTNS 5221

0:00This is the Daily Tech News for Monday, March 9th, 2026. We tell you what you need to know. Give you some important context. And today, we're really going to try to help you understand today. And Thropic found his lawsuit against the US calling it a supply chain risk. We break down the case and we're going to discuss it and let you know what's going on. Yeah. We're going to try to figure out if they have a chance of winning it. Marit. And I've done wood. Let's start with what you need to know with that big story. Thanks to Fred, who noted our big story on the DTS subreddit and Thropic filed two lawsuits on Monday against the United States government, one in California federal court, and another in Washington, DC's federal appeals court, I could not find the actual text of that one. Importantly, it challenges whether and Thropic can rightfully be considered an adversary,

1:00as is required by the definition of supply chain risk in 10 USC 3252. That's the law being used to justify the designation of Anthropic as a supply chain risk. The lawsuit filed in California, however, I did get a chance to look at the text of that one disputes the designation on five separate grounds. And the first one is one I talked about as a possibility on the weekend edition of DTS. If you haven't listened to that, I go into detail on the actual law, USC 3252. It is unclear that the Department of Defense followed the procedures of that law. They required that the department write down its reasoning about why the designation is necessary, why, why there is not a less intrusive measure that could do the same thing and deliver that to Congress. Now they've written it down in Senate 10thropic, but it's not clear they sent it to Congress. So Anthropic is saying they didn't follow the procedures of the law. That's point one. Anthropic also says that the designation is an attempt to suppress speech, and therefore

2:03a violation of the First Amendment, that the direction of agencies outside the scope of that particular law, overstep presidential authority, that the designation is a violation of due process and that it negatively impacts its business and therefore is an illegal sanction. The notification of Congress could be easily fixed. They could just write it down and send it to Congress. Then it would be on Congress to review that designation. It probably wouldn't do it. It probably would just let it sit there because it is a Republican controlled Congress. So if the Department of Defense makes that happen, then maybe that provision goes away. So the case in California probably rests mostly on the other provisions. All the analysis I'm seeing is focusing on that First Amendment as its best bet. Anthropic will likely use the fact that the government is using its tools and has given agency six months to wind down their use as evidence that it's not an urgent supply chain

3:03risk. Wall Street Journal has an in-depth article on the use of AI for the US war against Iran, which details Palantir's tools being used for non combat needs, including intelligence, mission planning and logistics. And Palantir is the one that has been using Anthropics Cloud models the most. So so Rob, I feel like there is a chance depending on what happens with filings next that they can make a case. The one thing the Department of Defense could do is prohibit judicial review, but they'd have to follow all the the eyes and teeth of the original law to do that. Yeah. The thing for me is that the government is saying that this is not suitable for use, you know, here by anybody who's doing business with us, yet they're still using it. It seems to me like if there if there was a real concern about this software being a real supply chain risk, they would have hit the red button in the data center to just shut

4:04Anthropic down immediately. Yeah. And that's not as possible. That's not happening. So the question becomes and when you start to think about it is that, okay, so so what has changed about the software? Anthropics saying nothing has changed about it. What has changed is the relationship with the federal government, with, you know, with the Department of Defense. We, you know, we broke down and deals and now it feels punitive that because our deal broke down that we can no longer do business with anyone, let alone the Department of Defense. And I'm not a lawyer. I don't play one on the internet, but I do wonder how that is going to play in front of a court. I think you're right. If they send something to Congress, I think that it'll just be sent to Congress and nothing will happen. But I do wonder what's going to happen in front of a court because this is costing Anthropic dearly to not be able to do business in the way that they've been doing business. There are other vendors that we just can't take the risk of working with you right now. We don't want to get on the bad side of the DOD. So there's a lot of things that, to me, are adversely affecting Anthropic and it looks

5:10like, you know, at least from the service room we've been hearing about is just that we, we decided to go a different route and we don't want anybody to work with you anymore. That's what it seems like in this particular case. Yeah. And Anthropic is making up some of its ground on the consumer side. The, the clawed model has become one of the most popular on the app stores, but it takes a lot more consumer revenue to replace enterprise revenue. So this is still hurting them. A couple of things I think are worth reminding folks of. One, one is that the dispute was that the Department of Defense said we want to simplify this and say we can use this for anything that's legal. And Anthropic said we're fine with that, except we would still like to hold back in cases of autonomous lethality and mass surveillance. And those cases we, we want to have a little bit of a check on that. Anthropic was not even saying you could never use it for them. They were saying, especially with autonomous, it's not good enough yet. I, we'd like to hold off on that until it gets good enough. And on the mass surveillance stuff, they, they were holding a little more of a line.

6:11And then this blew out of proportion as we've seen over the last couple of weeks. So it's not like Anthropic and the Department of Defense were that far apart. And I think that's what open AI was trying to say open AI, no friend of Anthropic, but open AI was trying to say you're not that far apart. Look, we can do a deal that preserves your concerns. And you should be able to deal that, do that deal too. And they have not been able to get those two sides to do that. The other thing I think is worth pointing out is the Department of Defense does not have to declare Anthropic supply chain risk to stop using them. In fact, they have a better case if they don't declare it a supply chain risk because the Department of Defense can just choose not to use whoever they want. The US government can even say, you know what, nobody use Anthropic because we just don't want to use them. And there would be little that Anthropic could sue about in that case that they went over the top to say, you're a supply chain risk. So now no one who does business with us can use Anthropic stuff in any way related to their business with the US government. That's just punitive.

7:12And that's why Anthropic is filing these lawsuits and that why they have a chance of succeeding and overturning this. I suspect that maybe the Department of Defense doesn't care and says, yeah, DOJ go fight that. Drag it out as long as you can if we lose it, who cares? We achieved our aim. We wanted to punish Anthropic and we wanted to have a chilling effect so that companies who use Anthropic pull it out, whether they have to or not. Yeah, that is that is the kicker here. This feels very punitive into into state your earlier point even more. Anthropic was actively working on doing these things. They just said that we're just not there yet. It's not that we don't want to allow you to do this stuff. We just don't feel that our software is capable of doing this stuff yet. So we want to hold back on it. And so when you think of it like that, it's like, you know, what has really happened? What did they put in the software that makes it now a supply chain risk? I don't think anything was put in the software. I think this, the filling about Anthropic changed and we're going to go nuclear option.

8:14We're going to make it so that we, you can't do business with anybody that does business with us. And, you know, as many people that are signing up Anthropic accounts and even cancelling open AI accounts, that's real nice. But there's no company that is in the talks of getting a DOD contract that wants to be out of the talks of getting a DOD contract and have your company taint because the DOD says, no one else who works with us can work with you. That's something different, you know, contracts are lost all the time. Companies win, companies lose. To be told that you can no longer do business, not just with us, but with anybody who does business with us, that is, that is, that is very, very different than what we've seen in the past. Yeah. And, and it is not necessary. I think the DOD as well within its rights to say, you, you want to hold the line on mass surveillance and lethality, we're going to go a different direction. We're going to find somebody else. That would have been fine. But declaring it a supply chain risk was not necessary to do that. And as we're going to see andthropic can continue to do business with companies that are doing

9:18business in, in areas that don't include the Department of Defense, you know, including Microsoft, which has said, yeah, as long as it's not in a project that we are using with the Department of Defense, we're going to continue to use cloud. DTNS has made possible by you, the listener, thanks to Matt Zaglan, Jeff Wilkes, Tim Deppity, and our new Patrons Sabrina Ron Luke and Estesie, we've got returning Patrons bow as and fallen in sea and we've got a raise by Jordan. It's been, it's been, it's been a busy week, excellent user names and excellent support of the show. Thanks, everybody. Estesie is top notch. That is a top notch name, right? And it's spelled ESDC for those who are, all right, there's more we need to know today. Let's get to the briefs. RwNAS let everyone on our subreddit know about this next one, Microsoft is using Anthropics Cloud co-work to power co-pilot co-work on the Microsoft 365 co-pilot AI platform, co-pilot co-work can build presentations, create spreadsheets, and set up meetings within your company.

10:23Cloud co-works launch on January 30th, cause concern for the future of companies like Salesforce and Service now, Microsoft is bringing similar agentic capabilities to word, Excel, PowerPoint, and Outlook as well as improving co-pilot chat. The company also launched its monitoring and governance platform, Microsoft agent 365 and the general availability for $15 per user per month. You can get all these tools together in the Microsoft 365 e7 suite for $99 per user per month. Yeah, so I think this is a big deal just because it's Microsoft getting into this area that caused some stock to tank at the end of January. I don't think this is going to cause the stock to tank because I think people have kind of reset their expectations of what these things can do, but it is the kind of thing that will not only further clause business because this is something Microsoft can continue to do and not worry about the Department of Defense stuff, but also it's going to make co-pilot better and it's going to bring some of these agentic capabilities into the 365 platform

11:24in ways that I think are more acceptable to enterprise admins. I can absolutely see why these big, giant software of the service type companies like Salesforce and ServiceNow that they would be in a bit of a, you know, a curfuffle right now that they've been in the middle of panic because when you think about what the promise of a agentic AI can do, it targets SaaS. You don't have to go build a SaaS solution to do these things to enter this data in to format it this way, to scrub it, to put it here to make it available to all these things that Salesforce and ServiceNow actually do. You literally just say, hey, I want this data to go from here to there and I want it to look like this and I want it to do that to your AI and it just creates it. And is it there yet? No. You know, I think there may have been an over, you know, reaction in the market and they've kind of come back from it, but that is ultimately where I do think agentic AI is going. Eventually, you won't necessarily need a service now. You won't need a HubSpot.

12:25You won't need a Salesforce because you can just say, hey, AI, do this thing with this data in this way. And it says, okay, cool. And it just does it. Yeah. The smart thing that Salesforce and ServiceNow, I think both have done is incorporate these tools into their side of things so that their future proofing themselves to say, yes, once the stuff that we sell the most of now is not needed anymore, hopefully we've developed some other tools that will be needed, right? So I think that's part of the correction you were talking about as people realized, oh, Salesforce is continuing to offer new things, right? It's not just going to stay the way it was five years ago for the rest of its life. Yeah. And the promise is not even close. It's not there yet. You're not replacing Salesforce with agentic AI. Not today. Not today. They've got some time to adapt. One from Harvard Business Review has conducted a study of 1,488 full-time US workers asking them how often they use LLMs and other generative tools in their work and how they felt it affected

13:29their thinking and feelings. They found, and there's some ways to measure burnout and mental fatigue that the survey was using as a metric, they found that burnout reduced when people used these kinds of tools. Now, mental fatigue increased and you may say, well, what's the difference? Mental fatigue is, I'm enjoying what I'm doing, but now I'm feeling slowed down. I feel a little confused. I need to take a moment. I'm not able to do as many things as possible. It's the myth of multitasking. When you multitask too much, you get mental fatigue. Burnout is, I can't do this anymore. Too much mental fatigue can lead to burnout, but what they're finding is the LLMs actually reduced the burnout. I don't feel like I need to quit anymore, but the mental fatigue that confusion increased. The most mentally fatiguing use of these tools was the increase need to monitor performance. In other words, have to check and see. Did you do it? Did you do it right? Are you doing it? Another factor was that use of the tools increased your workload, which increased the number

14:31of outcomes you needed to pay attention to. Well, now I've got 10 projects instead of two, because I can do 10 projects, but I had to pay attention to 10 projects. Study also found that using two tools simultaneously did increase your productivity, adding a third tool at once. You're using three of these at once, increased it, but at a slower rate. However, if you did more than three, it started to decrease your productivity. There's a classic bell curve going on. If you use too many of these tools at once, you are not being more productive. Three is probably the max. Use of the tools to replace repetitive tasks, however, decreased burnout. The authors were like, this makes sense. If the tedious stuff gets taken away and you don't have to monitor it because it's pretty repetitive, it's pretty automatic, you're not going to get burned out because you're not doing the stuff you don't enjoy as much. Highest incidents of mental fatigue occurred in users in marketing, HR operations, and engineering. Lowest was in legal management, product management, consulting, and customer service. The survey found that managers who take time to answer questions about these tools and

15:36provide a clear strategy of why you're using them saw employees with less mental fatigue. Whereas employees who felt pressured to use the tools at any cost felt more mental fatigue. So what I like about this Rob is that we're starting to see an outline of, ah, this is how to responsibly use the tools in the workplace to actually increase productivity and increase employee welfare. Yes. So what's interesting? We actually had a conversation last week, last Thursday, on DTNS Live Hangout with Nika Monfer, where a lot of what's happening with people in using AI is not, they are becoming a little bit more productive, but the workloads they're getting are, you know, increasing, which is probably leading to the mental fatigue, it's work that I can do, but I've got so much more of it is just like, hey, you know, can you slow some of this down? So that's really interesting. I think companies are going to have to ultimately learn how to balance the two. We can make you more productive, but making you more productive doesn't mean giving you

16:37necessarily more work. It may mean allowing you to work on other things that you enjoy doing at a higher production rate. And there is, you know, there's a tipping scale there, you know, one tool, two tools. That's okay. Three tools, maybe not so okay. I think companies are going to have to massage this for a while and figure out where they are with their teams. And as you said, one of the, you know, one of the keys here is that if just a manager talks to you about the tools that you're using, that can actually lower stress. So I think that there's a lot of good information that comes out of this and companies over the next few months, over the next few years, over the next, you know, the remainder of the decade or want to figure out how do we actually get the productivity that these tools promise while keeping the people who are using the tools happy and employed successful? Yeah. It's the same principle behind not having too many direct reports, right? A manager with too many direct reports can't effectively manage all of their employees. These are essentially direct reports, right? You got too many projects going on. You can't effectively manage them. So for the first time, I think companies have to look at the attention resource of an employee

17:42and say like, well, how much are we overtaxing their ability to pay attention to stuff? I'll bang on this again. You definitely need to measure the quality of the output of your employees, not how many times they're using an LLM because they'll just game the system and use the LLM a bunch. They'll use it wrong. They won't know why they're using it. They'll get mental fatigue. They won't be more productive because they'll be like, well, the more I use it, the better. So I'm going to use five of these at once and turns out that's going to lower their productivity. It's all right there. Like we're seeing it. You need to tell them why they're using it, answer their questions, have a clear strategy, and measure them on how well they do their job, not how many times they put in a query. Yeah, I think the companies eventually will be the winners in this area. Won't be the ones who are looking at if we can make them more productive. We can have fewer of them. We're seeing over and over again. That's not necessarily the way to look at it. If we can make you more productive doing the things you like to do,

18:43we can allow you to do more of those things. Yeah. And then the other people who like to do the other things allow them to do more of those things. And then by everybody really liking what they do and being more productive at the things that they like, the company overall does better. And I think that's where companies really need to get to as compared to just looking productivity. We can have less people do the same amount of work. That is probably not the mathematics that you want to apply here. People doing the right work with these tools that they enjoy and then bringing up that load of the stuff that they don't necessarily enjoy to the people who enjoy. Yeah. And then you improve the overall product of the company, whatever it is. Absolutely. Local government agencies in China have been done encouraging the use of open source, open claw platforms leading to renewed optimism in software companies. A shinson district is seeking feedback on a policy to encourage free open claw services. Tencent, for example, already offers free installation of open claw on his cloud platforms. China's ministry of industry and information technology still cautions that deployments

19:46can lead to security risk if not properly configured. I can't decide what my take on this story is, honestly, is it like, ooh, China willing to take a risk that maybe a lot of US companies won't because, you know, and try to jump ahead by using an open source or what a great Trojan horse the US has sent to China, giving them open claw, which is, you know, difficult to use responsibly and full of security holes that could then be exploited. I think because we look at this ginormous country, the biggest country on earth, we just look at it is just China. There's a lot of folks over there. And I think what's happening is that, hey, you know what, this looks kind of cool. Let's take a look at it and see. Let's figure out if there's, if there is, is there a there there seems to be what this is. So of course, when that happens, we're going to have significant overreaction to it. So, oh, why, why are they looking at this thing that we actually built that we tried to, you know, have be more beneficial to us to? Yeah. Well, folks, you got to thought on this.

20:49What do you want to hear us talk about on the show? You heard a couple of people who have let us know about stories today. You could be one of those people too to make your stories and vote on them at reddit.com slash r slash daily tech news show. All right, let's get into those quick headlines that we give you that'll make you look smart. Google was added a preference to Google messages called tap to draft that changes smart applies from instantly sending or creating a draft that you can review first. I personally would be a fan of this. Yes, no, turn that setting on it like the thing where you like want to take the suggestion and it's like, great, I sent it. No, not a good look. Oppo announced the Find N6 coming March 17th will use a titanium alloy hinge with a memory glass layer that will make the crease in its foldable difficult to see. In fact, some people call it zero appearance as if it's not there. We need to see it to believe it. Yep, but credible people are saying you can't tell there's a hinge there.

21:50And that that is that's some next level stuff. Because even when you're looking at some of the, you know, the phones that we know when you're looking dead on them, you can still see it. If they're saying you can't see it, that's that's something different. ABB robotics is integrating Nvidia's omniver simulation platform into his robot training software for autonomous industrial robots first being tested by Foxconn. Oh, good for Nvidia. Good for Apple, I guess. Keep an eye on the dispute between chip factory and experience China operations and its parent company in the Netherlands has been going on for a while now. The latest is that China's commerce ministry ministry has started warning that a fresh dispute between the two sides, which has been heating up again, could disrupt chip supplies even further. And if it did, it would be the Dutch side's fault. So they're already pointing fingers. After 11 years of Tesla, Thomas Demetrick, the man who oversees software for Tesla's robot taxi service is leaving true prior to his family. I'd love to see how this goes because it means one of two things.

22:50He's leaving true prior to his family or he's leaving for some other reasons. Could be some other reason. And he's not the first. Oh, there's one of the bigger ones to leave recently. After any Apple announcement comes the rumors of the next Apple announcement. And among them are some obvious ones, foldable devices and such. But I found one from Bloomberg's Mark Gurman, more notable. Apple is working on 3D printing aluminum to make cases for phones and watches more efficiently, aka more cheaply. Interesting. Researchers at the Granger College of Engineering at the University of Illinois have determined how to develop 2D magnetic systems that follow the same equations as mobile electrons and 2D graphene. They believe the finding could help make smaller microwave systems for wireless and cellular networks. Oh, that would be that would be interesting to see. Magnets is graphing. Linux 7.0 release candidate 3 is one of the biggest changes in recent history. And a lot of people are like, that's not necessarily a good thing because there are way more bug reports, Linus Torvalds, saying the bugs aren't scary, but the number, the volume of them

23:53worries him. UK data center developer in scale has added a few familiar names to its board, including former Med execs Cheryl Sandberg and Nick Clegg. Oh, yeah, okay. I know those names. And finally, TechCrunch's Zach Whitaker has done a bang up job collecting all the info about the ongoing salt typhoon hacks. They don't, they don't always make the headlines, but they're persistent out there. These are attacks usually against phone and internet providers worth your time to read and catch up if you want to know the latest. Well, folks, we in every episode of DTN, that's some shared perspectives. Today, Kevin is a thought on the whole and throbbing government contracts thing. Yeah, Kevin writes, I appreciated your coverage of the recent and throbbing versus open AI versus the government story. One thing that I think it's lost in the outrage is the question of, who would you rather have do this? Seems obvious that the administration will find a willing vendor in the space eventually. Would you rather have one that is at least publicly saying they're trying to do the right thing

24:54or somebody who is an unknown entity that doesn't even claim to have any morals? Yeah, that's a that's a thought. I think from the government standpoint, we want to work with whoever's going to do whatever we tell them to do. That seems to be what's happening here because in throbbing and opening AI really are not, I won't even say that that far apart. I think they're both thinking the same thing about this is just that in throbbing was a company that the government got mad at first and now they're going through what they're going through. Yeah, I think what Kevin's pointing out is people getting mad at open AI for stepping in and signing a contract and saying, well, we don't believe you that you want to do the same things. You've obviously compromised. And I think Kevin's saying, well, okay, maybe, but would you rather open AI sign the deal and not to say that they want to they want to keep those red lines? So I think it's a fair point. And I think a lot of folks too, they just have a problem with open AI signing a deal with the government. And as I said before, you do not want to not do business

25:55with the the federal government. You know, we're talking specifically about the Department of Defense, but ultimately we're talking about the federal government. You want to do business with the biggest employer of people in the United States. You just you just want to do that. And a lot of people weren't paying attention or didn't mind that in thropic was doing both the lion share of the business in AI with the Department of Defense with the US military until now. So for many, for many months, many, many months. Yeah. Well folks, we want to know what you're thinking about. So if you've got some thoughts or insights into a story, please, please, please share it with us at feedback at dailytechnewshow.com. Yeah, big thanks Kevin for contributing to today's show. Thank you for being along for daily tech news show. You folks keep us in business. You can support us by becoming a patron. If you're not already, patreon.com slash dtns. Talk to you soon.

26:55The dtns family of podcasts helping each other understand time and club hopes you have enjoyed this movie.