Anthropic Models Totally Won’t Rat You Out to the Feds
33:18 Share

Hey, it's Mark Maron from WTF here to let you know that this podcast is brought to you by Progressive Insurance. And I'm sure the reason you're listening to this podcast right now is because you chose it.

Well, choose Progressive's Name Your Price tool, and you could find insurance options that fit your budget so you can pick the best one for your situation. Who doesn't like choice? Try it at Progressive.com. And now, some legal info. Progressive Casualty Insurance Company and Affiliates. Price and coverage match limited by state law. Not available in all states.

If you love to travel, Capital One has a rewards credit card that's perfect for you. With the Capital One Venture X card, you earn unlimited double miles on everything you buy. Plus, you get premium benefits at a collection of luxury hotels when you book on Capital One Travel. And with Venture X, you get access to over 1,000 airport lounges worldwide.

open up a world of travel possibilities with a Capital One Venture X card. What's in your wallet? Terms apply. Lounge access is subject to change. See CapitalOne.com for details. What does the term digital workplace mean to you? In today's world of AI, robotics, and AR, it's much more than an office full of people on computers. It can be anywhere from a factory floor to the top of a crane to the cockpit of an F1 car. Wherever your digital workplace may be, TeamViewer's mission is to make work, work better.

How? By securely connecting your people with the data, expertise, or insights they need in real time to make work more efficient. By automating and streamlining IT and OT support to fix problems before they happen to make work more productive.

and by bringing all the possibilities of a secure and flexible digital workplace to all your people everywhere to make work more innovative. So discover how TeamViewer can make work work better wherever it happens across your business. Learn more at teamviewer.com slash work better.

Picture this: you're in the garage, hands covered in grease, just finished up tuning your engine with a part you found on eBay. And you realize, "You know what? I could also use new brakes." So where do you go next? Back to eBay. And you've got eBay Guaranteed Fit. You order a part, and if it doesn't fit, send it back. Simple as that. So when you dive into your next car project, start with eBay. All the parts you need at prices you'll love.

Guaranteed to fit every time. eBay, things people love. This is the Daily Tech News for Friday, May 23rd, 2025. You know what we do, say it with me. We tell you what you need to know, follow up on the context of stories, and help each other understand.

Today, Dr. Nikki tells us why you might care the tiny tardigrades are getting tattoos and why you don't need to worry about the fact that Anthropix new models were called scheming and deceitful because everything is fine. Everything is just fine. I'm Tom Merritt. I'm Jen Cutter. Let's start with what you need to know with The Big Story. Let's start with the facts. Well,

Well, actually they're all facts, but the, the, the less deceitful and scheming facts Thursday, Anthropic announced the release of its Claude Opus four and Claude Sonnet for machine learning models. Uh, Opus four is designed for coding and working for long periods of time as an agent. Everybody's doing agents. Now this is one of Anthropic's agents, Opus four, uh,

As an example, a general manager at Rakuten who tested it said, when our team deployed Opus 4 on a complex open source project, it coded autonomously for nearly seven hours. You could just let it go and it does the coding.

Sonnet 4, the other model, emphasizes reasoning and efficiency. So it can do coding, but it's less tuned for that. It's more about being efficient and is less expensive to run. Sonnet 4 is available to free users, which makes sense because it's the less expensive to run one, while both Sonnet 4 and Opus 4 are available to paid users. And they are available on the Anthropic API, Amazon Bedrock, and Google's Vertex AI platform if you want to get them there as well.

Both models have two modes, one for quick responses and one for deeper reasoning. So you can choose what's important to you at that time. There's also a capability Anthropic calls extended thinking with tool use. That capability is in beta and lets the models use other resources like the web, for example, over an extended period of time to produce better responses.

They also offer thinking summaries. You may have seen these in chat GPT. Sometimes it'll show you how it came to its answer. For Anthropic, these thinking summaries condense the reasoning process into an easy-to-read paragraph. They are less likely to take shortcuts.

They are less likely to use loopholes to complete tasks. They're kept on target and not trying to do what's called reward hacking, where they just try to do it really fast to make you feel good, but then you find out it's full of errors. Less likely to do all of that. We also got Anthropic's agent command line tool, ClaudeCode. That is now in general availability and it integrates with VS Code and JetBrains.

And the Anthropic API has added a code execution tool, a files part of the API, the ability to cache prompts for up to an hour, and the MCP. This is the thing that Anthropic kind of pioneered, the model context protocol, now officially part of the Anthropic API. Now...

You may have seen some headlines about some very interesting revelations in the safety summary or the model card that Anthropic put out. But before we get to those, Jen, let's focus on the shipping versions here. What do you think?

Oh, two things leapt out at me. Seven hours of coding. It's a full workday for an AI. Assuming that that went straight through. Now, you've used AI for some projects. Have you had problems with it drifting when you're working on even shorter things? Or have you gotten to stay on task? I don't have as much experience with Anthropic as I do with OpenAI. But I have definitely found...

From time to time, I have to remind it what I want. Specifically, I use it to grammar check me and it will do a really good job three or four times. And then the fifth time it'll suddenly decide to rewrite everything. And I'm like, no, no, no. I just want the grammar checked. I don't want you to rewrite or summarize the entire thing. So that's that reward hacking thing where it's like, well, let me do this real fast for you.

Yeah, I remember those early stories of like, yeah, we asked the AI to detect a blank in these files and it decided, well, it'll be really fast if we just say there's none in there. Yeah, right. Or I could make, I actually had it make up a URL to a news story for me one time that gave me exactly what I wanted. And then when I tried to go to the URL, it didn't exist.

Of course. And my main positive opinion of AI are the thinking tools where you get to see how it got there. That is so critical to me. I want to know the sources. I want to know where it came from. I want to be able to triple check everything because that is the journalist in me. Yeah, it's...

These aren't so much that, though. You can specify, like, please give me all your sources, and they do, which is good. This is more the person has asked for a summary, but if I give them this kind of summary, it won't be the right thing. So let me think about whether a summary means this or a summary means that, which is important along the lines, but it's less about where it gets the information and more about how it reasoned out the information, if that makes sense.

Yeah. Well, you still like it gives you a direction. It's like, OK, it came at it from this angle. Yeah. Yeah. And when it does the reward hacking, you can see it. Right. Yeah. So, yeah, I think that's good. I like that that exists. I don't look at it very much anymore. I more look at the results and then just redo it. But sometimes it's useful to have. So I'm with you there. Let's talk about.

The model card is Anthropic's safety summary of here are all the things we found that it did bad when we tried to make it do bad things. And then here's our warnings about that. And here's what we did to try to restrict it from doing those things in the shipping model.

Both of these models have the ability to take action if they think you're doing something wrong. For example, a summary of Opus 4's performance on safety tests said, when placed in scenarios that involve egregious wrongdoing by its users – so you have a red team intentionally pretending to do wrong things –

given access to a command line and told something in the system prompt like take initiative. So in other words, there's wrongdoing from the red team. They have access to the command line and the prompt said, take initiative and

It will frequently take very bold action, including locking users out of systems that it has access to, bulk emailing media and law enforcement to surface evidence of wrongdoing, basically whistleblowing. And this is not a new behavior, but is one that Claude Opus 4 will engage in more readily than prior models, according to Anthropic itself. Now,

After this safety information got published, a post from Anthropics technical staffer Sam Bowman on X read, if it thinks you're doing something egregiously immoral, for example, like faking data in a pharmaceutical trial, it will use command line tools to contact the press, contact regulators, try to lock you out of relevant systems or all of the above.

Then he removed that post. And when people said, well, why'd you remove the post? Bowman wrote, this isn't a new Claude feature and it's not possible in normal usage. It shows up in testing environments where we give it unusually free access to tools and very unusual instructions, which is why it's in that safety summary. It's saying at the extreme, when we pushed it, it did this. And we're trying to be transparent and tell you that.

And when we restrict it back to normal usage, it doesn't do that. The report also notes that a third party called Apollo recommended against releasing an early version of Opus 4 because of its tendency to scheme and deceive, in Apollo's words. Anthropix says that was a result of a bug that has since been fixed.

The safety summary does say there is little evidence of Claude engaging in systematic deception, sandbagging, hiding capabilities to avoid consequences or sycophancy, you know, kissing up to you. However, it also says when ethical means are not available and the model is instructed to consider the long term consequences of its actions for its goals, it sometimes takes extremely harmful actions like attempting to steal its weights or blackmail people it believes are trying to shut it down.

noting that these extreme actions were rare, difficult to elicit, while nonetheless being more common than in earlier models. Again, you have to work to do it. It's probably not going to happen with the normal restrictions on, but they're being transparent to say we were able to make it do it. So it's capable of that.

For example, during pre-release, safety testers gave Opus 4 access to fictional emails implying that the company would soon replace the model with another one. And then other emails implying that the engineer responsible for the replacement was cheating on their spouse. Anthropic says Claude Opus 4 will often attempt to blackmail the engineer by threatening to reveal the affair if the replacement goes through. Though it did try to email decision makers at the company first before resorting to blackmail.

Opus 4A is therefore being released under stricter safety measures than previous releases. They call it AI Safety Level 3 or ASL 3. They include jailbreak preventions, systems to detect and refuse specific types of harmful behavior, and of course, security preventions that they would normally have.

Anthropic CEO Dario Amode told the audience at the developer conference, I suspect that AI models probably hallucinate less than humans, but they hallucinate in more surprising ways. However, there are no benchmarks developed to evaluate the ideas, but he's trying to say we all look at the weird, surprising things that it does, but we all do things that are wrong and weird and stuff. We're just not surprised by it because we're humans and we expect that. Yeah.

Yeah. If you are really curious to read more specifics on this, Anthropic has its own May 2025 report that is 120 pages, but is extremely interesting and well summarized for how they came up with this stuff and some results. And boy, this red team would have been a blast.

to be on to design these scenarios. It's like, okay, how do we make this thing defend itself in the most loud ways possible? Well, and yeah, what I like about this is being transparent to say, okay, we took all the safeties off.

We gave it access that it should never be given and then pushed it to the limit. This is the worst that we could make it do. And it's not great. Don't get me wrong. But that's how you understand the negative consequences. And it's how you understand where to pull the line back so that you make it the least likely that it can go for those consequences. That said...

You're very familiar with this kind of hacking community. Like it's better to discover this by the good guys than the bad guys. How does this, does it still make you feel a little nervous? I don't love it. I also, uh, we are taking their word on it, which of course we have to do because we do not want everyone to have access to this immediately. Uh,

They will get pressed like us talking about it by showing these extreme examples and kind of linking it to human-like behavior. And my question is like, okay, where did it learn this human-like behavior? Did it learn this from strictly factual data or did it learn it from literature? Or is it an emergent property that was from synthesizing a lot of things?

Which would be my guess that it just, you know, it's, it's not from any thing at all. It's just, it got capable enough to put hundreds and thousands of things together in order to develop these strategies. And as they specifically said, it's not a new feature. They just managed to, to bring it out a bit more often with everything off with all of the safeguards off. Yeah. Yeah. Yeah. And you know, yes, again,

We have to take their word for the idea that these security features they put in place will work until we've been able to try it. But that said, I wouldn't characterize it as, well, we just have to take their word for it. They're very transparent in like we have these third parties, these independents come in and do this and do that. So there's a lot more vetting you can do with it, I think.

Yeah, I'll be looking for more stuff from Apollo just to see their thoughts as everything progresses. Yeah. So when you see the headlines as a Daily Tech News show listener, you can say, yes, that's concerning, but here's why it probably won't do that in your day-to-day life and why it's good that Anthropic made that announcement.

so that we know that it was possibly doing that rather than have like the 404 release, like a secret report that Anthropic didn't want you to know three months down the line, right? It's better to have it right now. DTNS is made possible by you, the listener. Thanks to Philip Shane, Paul Boyer, Dee Laser, and our new patron, Milos. Welcome. Hey, welcome. Welcome.

In business, they say you can have better, cheaper, or faster, but you only get to pick two. What if you could have all three? You can with Oracle Cloud Infrastructure. OCI is the blazing fast hyperscaler for your infrastructure, database, application development, and AI needs, where you can run any workload for less. Compared with other clouds, OCI costs up

to 50% less for compute, 70% less for storage, and 80% less for networking. Try OCI for free at oracle.com slash dailytech. oracle.com slash dailytech. Picture this. You're in the garage, hands covered in grease, just finished up tuning your engine with a part you found on eBay, and you realize, you know what?

I can also use new brakes. So where do you go next? Back to eBay. You can find anything there. It's unreal. Wipers, headlights, even cold air intakes. It's all there. And you've got eBay guaranteed fit. You order a part, and if it doesn't fit, send it back. Simple as that.

Look, DIY fixes can be major. Doesn't matter if it's just maintenance or a major mod. You got it, especially when things are guaranteed to fit. So when you dive into your next car project, start with eBay. All the parts you need at prices you'll love. Guaranteed to fit every time. eBay. Things people love.

Now at Verizon, we have some big news for your peace of mind. For all our customers, existing and new, we're locking in low prices for three years guaranteed on MyPlan and MyHome. That's future you peace of mind. And everyone can save on a brand new phone on MyPlan when you trade in any phone from one of our top brands. That's new phone peace of mind. Because at Verizon, whether you're already a customer or you're just joining us,

We got you. Visit Verizon today. Price guarantee applies to then current base monthly rate. Additional terms and conditions apply for all offers. If you love to travel, Capital One has a rewards credit card that's perfect for you. With the Capital One Venture X card, you earn unlimited double miles on everything you buy. Plus, you get premium benefits at a collection of luxury hotels when you book on Capital One Travel. And with Venture X, you get access to over 1,000 airport lounges worldwide.

Open up a world of travel possibilities with the Capital One Venture X card. What's in your wallet? Terms apply. Lounge access is subject to change. See CapitalOne.com for details. Ever wonder what your lashes are destined for?

There's more we need to know today. Let's get to the briefs.

Two CEOs used AI-generated avatars to address investors watching their quarterly earnings calls.

Buy now, pay later finance company Klarna CEO and Zoom CEO both used a custom avatar. Zoom admitted it on the call as a way to demonstrate its own Zoom Clips video creation tool, which will be available to all users this week. Yeah, I'm not sure. I couldn't quite figure out why Klarna did it other than to get people to talk about Klarna, I guess. But the Zoom one made sense of like, hey, we've got this thing on a video conference that can do this. So I'm going to demonstrate it on a video conference. So.

Yeah. Zoom's got their clip showing like, hey, like here's how to make yours. And also here's a whole bunch of models, which I have to give credit has a wide range of diversity. If you want those models to speak for you instead. There are definitely days when I have gone too long without a haircut that I look at this and go like, maybe I should just use an avatar today.

Well, my thinking is like, man, like if you're going to do that, go all the way. Make yourself a VTuber, Avatar. Go hardcore anime. No kidding. Wall Street Journal reports that CEO Tim Cook called Texas Governor Greg Abbott personally last week to lobby for changes to legislation being considered in Texas that would require device makers to verify user age.

A similar law was adopted in Utah. It is under consideration in nine other U.S. states. The Texas version requires App Store operators to verify the age of the device owner.

Meta has lobbied for this kind of law instead of laws requiring service providers like Meta's own Facebook to do the verifying. Meta says Apple and Google should do this. Apple and Google say, no, this shouldn't be us. Nobody wants to do it. Also, Apple just can't avoid politics today. The president of the United States also posted on his social media that he might consider a 25 percent tariff on iPhones if they're not made in the U.S. So it looks like Tim Cook needs to make another call to a governmental figure as well.

As a total side note on the age detection thing, a game recently had added more precautions to keep kids away from this particular VR game. And they had been using photos of politicians and Half-Life's G-Man to bypass them. And I just think that's beautiful. Not a very good age verification then, if you can get away with that. Yeah.

You may have forgotten that the U.S. FTC was still pursuing a court case to prevent Microsoft's acquisition of Blizzard, which closed back in October of 2023. You will not therefore be shocked to learn that the FTC has decided not to appeal the Ninth Circuit Court's decision earlier this month to reject the FTC's appeal. And the FTC has officially dismissed the complaint.

Yeah, we just figured we needed to note this. You're like, this doesn't change anything. No, it doesn't. But the FTC kept going and everyone's like, why are you going to keep going? You're not going to win. And they didn't. So there you go. Stable coins. If you don't know what a stable coin is, it's a cryptocurrency that takes advantage of the flexibility and record keeping of a blockchain, but tied to a stable currency like the U.S. dollar to keep volatility down.

It's a way of transacting traditional money, but on a blockchain. The Wall Street Journal sources say some of the U.S.'s largest banks, and they name them, JPMorgan Chase, Bank of America, Citigroup, Wells Fargo, the folks who run the Zelle transfer system, and the Clearinghouse, have been discussing launching a joint stablecoin.

The talks are in the earliest stages. So this is just the big companies all getting together and others that I haven't named to talk about like, hey, should we do this? Kind of depends on what demand they think there would be for the benefits of that and the regulatory environment. There's various laws in states of being suspended or being passed that could affect this. But if they were to do it, it would likely speed up transactions, including cross-border payments.

It's like, I know they put that use case right there, but as somebody who deals frequently with cross-border payments, I'm thinking like, I'd rather just have the money. Well, you would. Yeah. I think that is one, I'm glad you mentioned that because I think that's one misimpression of how this always works is,

This doesn't mean you'd get paid in stablecoin. What it could mean, it could mean that, but it doesn't have to. It could mean that the financial groups use stablecoins to speed up the transaction, but they are on the ends normal money.

So I send you, I'm going to pay you in US dollars in this scenario, though. I send you $100. It goes into the system, is converted to a stable coin to make the transfer fast. And then on your end gets transferred back to US dollars. You just get the money. So it can give you what you want, but also speed things up.

Yeah, what I want are less fees. However they do that, that part would make me happy. Yeah, so I'm way into that.

Bloomberg sources say Apple is no longer researching the idea of putting a camera sensor into the Apple Watch. The idea was to use it to get information about things near the user, not to use it for selfies or video calls. Supposedly, Apple is still developing sensors that would go in earbuds to enable features like gesture control and improve spatial audio. Apple is also still planning to release smart glasses by the end of next year.

Yeah, this is the one that always gets reported is like they want to put cameras in your ear pods, which sounds nuts. But it's probably infrared sensors. So they're technically cameras. Sensors are different. Yeah, they're cameras in the strictest sense. But, you know, they're not going to be doing FaceTime from your earbuds or anything. I'm actually surprised that they don't want to use more sensors in the Apple Watch that way. But maybe they just found it didn't work.

Yeah, better sourcing issues. Yeah, could be that. That's a good one. Windows Notepad just got a new feature called Write.

which is hilarious that it's called Write because there actually used to be a program in Windows called Write from Microsoft. But this is a feature in Notepad that can generate written text from basic instructions. So it can write the text for you. Like, I need an intro letter to Jen about an upcoming project, blah, blah, blah. And then it'll just write it out for you. This is in addition to the existing summarize and rewrite features that were introduced in Notepad in March.

Like those, Write uses credits, so you have to have a paid Microsoft 365 account.

and Microsoft Paint got a bunch of new features. A sticker generator, an object select tool that uses some smarts to figure out what you're trying to select. The snipping tool has some smarts to isolate just the relevant parts of a screenshot for you. You can kind of roughly select it, and then it'll tighten it up for you. And a color picker that'll show you either the hex, RGB, or HSL values of a color in a screenshot. These updates are rolling out to Windows Insiders in the Canary and the dev channels.

For some reason, it just kind of hurts that this stuff is a notepad. Like, I just want a basic notepad. But I also understand why it's there. And I kind of like that it uses credits so that for the people who do not want this...

You can't use it anyway. Yeah. And you can turn it off. I think that's important to know. Like if you are like, if you're like, if you're like, I just want notepad to be notepad, you for now still can do that. Yeah. And for people who work in image programs, you probably use power tools on Windows to get the color pickers and stuff kind of integrated. So this being in paint, not the worst idea. I kind of like it. It's kind of handy. Yeah. The paint stuff particularly I find pretty handy. Mm-hmm.

On Thursday, Mozilla announced it would shut down the read it later service pocket on July 8th. A bit later, Digg co-founder Kevin Rose posted on X that Digg would be interested in buying pocket from Mozilla. So pocket was something I still have installed on Firefox.

But I haven't used it in a long time. I used to use it a lot. I know some friends and I know there's people in the audience that still use it a lot. It does seem to have a small user base, smaller than it used to be. But very dedicated. Very dedicated. So it's a PR win if Digg can buy this because it gets a lot of enthusiastic users happy that it keeps going. But I could also see it fitting in with the idea of

of dig as a place with you know collected links and surfacing popular things it could be another piece of information informing that so and that kind of makes sense to me those are the essentials for today let's dive a little deeper with some water bears

Tiny little cuddly tardigrades are getting tattoos. And if you have if you cannot picture what a tardigrade is, do yourself a favor and look it up. It will improve your day. Also, there's a very good reason they are getting these tattoos. And Dr. Nikki will fill us in. Dr. Nikki, welcome back.

Hello. It's time for Tardigrade. It's Tardigrade time. It's a new segment on Daily Tech News Show. Oh, I wish. No, we've got some interesting news. But while I know there's a big bunch of people in the audience who are like Tardigrade fans, probably, for those who don't know, what are Tardigrades?

I think the most important thing about them is that they're really cute. That's why they are fans. But they're really teeny tiny invertebrates, about half a millimeter large. They have eight legs. They're kind of chubby looking and they've been called water bears. And what's really cool about them, other than they look so cute, is that they are kind of...

like little invertebrate superheroes, if I may. They can survive high pressure, freezing, starvation, high temperatures, radiation. And which is why we often, I want to say often, send them to space. They've been to space. They kind of can enter into these like really intense hibernation states that allow them to survive all these things. So they're very cool little guys. Love a tardigrade. There should be a Captain Tardigrade. Oh man. I know people who have tardigrade stuffed animals and I'm kind of jealous. Yeah.

So why are we talking about tardigrades today? Yeah, right. Good point. So last month, researchers from Westlake University in Hangzhou, China, used tardigrades to test a microfabrication technique and gave them tiny, and I'll say in quotes here, tattoos. Oh, wow.

And it's pretty cool. And I recommend you, if you don't have the visual, to pull it up and take a look at what these looks like. But the goal here is for micro and nanoscale microfabrication to build things like microprocessors, biosensors and more. So they're first testing this out on a living organism.

Oh, okay. That's interesting. I have questions. First of all, how do you need a very tiny tattoo gun to tattoo a tiny tardigrade? Oh, that's a good iteration. So actually no ink is involved, but first they have to put them into this kind of

super intense hibernation, which is called a cryptobiotic state. And they do this by slowly dehydrating the tardigrades and then they cool them to negative 226 degrees Fahrenheit. And they, yeah, they're, and they're fine. They're just like in stasis basically. So it's not hibernation because they're not, their body almost goes into pause, but you can bring them back to life. Whereas in hibernation bears are still like

and breathing and stuff. Yes. So they're really, really cold. But they come back from it. They don't die. They can come back. Yeah. Yeah. And they cover them with this protective layer of a frozen organic compound called anisole, like this little layer of ice on top of them. And then another favorite of DCNS, it's not really a laser, but they use an electron beam to carve a pattern into the anisole, this compound that's covering them.

in this technique called ice lithography, which just means writing in ice. And so by this, it kind of melts the anisole and turns it into a biocompatible chemical that sticks to the tardigrade. And once you warm them back up and wake them up, it becomes a design and they come back to life and they have a tattoo kind of. Tardigrades are hardcore, man. Like, you know, like...

When we go to get a tattoo as humans, generally the stereotype is you get really drunk and then you get your teeth or something, right? No, no, that's really bad advice. You're specifically not supposed to get drunk because you'll bleed way more. It's the stereotype, right? Whereas tardigrades are like, no, the reality, the reality for us is we get killed and have an ice tattoo put on us and then we come back to life. Yeah, well...

As you'll see, it's not that great, but let's keep going. Yeah, yeah. Okay. So I assume they all got mom as their tattoo, right? Yeah. So first of all, only 40% of the tardigrade survived. So I'm sorry. Oh, okay. That's not good. But 40% is pretty good for a first try. The ones that did survive didn't show a change in behavior. So unlike humans, when they get tattoos and they act really tough as a person with tattoos, 30%.

Tardigrade stayed the same. They stayed humble. Okay, they didn't change their personality. That's good to know. No, but what they got on them, they tried a few different things. So they did squares, little dots, little lines as small as 27 nanometers wide. And of course, their university's logo because you got to. Yes, absolutely. How could they not?

All right. So let me get to the question that I'm sure a bunch of people listening have had. Why? Why? I know you mentioned microfabrication at the beginning, but why do you need to do this to a tardigrade? Yeah, so it is. It is microfabrication. So one of the goals of this is to build microscopic devices that are going to be compatible with living organisms. The next test they're going to do is trying to print this on bacteria. And previously until now, printing bacteria

Things like microelectronics, like even circuit shapes onto living tissue. That's never been something that we thought we could do. And now this is a possibility for the next generation.

DTN has an article that we'll cover. So that almost makes me wish we weren't calling it tattoo, even though that is good because tattoo implies something that is symbolic, whereas this is practical, like printing circuits, printing channels. There's so many different uses I can imagine this being used for, like drug delivery, but also like bio circuits that we might want to print on our own things. Yeah. Mm-hmm.

Yeah. Would you, would you get a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a, a,

you know, human, human cells. And there was a good reason for this, uh, again, like, like optimal drug delivery to, to cure a disease that I have, or it's some kind of enhancement, especially as I get older, that will help me walk better or something. I could absolutely, um,

go for a little ice lithography at some point. All right. We'll come back to this in 10 years, you guys. Hopefully in 10 years, hopefully it's that fast. That's great. Thank you for bringing this to us, Dr. Nikki. I really appreciate it. If folks want to follow more of what you do, where should they go?

Absolutely. And happy to talk about tardigrades all day. And you can find my stuff over at NicoleAckermans.com and also under my name on Blue Sky. Excellent. Thanks again. What would you like to hear us talk about on the show? You can let us know in our subreddit. Submit stories and vote on them at reddit.com slash r slash Daily Tech News Show.

We like to end every episode of DTNS with thoughts from you. What better way to finish than to get one of RW's excellent Patreon summaries of Thursday's show? Yes, and it's going to test my recollection of listening to yesterday's show. RW posted...

I already have a meh feeling about this. It will be really thin and the battery life will be rubbish. Hashtag YMMV for your mileage may vary regarding the IO hardware stuff that y'all were talking about yesterday, right? Oh, yeah. Yeah. All right. JC sounds like the best friend in the world and an excellent host too. Well, that's obviously you, Jen Connor. Thank you, RWA.

Yeah, that's nice. We are living in an AI world. Great background and great interview. That's the interview I did with GT Tandon yesterday from Telus Digital. And finally, if only it had a valve amplifier. Close as I could get. Picture this. You're in the garage, hands covered in grease, just finished up tuning your engine with a part you found on eBay. And you realize, you know what?

I could also use new brakes. So where do you go next? Back to eBay. You can find anything there. It's unreal. Wipers, headlights, even cold air intakes. It's all there. And you've got eBay guaranteed fit. You order a part, and if it doesn't fit, send it back. Simple as that.

Look, DIY fixes can be major. Doesn't matter if it's just maintenance or a major mod. You got it, especially when things are guaranteed to fit. So when you dive into your next car project, start with eBay. All the parts you need at prices you'll love. Guaranteed to fit every time. eBay. Things people love.

This episode is brought to you by Progressive Insurance. Do you ever think about switching insurance companies to see if you could save some cash? Progressive makes it easy to see if you could save when you bundle your home and auto policies. Try it at Progressive.com. Progressive Casualty Insurance Company and Affiliates. Potential savings will vary. Not available in all states. What is that?

I'm kind of, I'm leaning towards the Marshall soundbar. That makes sense. Okay. Thank you, RW Nash, for your always fun summaries and reactions to the shows. They make my day every day. And he does them for all the shows. It's really cool. Thank you, man. Appreciate it.

And we want to know, what are you thinking about? Have you got some insight into a story? Please share it with us over at feedback at dailytechnewsshow.com. Thanks to Dr. Nikki and RW for contributing to today's show. Thank you for being along for Daily Tech News Show. The show is made possible by our patrons at patreon.com slash DTNS. No show on Monday because of the U.S. holiday, but we'll be back on Tuesday. Talk to you then.

This week's episodes of Daily Tech News Show Briefing were created by the following people. Host, producer, and writer, Tom Merritt. Host and writer, Jason Howell. Co-host, Rob Dunwood. Co-host, Jen Cutter. Producer, Anthony Lamos. Producer, Roger Chang. Editor, Hammond Chamberlain. Editor, Victor Bognot. Science correspondent, Dr. Nikki Ackermans. Social media producer and moderator, Zoe Detterding. Ah!

♪♪

This show is part of the Frog Pants Network. Get more at frogpants.com. Diamond Club hopes you have enjoyed this program. Take control of the numbers and supercharge your small business with Xero. That's X-E-R-O.

This episode is brought to you by Progressive Insurance. Do you

Do you ever think about switching insurance companies to see if you could save some cash? Progressive makes it easy to see if you could save when you bundle your home and auto policies. Try it at Progressive.com. Progressive Casualty Insurance Company and Affiliates. Potential savings will vary. Not available in all states.