Microsoft’s 10,000-year Storage Tech - DTNS 5209

0:02This is the Daily Tech News for Thursday, February 19th, 2026. We tell you what we think you need to know. We give you some important context, and hopefully that's going to help each other understand. Today, Microsoft has developed a storage alternative that lasts 10 ,000 years, and it's made of glass. But will it help us with Ramageddon? Who knows? I'm Tom Merritt. I'm Wint Witt Dow. Let's start with what you need to know with the big story. All right. Well, with hard drive prices rising, you might think it's better to just, you know, go back to engraving stone tablets. Guess what? That's actually cutting edge.

0:40Tom sat down with Dr. Nikki to understand how Microsoft has developed its long-term storage. Dr. Nikki, thank you so much for joining us. I'm happy to be here and talk about ancient or modern tablets. Yes. So should I pull out my stone tablets? What are we talking about here? Why are we talking about engraving tablets? I was hoping so, but no, we're talking about data storage. As ancient tablets were also for that, but now we're talking about that on a slightly larger scale. You probably generate a lot of data every day just in DTNS recordings. As a species, we basically double our data production every three years.

1:21So it's exponentially growing and preserving that data, or at least some of it for future generations is becoming challenging. We have current storage solutions like magnetically encoded hard drives. But as you may have experienced, they tend to fail usually about five to 10 years out. Yeah. And usually when you least want them to. Yeah. Right. When you need to start them back up for something you fried your computer for. Yeah. And then there's also magnetic tapes, but those can also degrade after about three decades. So people have started to turn to optical storage like laser writing specifically on a very kind of durable material, which would be glass.

2:00And they think this may have the potential to increase the longevity of your storage. So it is a little bit like engraving tablets, right? I mean, it's almost exactly the same. This is exactly new. Yeah. Yeah. I mean, people have been writing on tablets. In fact, the first form of writing was on a tablet from 2000 BCE. So, you know, not a new technology. But as far as the glass engraving, people have been working on forms of this since 2013. On the Wikipedia article, you can see there's like a Magna Carta on a little glass disc and a Bible that people have engraved. So they've been working on it for a minute.

2:34But now this recent article in Nature, which was published by the Microsoft Research Project Silica team, actually came up with a end-to-end system for this where you can write your data, store it, and retrieve it from these glass tablets. And it's all automated. So it's kind of like a bundle system. So what is Project Silica? Yeah. So cool name. This is what Microsoft has called this storage technology. It's based on femtosecond laser direct writing, which is cool. Basically, you can modify a material's properties at the nanoscale. They chose glass because it's so durable, but you could do this on like a strong plastic if you wanted to.

3:20Potentially plastic is more degradable. How this works is each modification, and we'll get into what a modification is, but your data is written in a volume of a pixel, which is called a voxel. And these voxels are super densely packed onto the glass cube because the laser is very accurate. Some might call it laser accuracy. Laser precise. And this gives you basically super or rather ultra high capacity 5D optical storage. 5D? Yeah, I checked. I was like, are they just saying this to say like exaggerate? But the 5D comes from the fact that the information can be coded by also changing the laser's wavelength and polarization.

4:02So you can get like infrared or, you know, differently polarized the way that you engrave the data. So it adds two more dimensions. Some people say that this is similar to the Superman memory crystal. Ah, okay. So this is Kal-El's technology. That's where we got it. Yeah, we're getting there. All right. So obviously this is not writing runes on here. How does this actually work on the data level? How do they encode the data? So you could take your runes and type them into Word, I guess, and then convert them into data. The data gets received as bits. These bits are then grouped into specifically coded symbols, and each symbol corresponds to a voxel that stores a little bit more than one bit.

4:49So small. Yeah, really small. And these voxels are written into two-dimensional planes and they're stacked three times into a 3D volume. So you get like a 120 millimeter square that's two. So you're not just reading the surface, you're like scanning. Yeah, it's also stacked into depth in this glass. And you can read it with normal wide field microscopy. It's what I have in my lab right now. And you then decode the symbols using a convolutional neural network back into bits. And you can add layers of encryption to this if you add like a static layer in between and the convolutional network can like decode that.

5:25they also while they're doing this introduce redundancy into the data so that if some of it can't be uh decoded or recovered you can still get most of the information because you like duplicate it and i could see that being similar to like how a qr code works where even if you have a portion of the qr code you can still usually get all the data out of here um what are we talking about as far as specs you know like how much can we store this way little square of glass 120 and 20 millimeters square. That's not very big. Uh, two millimeters thick is 4.8 terabytes. That's better than the hard drive that I've got in my lab.

6:01Um, and that's the equivalent of 2 million printed books. Uh, what's, what's your latest book called? Uh, Sinked. So 2 million Sinked printed. There's the plug for you. Yeah. Let me sell that many someday. And 5 ,000, uh, Ultra HD movie. So Superman. Okay. Yeah. And it's relatively fast. The laser emits 10 to 25 million pulses per second. I say relatively, cause it's not as fast as, as current writing, but you know, that's a lot of pulses. It's not meant to be used as your daily hard drive. It's meant to retrieve long-term storage. So it's fast for that actually, like compared to tape drive or something.

6:40Yeah. Yeah. Also they predict that the data the lifetime of this is 10 ,000 years at room temperature. Now they haven't tested this because it would need to wait 10 ,000 years. 10 ,000 years, yeah. But they basically heated up the cubes to 500 degrees Celsius and said, you know, did some math. And we're like, this could last 10 ,000 years. But they're not quite sure, you know, if you vary the conditions, you freeze it or something, if that might change that. Yeah. So it kind of seems obvious some of the advantages to this. But beyond the obvious, what are the pros to doing it this way? So, yeah, I guess some of the obvious, I mean, for me, I think it would look really cool.

7:18Yeah, exactly. A facility where everything's just like glass cubes would be pretty awesome. I've been thinking about that the whole time I was writing this. I was like, oh, you got to put energy lights. I mean, I'm making a fortress of solitude out of these for sure. But of course, it saves money to do it this way. I mean, lasers and glass are not very expensive. Equipment and energy, since it's not very energy used, they don't use a lot of energy to do it. And of course, it lasts a lot longer. Glass itself is thermically and chemically stable and resistant to moisture and electromagnetic inference, which other types of hard drives are not.

7:50Yeah. I mean, basically, they tell you after 10 years, all bets are off on a typical magnetic hard drive, right? Exactly. I've got hard drives that have lasted longer than that, but that's by accident, not by design. Yeah, you didn't heat it up to 500 degrees. Now, obviously, it's not all roses with any technology. What are the cons? What are the challenges to this? Yeah. So while they say, you know, this is a fast, dense and efficient writing technique, it's no competition for current modern day data centers in terms of the capacity or the writing speed, but the longevity is better. It is difficult to actually test that longevity because you just have to wait.

8:28And there is a lot more research needed to kind of figure out how extreme conditions would impact this. So like I said, if it's really cold or there's an earthquake or, you know, whatever. Also, this just got invented. So you would need to really widely distribute this and create a supportive ecosystem so that it gets adopted everywhere. It's not useful to have your laser disc and be the only person that uses that and nobody else uses it. Yeah. You don't want it to be the beta max of storage. Exactly. And the researchers were thinking like maybe in the future, this can be adapted to fit inside a smartphone for writing inside a smartphone.

9:04And then everyone would have one. I don't know how far away in the future this is. Do we need them inside our smartphones? I guess just for archival purposes? Potentially. Because it's smaller. Yeah, like writing everything depends how much data you want to save. And they said, you know, this would be great for historical data, archival data, or for example, climate data, which is like a lot of data and we need to keep for a very long time. So those are pretty interesting examples. Yeah, yeah, yeah. I don't know if you're familiar with Anasir. No. Okay, so one of an old like engraved text that we found from Anazir, who's a copper merchant, was basically someone making a Yelp review about how terrible his copper was.

9:47And it's like a huge long rant. It's like one of the best preserved texts that we have. I didn't remember the name, but I have heard about this now that you've mentioned it. So I really want someone to engrave that into a glass tablet so that it's like a circle. That's fantastic. Well, this is like, even if it's just for archival purposes, this is a very promising technology, which honestly, if adoption is your biggest downside and the fact that you haven't tested it for 10 ,000 years, that's pretty good. So I'm pretty bullish on this. Thanks for bringing this to us, Nikki. I want a glass hard drive.

10:18I do too, right? My PC wants a heart of glass. Oh, yeah. Uh, if, uh, people would like to help you, uh, along the way to getting your glass hard drive, uh, where can they support what you do? Uh, you can find everything I do at NicoleAckermans.com and I have the same handle on Blue Sky. Fantastic. Thanks, Nikki. This is a, uh, a pretty cool technology, uh, that, that is, is going to be useful in a myriad number of ways that we don't even know yet, honestly. I love that the 5D aspect was real, that you can have different wavelengths. That is so cool. I can't even describe how cool it is to actually have something that was 5D.

11:06And I really, really want a glass SD card for my smartphones. I need this now. I really, I'm waiting for that. Just to store four terabytes of data? Of whatever data. Yeah. Yeah. No kidding. I can hear Joey Image already calling Microsoft saying, I need this for my vast amounts of storage. I do have a question. What happens after 10 ,000 years? We don't know. We don't know. They're like, at least, it's at least 10 ,000 years, maybe longer. I know how those numbers come up. I'm just always curious, like, in terms of like a model or extrapolation. Turn into dust. Immediately, 10 ,000 years, we're done.

11:45It just turns into sand. Yeah. Oh, gosh, that was amazing. And I feel very lucky today that I got to listen to such an amazing story because of you guys, the listeners. DTNX is made possible by you. And thank you to Paul Boyer, D. Laser, and Brad.

12:07There's more we need to know today, so let's get to the briefs. Okay, so the information sources say that Meta plans to release its first smartwatch this year, codenamed Malibu 2. It would, of course, include Meta AI and health tracking. The information has been reporting on Meta's smartwatch development since 2021. Meta reportedly paused the project in 2022, but started it back up again this year as part of the company's focus on wearables. Meta already makes virtual reality headsets and smart glasses and is working on more advanced glasses that can do augmented reality. A mixed reality headset is reportedly targeted for 2027.

12:48Yeah, so Meta switching back to saying, oh, wait, wearables is what we want to do. Let's get that watch going. Hence the Malibu 2. The earlier project was just Malibu. I'm still waiting on Malibu Stacy. Casey, it is interesting to me to see more of these kinds of efforts from the big companies. Like, you know, Google could tell Meta how hard it is to get into hardware, but it is making inroads and Meta is placing more bets on devices. Does that mean people think Apple is ripe for the taking in new categories, right? It's much harder for Google to make way in pixels, but maybe in other stuff, maybe in glasses, maybe in the new categories that haven't been taken yet.

13:38Oh, that's a good point. First, that's just the rule. That's just it's been the Internet tradition, right? Being first somewhere or being one of the first and finding fresh territory. It kind of makes sense. I mean, it's like an extension of like what we do every day as like software developers at like, you know, services as a wait, as SaaS company. Software as a service. Software as a service companies is, it's all about, you know, increasing like users and in this case, buyers. And sometimes it just takes like finding new things and that could be hardware. Although, yeah, the capital investment for hardware is a lot different.

14:09So it is interesting to see what happens. I mean, bad software just apparates like it was written on 20 ,000 year old class slates. And it turns to sand. It turns back into sand. And there's like a genie from Aladdin joke there, 10 ,000 years. But anyway, yeah, it is interesting because, yeah, hardware was never one of those things that seemed like an easy prospect. But, I mean, Meta, as you said, is making an inroad, so we'll see. Yeah, I mean, they bet on buying the Oculus Rift. And you could say they won because they dominate the category, even if it's not a very big category. And so they're willing to take some more bets because I feel like they're certainly currently dominating the smart glasses category.

14:52They may end up being the BlackBerry or the Nokia of that category. We'll see. But yeah, it makes sense that they would want to get into a watch because of the wearables. But the watch is already a very competitive category. So we will see. Speaking of Meta, CEO Mark Zuckerberg took the witness stand in Los Angeles on Wednesday. This is the case brought by a 19-year-old identified as KGM along with her mother against currently Meta and YouTube, alleging that their algorithms cause mental health damage. The case was originally filed against TikTok and Snap as well, but those two companies have settled.

15:27In an attempt to portray Meta as more proactive about safety, Zuckerberg testified that, yes, he had exchanged emails with Apple's Tim Cook to discuss the, quote, well-being of teens and kids. When asked by the plaintiff's lawyer why he lifted a ban on cosmetic surgery filters for Instagram, He said, I genuinely want to err on the side of giving people the ability to express themselves. So far, so much you would expect from Zuckerberg. He also asserted there was not enough casual evidence of harm, especially relating to these kinds of filters. And when the plaintiff's lawyer pressed him on this to say, hey, are you an expert?

16:08Do you have a college degree that would indicate you're an expert in causal evidence? Zuckerberg said, I don't have a college degree in anything. Zuckerberg also claimed that Meta doesn't have an engagement objective for its users, which I think all of us collectively could question whether he was really sincere, although he was under oath. And he said that it is, quote, very difficult to enforce Instagram's age limits, which I think he does believe. Meta also being sued by the attorney general of New Mexico over an alleged failure to ensure children are safe from online predators. Another trial in Northern California will begin this summer, alleging meta in YouTube's apps also damage mental health.

16:49The one we're talking about today is from Southern California. There's one that's going to happen up north as well. And there's a worldwide trend to completely ban children from using social media apps. We've talked about Australia, which just began enforcing its ban in December. Brazil's child safety law has come into effect. It will be enforced in March. That will require age verification, mostly for adult sites, not for social media, but people think it might get expanded. More than a dozen countries are currently considering bans for children, including Austria, Bulgaria, Czechia, Denmark, Finland, France, Germany, India, Italy, Malaysia, Norway, Poland, Slovenia, Spain, the UK, and in the US, Florida has begun enforcing its ban for users younger than 14.

17:33Now, some child safety advocates believe this distracts from the problem rather than helping increase safety, That it's easy for a government to just put in place a ban because somebody else has to do the hard work and then they can say we did something. Apar Gupta, a founder of the India's Internet Freedom Foundation, told Rest of World platform design choices around recommendation algorithms, data harvesting practices and addictive features cause far more harm than anonymous access to information. uh and gupta is saying you know you you can you can ban the kids but they'll often find a way around it and then if you haven't fixed the underlying problem they'll still run into it yeah we we talked about this like i think like last week and a week two weeks ago where and i i think i have like the same and i i had the same like initial like reaction as a human being it's like oh of course certain things like this are addictive but that we don't really have specific ways of measuring this and studying this because it's like very it becomes very nebulous where we want to do something to help.

18:30And we try things, but they may not be. And I agree, like just banning something. And I'm very sympathetic to the idea that, well, banning it doesn't, is more of, I guess, supposedly tactical in that it's structural problem, but also kind of not really getting to the root of the issue. And that something else will kind of, that these practices, which by the way, I wish you could see the face I made when you said that Meta, like Meta, that Zuckerberg claims Meta doesn't have engagement objectives. As a software engineer, the engagement is one of the words that haunts my nightmares because that is what drives a lot of like sass and and social like that's how ads happen that's how a lot of things happen so i'm not saying they have mosseri on the stand adam mosseri that runs instagram earlier in the week talking about engagement as a metric i'm not sure what he thought he was going to try to get away with there i i don't i mean and i i'll give him credit that maybe he i just i'm not understanding what he's trying to say in regards to what their objective is.

19:24But it's something that I guess I see in that companies, software companies are so hyper-focused on engagement because really it's a hard thing to understand, especially when you have something like software or services and stuff like that. What does it mean? What brings in the money? What makes you successful? And engagement and these kind of metrics are the kind of thing that we have hyper-focused on. And like anything else, line go up. we've connected engagement and certain kinds of like behaviors as line goes up for the company. It's just, to be fair, it's not very sophisticated and potentially dangerous, but it's kind of like the metric that we have.

20:01And I think just, I guess, to tie it back to this, that, yeah, there's something in there that has become an issue. And yeah, I don't, like, it's hard because like we don't, it's something that's so widespread, so like diverse in all the different cases that it can happen. We don't really understand it. And I just wish you could have seen the face I made because I just – Yeah, man. I mean, engagement may or may not be the best metric for certain things. And maybe that's what he was trying to say is like we are not trying to maximize engagement at the expense of everything else. Yeah, that's fair.

20:31But what he actually said was it wasn't a goal of Meta. It's a goal. I'm pretty sure it's a goal for every company on the internet, including Daily Tech News Show, right? I mean that's – I guess like to me put a positive spin on it. Maybe we all as an industry and as like – I'm going to say tech is an industry, but like this tech community, we could do better to introspect on like different ways of increasing whatever it is that we're trying to increase. Right. Engagement is a metric. It's a metric, yeah. And certain kinds of engagement are not a goal of DTNS, right? Yes. What we do is we look at it to say, oh, are people engaging with our news the way we want them, the way we intend it to be engaged with?

21:14Yeah. Wow. All right. Well, we want to thank Motang for posting this on the subreddit. We mentioned this, I think, last week, but Ring took a lot of criticism for its Super Bowl commercial featuring Search Party, which was an image recognition feature that would be used to help find a lost dog. 404 Media then obtained an email from Ring founder Jamie Siminoff that said, quote, Search party first for finding dogs will end up becoming one of the most important pieces of tech and innovation to truly unlock the impact of our mission. You can now see a future where we are able to zero out crime in neighborhoods.

21:47Ring confirmed to The Verge that the email is genuine. The New York Times reports that Siminoff is conducting interviews to apologize. Yeah. Yeah. Listen, he did not say in that email, and we will go from dogs to looking for people. But clearly that is the plan. When you're saying to zero out crime, I don't think he's talking about crime committed by dogs. He's talking about crime committed by people. So, and I don't think any of us have a problem if we felt that it would be limited to actual crimes and actual criminals, Where it gets people upset is, OK, but then once this is in the hands of the authorities, will they use it improperly?

22:32And I understand both all the different various perspectives on this, which is some people thinking it's exaggerated that your local police department will misuse this. Others clearly have evidence that their local police department has misused information in the past. And so they have no trust that they won't misuse it in the future. but Ring is pretty bad at handling this sort of thing, I have to say. Yeah, not optimistic. And I mean, obviously because, as the New York Times reports, if Siminoff is apologizing, it kind of lends credence to the idea that there was something to apologize for. Yeah, yeah.

23:11This next one, I think, is a good cleanser for these last two stories. I saw a link on Hacker News today to a subreddit post from a user called Anvarezizov. Anvar Azizov is a Ukrainian user that described how they're using two low-RA radios. Those are long-range, low-power radios to control home assistance, a roll-your-own home automation system, when Russian bombings take out internet and cell towers. Now, they have a home battery backup. So when the electricity goes out to the Internet, this person still has home battery and then has a LilyGo T-Echo radio plugged into a Mac Mini and then takes a second T-Echo radio with them and has a Python listener demon routing commands and sensor data between the radio and the home assistant.

24:04so they can still turn lights on and off, lock doors and turn on the security system, et cetera, even when they are away from home and the power has gone out. Amazing. Wow. That's just really cool, dude. Go check out this Reddit post because this person has basically put all of the specs of like, here's what I use, here's what I plug into what. This is, it's like 30 bucks for the radio. This is how I make it happen. And the conditions under which this is necessary are horrible, but the technology and the workaround are impressive. That does – I regret the situation in which he had to do this, but just the ingenuity warms my heart, and that's pretty awesome.

24:47Yeah, he's got very detailed instructions. That's cool. We'll have a link to this Reddit post in the subreddit, or you can look for it on Hacker News as well. real quickly folks if you want honest reviews from people who actually buy and live with tech you need live with it sarah lane hosts a weekly look at the tech both she and others have been using in their daily life this week's episode is about the phi dog collar tracker that her vet actually hooked her up with and she's been using with her dog rex she's got a very interesting perspective of what it might be good for and why it's probably not great for her and her dog listen to live with it wherever fine podcasts are found, or you can watch it at youtube.com slash daily tech news show.

25:31All right, let's get some quick headlines underway. These are little tidbits that are going to make you look smart. So pay attention, pay attention class. Well, Mastodon announced plans to improve the platform and to make it easier for new users, especially content creators, promote small and also promote smaller servers and help server operators handle maintenance and moderation tasks more easily. Yeah, Mastodon's been staffing up, so this is kind of the first of the fruits of that. In the latest Windows 11 beta build, Microsoft is testing, making speed test available by right-clicking the network icon in the system tray.

26:05It would then open in your default browser, but that's pretty fast. Oh my gosh, yes, please. I do speed tests all the time. Right? Yes, yes, yes, yes. That one might, Wait, would that get me to use Windows 11? Sorry. Google has added the camera to the iOS version of SnapSpeed in the top right corner of the app, along with other improved features like manual control. A lot of people love SnapSpeed. Bloomberg reports that European officials say Russia's army is facing evident communications difficulties. It's changing how the front lines are moving after both the loss of Starlink access as well as Russia's own block of Telegram services.

26:46Google Maps now uses a limited mode if you are not logged into Google. That mode doesn't show things like user reviews, photos, related locations and features like popular times or how busy a location is. Accenture confirmed to CNBC that, yes, the company has informed some senior staff that they must regularly use AI tools. Their logins will be logged if they want to be considered for promotion. Google announced the launch of Gemini 3.1 Pro, its first point upgrade to Gemini, which it claims has more than double the reasoning performance of 3 Pro. Amazon has passed Walmart as the biggest company in the world by revenue.

27:26So if you're like, wait, I thought NVIDIA was the biggest company. That's by stock price. This is by actual money brought in from selling things, which is one of the reasons Walmart is so busy trying to be like Amazon, I guess. Well, Etsy has announced it will sell secondhand clothing market Depop to eBay. I guess that makes sense. The United States is creating a portal called Freedom.gov that will have a built-in VPN intended to let users access content that is banned in their country but is legal to publish in the United States. And Apple lacks correlation to the rest of the Nasdaq stocks as it's not seen as a participant in the AI race, which means currently that it's being seen as a safe haven from the volatility around other tech stocks.

28:14That's a silver lining if I've ever heard one. And much is being made of a photo taken at India's AI Summit 2026, where Prime Minister Narendra Modi asked 13 business leaders on stage with him to clasp hands and raise their arms in solidarity. OpenAI CEO Sam Altman clasped Prime Minister Modi's hand. Good positioning for Altman. Anthropic CEO Dario Modi clasped the hand of Alexander Wang, the CEO of Scale AI. But Altman and Emody held their hands in fists next to each other rather than actually touching. And everybody's making a big deal out of it. This is like some high school mean girls type stuff.

28:57And I'm here for it. Yeah. It's amazing. I wonder which one tried to grasp the other's hand and the other one was like, no. No, we're not. All right. Well, we end every episode of DTS with some shared perspective. Today, Jay has another example of how putting in the work for the prompt made all the difference in getting a result from an LLM that was useful. Yeah, Jay wrote, your feedback from Steve today inspired me to share an AI-assisted coding anecdote of my own. I've been a software engineer since the 90s. Very recently, I had an idea for a new feature that would realistically take a week or two to implement.

29:31But after some internal, let's politely call them debates about AI, I decided to see just how far I could get letting an LLM do all the coding for me. But one of the immediate challenges was that the platform and UI frameworks I was using were proprietary and not well documented. So I spent the first couple of hours using LLMs to look at existing code and usage to construct a small set of fairly comprehensive docs on how to build features for this system. I then used Codex in planning mode to work through a solid requirements document, occasionally bringing other models in to critique the requirements or raise questions I might not have thought of.

30:08With over half of the day already gone, I eventually had a strong set of documents explaining what I wanted and fairly decent directions on how to build it. Then, and only then, did I let an LLM loose on actually implementing the feature. After several minutes, it had built the entire feature, front-end and back-end, well-organized, well-structured, and following the same patterns and best practices used across our organization. Best of all, it actually ran. The UI was aesthetically terrible, and there were a handful of code quality points that I quibbled with the LLM on, but by the end of the day, I had close to a week's worth of work not only done, but done to the standards I would have expected of myself.

30:52The key to all of that was making sure the models clearly understood those expectations and had enough information to meet them. Great show as always, Jay. Jay, I really do appreciate, again, like kind of the focus. And this is where I think I would like to see, I'm trying to be as positive as possible because, again, I am the AI skeptic around here. Very much so, especially as an engineer. I do really appreciate, though, like your emphasis on like how the tool is built and how like those steps that you did. I think if anything, this is like the kind of stuff again, and I've said that I said this last week in response to Andrew's emails, but this is the kind of stuff that I love hearing about people approaching these technologies and how to use them.

31:32And I imagine that in an ideal world, this would be what like our jobs will become kind of like a conductor or an editor, as we've kind of mentioned the show before. And that, of course, you put guardrails, like quality, you know, quality expectations, like expectations around what it's building. And of course, you obviously have enough experience to understand what you really wanted the output to be. So I can only say this sounds great. Yeah, I think Jay is a good example. Another good example. We now have Steve and Jay and Andrew all saying like, yeah, you have to put work into it. You can't just say senior executives, use the AI and otherwise you won't be promoted.

32:12You got to give people like this is how to use it. This is the proper way to use it. This is not the proper way to use it. And so I love Jay's example because he's honest about that. Like, yeah, the interface was terrible. I had to do a lot of work to fix that. It's not like it just coughed out a perfect thing right from the beginning. Yeah. And I think so just to reiterate some things I said is that number one, especially as engineers, the job is to kind of understand how the tool works best and understand what is the most efficient way to use it. And my big problem is that not people like Jay or Andrew or anyone else we've talked to and Steve, who are doing the engineering thing and discovering how this tool can work and what is the best way to optimally use a tool.

32:49Unfortunately, a lot of us are put in situations where someone is saying, hey, you have to use AI, the exception thing. And not only that, but as people with jobs, our performance is being evaluated by, and as I mentioned, one of my friends said, yeah, we're going to get rated on how many tokens we use. Yeah, that's ridiculous. Ridiculous. And if you're a football fan, and this is the metaphor I've been working on, This is like trying to grade a wide receiver on an American football team by how many steps they run. There is a correlation there between a good wide receiver and how much they run, but it's just – it's a correlation.

33:21It's a small part of it, and there's so much nuance, and that's not measuring output, like the quality of output and the actual results that you're getting. And unfortunately, people like – so Jay and Steve and Andrew are doing the work that I love to see from engineers. But unfortunately, we're being placed as human beings with jobs that need to have – make a living into situations where the incentives are not to do what they are doing but to just use the AI, crank it out, and be 3Xers or 10X output. So that is the problem I'm having. You guys are giving me hope that we'll hopefully find a way to make this work because that's the future that I would like to see.

33:54So focusing on the Accenture thing with the senior executives and any other systems that are similar out there, I would use the technical term bass-ackwards. Those are – because they're putting – you're putting the horse behind the cart. You're not going to push the cart doing that. That is a technical term. I will continue to use it. All right. What are you thinking about? Do you have some insight? We would love for you to share it with us. Feedback at dailytechnewsshow.com. Thanks to Dr. Nicky and Jay for contributing to today's show. Thank you for being along for Daily Tech News Show. You folks keep us in business.

34:31Thank you if you're a patron. And if you're not a patron, go become one. It's easy. Patreon.com slash DTNS.

34:41The DTNS family of podcasts. Helping each other understand. Diamond Club hopes you have enjoyed this program.