Yann LeCun’s World Models Raise $1 Billion - DTNS 5222

0:00Tom Merritt: This episode is brought to you by Progressive Insurance. Do you ever find yourself playing the budgeting game? Shifting a little money here, a little there, and hoping it all works out? Well, with the Name Your Price tool from Progressive, you can be a better budgeter and potentially lower your insurance bill too. You tell Progressive what you want to pay for car insurance, and they'll help you find options within your budget. Try it today at Progressive.com. Progressive Casualty Insurance Company and Affiliates. Price and coverage match limited by state law. Not available in all states.

0:28Jason Howell: Guys, it's no use putting it off. The best time for an underwear refresh is now. Tommy John Underwear is designed for a perfect fit that stays put all day. There's zero chafe thanks to four times more stretch than competing brands. And their innovative horizontal quick-draw fly is a game changer. With over 30 million pairs sold, there are thousands of men out there more comfortable than you. Don't settle for less. Go to TommyJohn.com today for 25 % off your first order with code COMFORT. That's TommyJohn.com slash comfort. TommyJohn, comfort perfected. This episode is brought to you by Capital One.

1:04Jason Howell: Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade-in value. advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. Parles-tu français? Hablas espanol? Parli italiano? If you've used Babbel, you would. Babbel's conversation-based technique

1:39Tom Merritt: teaches you useful words and phrases to get you speaking quickly about the things you actually talk about in the real world. With lessons handcrafted by over 200 language experts and voiced by real native speakers, Babbel is like having a private tutor in your pocket. Start speaking with Babbel today. Get up to 55 % off your Babbel subscription right now at babbel.com slash ACAST. Spelled B-A-B-B-E-L dot com slash ACAST. Rules and restrictions may apply.

2:17Jason Howell: This is the Daily Tech News for Tuesday, March 10th, 2026. We tell you what you need to know, give you the important context, and help each other understand.

2:26Tom Merritt: Ah, yes, that's what we try to do. And today, Jan LeCun is betting that LLMs are dead. They're a dead end. They're not dead. They're a dead end. But world models, world models are the future. Time to learn a new buzzword, folks.

2:40Jason Howell: Yeah, we'll talk all about it. I'm Jason Howell. I'm Tom Merritt. Let's start with what you need to know with the big story. advanced machine intelligence aka ami labs that is jan lecun's new world model company has raised 1.03 billion dollars so let's just say a billion dollars and some change uh jan left meta you may remember last november where he was chief ai scientist for a number of years He left to co-found AMI and concentrate on world models as his bet for achieving human-level AI. He argues that LLMs alone are not going to lead by themselves to AGI and that world models are the real missing piece here.

3:28Jason Howell: Now, world models learn from video, they learn from sensor data, interactions in the real world to build an internal simulation of how objects and agents and time and other real world entities, let's say, actually behave. Lacoon compares this to large language models where he says they mainly predict the next word based on patterns in the text, right? So think of world models as something along the lines of common sense, let's say, common sense knowledge of the world. The development is directed at cause and effect in real environments. That basically means it has a potential to be very impactful, not only just for general AI applications, as Jan LeCun believes, but also once we start getting into, you know, throwing it in air quotes, real world stuff, we're talking physical AI.

4:25Jason Howell: We're talking robotics. And yeah, Jan LeCun, he's been very vocal about his belief that these LLMs, it's not that he says LLMs are pointless or useless or anything. He just says they're really good at doing that thing, that predictive thing. But that isn't enough to have a true, honest understanding on a technical level of how the world works. And that's what's needed to get truly to this human level AI point that people are putting onto a calendar and saying, let's get there by this date.

4:58Tom Merritt: But what has Jan LeCun ever done other than revolutionize optical character recognition and provide some of the bases for the machine intelligence that we have today? Good point. What would he know? No, I think he's worth paying attention to, of course. A lot of smart people can be wrong. I am usually wrong, but I look at this and I've been thinking like, look, large language models are very good. They are better than we expected for a lot of things. But there is a temptation, I think, by some, both in and outside the industry, to think they can do everything. Like, oh, they're really good at this.

5:37Tom Merritt: What if we make them do that? And we've seen that you can do a lot of things like multiple agents, deep thinking, thinking models where you have them double check each other that do make them better than just predicting the next word. You can do some double checking and get better results out of them. But there is diminishing returns the farther you go down that road. And I think that's what Yalan Kun is rightly pointing out is there are other models out there. There are deep learning models out there that we have seen play video games. Those are not LLMs. I think a lot of people in their mind, they just class that, oh, that robot that plays Go, the thing that can play StarCraft, that is a deep learning model that is different from LLMs.

6:20Tom Merritt: A lot of people taking LLMs and deep learning models and playing them off each other. That's an interesting thing, too. There's also computer vision, which Jan LeCun has put a lot of work into over the years, which is being able to recognize objects and learn how to recognize objects. So the commonality amongst all these models is reinforcement learning, being able to improve the model in a way that's similar to how we learn. But I think he's right that world models could be the next thing. Then you get into the question of like, well, what do you mean by world model? And I think you did a great job of explaining what Jan LeCun means.

6:57Tom Merritt: But don't be confused if you see someone else out there talking about a world model and you're like, wait, that doesn't sound like what Jason said, because there's really not a definition of them yet. And even LeCun said like in a year or so, you're going to see everybody claiming they have a world model because he thinks it's going to be buzzy like that. And so we're going to have to come to a definition of what we really think this means. But generally, I think a usable definition, a usable way to think of it is pretty close to what you said, which is a model that isn't just kind of simulating language, but a model that is learning how to adapt to the world around it.

7:32Tom Merritt: And digital twins are a kind of world model, if you've heard of those. But sort of that ability to not only simulate but recognize the world, I think is fair to say like, okay, now you're talking about some kind of world model.

7:46Jason Howell: Yeah, it's really interesting where we are right now with LLM technology. I saw a video and kind of the limitations that we run up against, even though they seem in one hand brilliant in certain ways, brilliant applications for certain things. And then you run up against this common sense idea. And it's like, well, it doesn't necessarily understand what common sense is, right? Like I saw a video just the other day of a guy who has a pen in his hands, and he's holding it with two hands on both sides, and he's talking with ChatGPT, and he's basically saying, what will happen when I let go with one hand, ChatGPT?

8:26Jason Howell: and it's like, oh, well, the pen is going to fall to the ground. It's going to fall down because of gravity. And even when he pulls his hand away, the pen, the chat GPT is like, yeah, see, it fell to the ground. He's like, no, it didn't. My hand is back here. So on one hand, these things are brilliant. But on the other hand, they don't quite have that connection that even babies, when human babies are born and learn in their limited capacity, their sensory capacity, these common sense things, LLMs don't always necessarily get there. And I think part of Jan LeCun's, he even uses the human baby example as kind of part of the understanding of what the importance of a world model, however you define that, however you get there, what the importance of that is to the development of all this stuff to get to this point to where these things are truly human knowledge capable.

9:20Jason Howell: And based on what we've seen out of LLM right now, there are some serious limitations, as brilliant as they can be. Yeah.

9:26Tom Merritt: And honestly, neural networks try to learn the way a baby or any human learns, right? By reinforcement of neural connections and simulating that sort of thing. I honestly think that world models, whatever they end up meaning, aren't the thing that gets us to super intelligence or AGI or whatever you want to call it, because we don't have a single model running in our brains, right? We have a language cortex. Well, guess what? We've got a real good language cortex with LLMs. We have a visual cortex. Well, computer vision, getting that. I wonder if we also need to combine models and say, well, it's not just a world model.

10:07Tom Merritt: You need a world model and an LLM and a machine learning model and a deep learning model and a neural network. and then you combine all those things together. Because it's one of the ways that LLMs get better is by double checking them and using models playing off of each other. Yes. You know, I think daisy chaining models like that, and I know there's people working on that sort of thing, could be a path forward as well. So the one thing that I absolutely think that Lacuna's nailed is LLMs are going to continue to improve, but they are not the end of AI. They are a thing in AI. This is why I constantly try to use AI as a collective for all the models, not just a substitute for LLMs, because there are others that can be useful in other ways.

10:55Yep.

10:55Jason Howell: And I should also mention on one of my other podcasts, AI Inside, my co-host Jeff Jarvis and I have chatted with Jan LeCun on a previous episode. I can't remember the episode number, but if you just do a search for AI Inside Jan LeCun, you'll find the episode. It's on YouTube. It's on our show page. And he talks a lot about what his definition is and what his ideas are around world models. Obviously, he was still at Meta when we talked with him. Now he's not at Meta. Now he's doing his own thing. So who knows? Maybe I'll get the chance to talk to him again. Hopefully.

11:27Tom Merritt: Yeah, you should talk to him again. It would be great to get an updated look at what he's talking about.

11:34Jason Howell: Yeah, what that means now.

11:35Tom Merritt: Yeah, yeah. But if you want to find that, there are three N's in Jan LeCun. see i am not an llm so i can properly identify the number of ends yes good shot tom thank you that's how i prove it now there are there are you are in strawberry okay three dts is made possible by you the real human listener thanks to you brandon brooks johnny hernandez and high tech oki it's great to have all of you patrons along for every episode

12:11Jason Howell: Youth mental health is a complex challenge that requires comprehensive solutions. We must strengthen after-school programs. We must make digital literacy tools available in our schools. We must work with mental health professionals to support children. And we must empower mentors, educators, and parents to keep kids happy. Learn more about our commitment to finding lasting solutions at empowerourfuturecoalition.com slash solutions. Paid for by the coalition to empower our future. This episode is brought to you by Capital One. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one.

12:50Jason Howell: It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. When you manage procurement for multiple facilities, every order matters. But when it's for a hospital system, they matter even more. Grainger gets it and knows there's no time for managing multiple suppliers and no room for shipping delays.

13:24Jason Howell: That's why Grainger offers millions of products and fast, dependable delivery. So you can keep your facility stocked, safe, and running smoothly. Call 1-800-GRAINGER, click grainger.com, or just stop by. Granger, for the ones who get it done.

14:06Jason Howell: TommyJohn.com today for 25 % off your first order with code COMFORT. That's TommyJohn.com slash COMFORT. TommyJohn. Comfort perfected. Parles-tu français? Hablas espagnol? Parli italiano? If you've used Babbel, you would.

14:22Tom Merritt: Babbel's conversation-based technique teaches you useful words and phrases to get you speaking quickly about the things you actually talk about in the real world. With lessons handcrafted by over 200 language experts and voiced by real native speakers, Babbel is like having a private tutor in your pocket. Start speaking with Babbel today. Get up to 55 % off your Babbel subscription right now at babbel.com slash ACAST. Spelled B-A-B-B-E-L dot com slash ACAST. Rules and restrictions may apply.

14:57Jason Howell: All right, there is more we need to know today. So let's get to the briefs.

15:01Tom Merritt: Yes, indeed. Indeed. Amazon has had several small issues recently. We've talked about these. There was a six-hour disruption of the retail website. Didn't impact everybody. We didn't talk about that one because we don't talk about outages. But worth noting, somebody was able to jailbreak its chatbot to answer questions unrelated to shopping. I think it's called Rufus, the chatbot on Amazon. So they were able to break in there and get it to say things that were not related to Amazon retail. The thing we did talk about, two minor outages, a cost calculator in China went down for Amazon and an internal tool, I'm sorry, for AWS, an internal tool at AWS were both attributed to the use of coding assistance.

15:44Tom Merritt: Well, the news today is that Senior Vice President Dave Treadwell sent an email out to all the engineers at Amazon and among other things wrote, folks, as you likely know, the availability of the site and related infrastructure has not been good recently. It's been fine, right? It's not like Amazon's been down all the time, but not good enough for Dave Treadwell. So they had a meeting of engineers. They actually have this meeting regularly, but they said, look, this is a special one. Usually it's optional. This time it's not optional. We need everybody to be there. And they wanted to discuss these recent incidents.

16:20Tom Merritt: One of the contributing factors to the problems, according to the email, was the use of generative models for which best practices and safeguards are not yet fully established. So as a result, AI-assisted changes at Amazon must now be approved by senior engineers before deployment. They have to review the code. It's a code review of your AI-assisted changes. You can't just push that stuff out there. And just in time, Anthropic announced a new code review beta featuring CloudCode that can automate that for you. It uses multiple agents to analyze blocks of new code for bugs and other problems.

16:54Jason Howell: then you got to ask yourself do i trust claude's code review agent to catch all the bugs or do i

17:00Tom Merritt: still have to check after that code reviews must be reviewed by a senior senior engineer god did we ever get to a point to where it's fully trusted you know what i mean like yes i think we do it's all a question of when right yeah but but i i think you always need to have a human in the loop that takes responsibility for it yeah that's that's a really good point takes

17:22Jason Howell: the responsibility for it. And if something like this happens, is willing to speak to it and implement changes to address it.

17:29Tom Merritt: Because not every pull request gets reviewed. I mean, engineers, raise your hands if you've never approved a pull request without actually checking all the code. No, it happens all the time because you put together like, oh, I trust that person. This is a small bit of code. I think I can go ahead and push that one through code review. is approved, you kind of get a sense of when you need to pay close attention when you don't. I think that we will get there with these sorts of tools. And I think the Cloud Code Review, I don't mean to throw it under the bus, that is a great way of looking at it because it's using multiple agents to go over it multiple times.

18:09Jason Howell: It's kind of what we were talking about earlier, as far as using the LLMs to check each other and to make each other better.

18:15Tom Merritt: And I know a lot of coders who actually do that themselves. They write code and then they ask that agent or a different agent, hey, look at this and tell me where the bugs are, and they'll find it. They'll say like, oh, yeah, no, there's a bug in there that I didn't see before. So I think that's good. But in the end, you do have to have somebody who's taking responsibility and saying, okay, I'm going to push this out, and I don't expect it's going to cause a problem.

18:39Jason Howell: Oh, boy, if you're the person who pushes out the code because, oh, no, it's fine. I trust that person. And then Amazon's site goes down. It's got to suck to be that person.

18:47Tom Merritt: Oh, it always sucks to be that person. Yeah. That isn't even really what we're talking about where Amazon site goes down. It's just availability and bugs and things that are disrupting availability. When I say availability, that does mean the site went down, but it's not like the site went down. Like we've had those AWS outages that took out half the United States, you know, half the world's websites. Obviously that's going to be a horrible thing. But they don't even want minor stuff. They don't even want this like calculator to go down that shouldn't happen. And so they're saying, look, we have to have somebody who's paying attention.

19:19Tom Merritt: You can't just push your code out there because we're getting too many bugs in the system. And I think that's good. You should be judging on output, not just making people use these tools.

19:31Jason Howell: Right. And not just the goal is all of your productivity gains. All that goes out the window if your productivity is great, but you're producing crap. It's not good. Yeah. Google is bringing some new Gemini-powered functionality to Docs, Sheets, Slides, Drive. Honestly, I felt like some of this stuff already existed, but maybe they're just giving it a new name. The new features can pull some of your data directly from Gmail, Chat, throughout Drive, of course, the web. So that basically means that users can use a single chat bot experience to create your drafts, your slides, your sheets, just do it all into an integrated single chat.

20:08Jason Howell: So Docs is getting help me create and help me write. So making first drafts, refining those existing documents that you might have, the content. Docs also get, this is interesting, match writing style, which I think a lot of people use LLMs for right now, and match the format for tone alignment, for structure. especially if you're using a lot of different collaborators on a single thing. You can kind of get everybody synchronized into one style. Users can also mirror an existing template using live data from a Google account. So you've got a template file and you're like, I want all the files to kind of have the same look and feel.

20:45Jason Howell: You can do that. In Sheets, Gemini can spin up a fully formatted sheet based on a single prompt. And then there's fill with Gemini for auto-completing things like columns, live search data into your sheets, categories, summaries, that sort of stuff. Slides is going to get a generation and restyling functionality so that you can take a Dex theme and apply it globally, essentially. The team plans to bring full prompt-based creation to Slides in the future, though that is not there quite yet. And then Drive itself is getting its own AI overviews. So if you haven't had enough AI overviews, you're going to see it in Drive, along with an Ask Gemini and Drive Search for answering questions about all the contents.

21:29Jason Howell: And all these features are available first to Google AI Ultra and pro subscribers and kind of a limited beta, at least initially. And for those who don't like the AI stuff, this part's your reward for sticking through to the end. Google is bringing to its Google Photos the ability to disable the Ask Photos feature. So if you didn't like the upgraded AI search that was integrated into Google Photos, Well, you're going to be able to toggle that off and go to their classic search experience. It only took two years for them to give you that option.

21:59Tom Merritt: There you are.

22:00Jason Howell: Yeah.

22:03Tom Merritt: Well, first of all, a lot of these things you could do before, but Google is now making them easier to do in Drive. And they have also said they work now. I think before some of them were like, yeah, you can do that. Totally. And now they're designed to do that. But one of the ones that I want to try the most is making slides. Right now, you can only make a particular slide, but eventually you'll be able to say, like, make this whole presentation. And if you've tied it into the rest of your Google account, it can go find things in Gmail. It can find other documents and know what you're talking about.

22:37Tom Merritt: When you're like, I need you to make a presentation that pulls in these reports and emails from these people, and it will be able to make all that happen. Obviously, you still need to review all that at the end, like we were talking about before. But yes, I think this becomes a lot more practical for people because it's in the document that they're working on. In fact, I guess we don't have it here on my workspace plan because I'm not seeing those tools show up here yet. But I'm looking forward to those because it's not that I would want it to rewrite anything, not for us. That's too personal for the way we do the show.

23:15Tom Merritt: But doing some formatting thing within the doc, being able to create stuff within the doc without having to copy and paste and pull it into Gemini and stuff would be pretty handy.

23:24Jason Howell: Yeah. And depending on the prompt, being in a doc and knowing you've got something in your drive somewhere that has information that's referenced, just being able to say, get that information and throw it in here versus having to go and dig and stuff. I think the interesting thing for me about this is that I have friends who are outside of the daily tech sphere, let's say, and they have a workspace account. They see that little Gemini icon, and they keep hearing everything about this AI, and so they try it. And their opinion to me has been like, yeah, I tried Gemini in my drive, and it was not good.

23:56Jason Howell: And so I wonder for people like that, in that experience, does something like this, does it actually lead to a pathway of trying again? Or are they just kind of like, no, AI and Gemini and Docs sucks. Why would I want to do that? It's like, well, this might solve your problem, but you just have to be open to it.

24:13Tom Merritt: That stuff always happens, though, right? I remember people complaining about smartphones, like, yeah, those apps don't work, or you can't do this with it. And then eventually they do, and that permeates out. And like, oh yeah, no, that works now. You can actually, you can make that happen. So it's not, yeah, I feel like that's just one of those things of, if it ends up actually working, then everyone will start using it because word will get around like, oh no, you can do that with that now. That's powerful. Yeah, we'll see if that's the case in this.

24:45Jason Howell: Indeed.

24:46Tom Merritt: If you have found a way to make it work, why don't you tell us in our Discord? You can join a conversation in our Discord by linking a Patreon account. I pretty much go in there every morning when I start sitting down to work on DTS and say hi. So if you'd like to say hi back, link your Patreon account to Discord at patreon.com slash DTS.

25:12Jason Howell: This episode is brought to you by Capital One. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. Youth mental health is a complex challenge that requires comprehensive solutions. We must strengthen after-school programs.

25:49Jason Howell: We must make digital literacy tools available in our schools. We must work with mental health professionals to support children. And we must empower mentors, educators, and parents to keep kids happy. Learn more about our commitment to finding lasting solutions at empowerourfuturecoalition.com slash solutions. Paid for by the Coalition to Empower Our Future.

26:14Jason Howell: All right, we got some quick headlines. Going to make you look really smart at the dinner table tonight. Yes.

26:19Tom Merritt: When you lean over and say, well, dear, did you know Meta is acquiring Maltbook? And she'll say, yes, of course. I listen to DTNS2, dear. The AI agent social network will become part of Meta's super intelligence labs or MSL. No details other than that. But yeah, I always like it when Meta spends a lot of money on something I'm not sure they'll make good use of. But they'll either not make good use of it or turn it into something really boring.

26:48Jason Howell: Yeah, completely different. Interesting. NVIDIA, speaking of claws and molts and all that, NVIDIA is reportedly building Nemo Claw, an open source enterprise AI agent platform for deploying its own claw-like agents across the workforce, regardless of whatever the hardware that underlies your system is using.

27:10Tom Merritt: Oh, man. Everybody's doing it. All the Chinese cloud providers, now NVIDIA. Everybody's clawing. Got their claws out. Oppo and OnePlus are warning of impending price hikes for many of their existing smartphones, of course, because of rising costs of components. And those are set to kick in on March 16th. So buy before March 16th.

27:32Jason Howell: Yeah, you got a couple of days. Sonos announced the new sonos play it's a$299 era 100 class portable wi-fi bluetooth speaker with full sonos ecosystem support um definitely a lot more portable ip67 rated along with a cheaper mic free era 100 sl for

27:52Tom Merritt: 189 that only took a couple years for sonos to get back into hardware making good job yeah better late than never uh some cat images of the upcoming pixel 11 pro fold have leaked out reportedly showing a familiar design to the current fold, but with light changes to the camera bump and a slightly thinner body. Mostly it doesn't look that different though.

28:13Jason Howell: No, it's pretty much the same. Blue Sky founder Jay Graber has stepped down as CEO and is handing interim control to former Automatic CEO Tony Schneider so she can focus on building out the underlying technology.

28:27Tom Merritt: Yeah. If you're a Blue Sky fan, that's a big deal. Jay has been a huge part of that community for along. YouTube is expanding its AI deepfake detection to now include politicians, government officials, and the most important people in the world, journalists, so that you can now, if you're one of those classes, also verify your identity and then request videos be taken down.

28:51Jason Howell: Canada has reversed its earlier order that TikTok shut down its office in Canada and will let it stay in the country as long as it adopts stricter data security audits and stronger protection for minors.

29:03Tom Merritt: Yeah, Canada wasn't even going to shut down the service. They just weren't going to let them have an office in the country, but now they can even rent an office. That's nice. Adobe is rolling out its Photoshop Assistant, first announced at max last October. That's the one that lets you edit images and generate contact with a regular old language prompt.

29:21Jason Howell: Be curious to play around with that. Microsoft is adding phishing-resistant passkey sign-ins via Microsoft Intra and Windows Hello, starting as an opt-in public beta, or sorry, public preview later this month.

29:34Tom Merritt: Yeah, there's a story kicking around that if you jailbreak an Android device, Microsoft invalidates your entry keys if it detects that as part of its security. So watch out for that. Watch out. A US judge has temporarily ordered perplexity, basically issued an injunction, saying you can't use your agentic comet browser to make Amazon purchases for Prime subscribers until this whole court case is done. That'll go into effect starting next week.

30:03Jason Howell: Mm-hmm. All right, we end every episode of DTNS with some shared perspectives today. Josh has thoughts on folks calling the MacBook Neo a Chromebook killer.

30:13Tom Merritt: Ah, yes. I don't know that I ever called it a Chromebook killer, although I have said that it's meant to compete with Chromebooks and could compete with Chromebooks. A lot of people are saying it. Yeah, yeah. Many people are saying it. Josh writes, hey team, now that the MacBook Neo craze has settled down a bit, I wanted to get your opinion on the whole MacBook Neo being a Chromebook killer in schools. Here's my counter argument. Most school districts are still underfunded. A device double the price is just not feasible. I guess a lot of Chromebooks are 250 bucks. So, you know, with the educational district discount,$500 would be double that.

30:47Tom Merritt: Josh also says most of the Chromebooks are very repairable. Pretty much every component can be replaced from another Chromebook. I don't think the Neos are the least repairable Macs out there. Macs have gotten a little more repairable over the years, but certainly a lot of Chromebooks would be more repairable. So that's a decent point too. And then Josh says, why would you buy a$400 MacBook? He even gives them a discount to$400. Why would you buy a$400 MacBook just to run Chrome on it to use your school apps? Because everything is on Google Classroom anyway. That is an interesting point. You buy the Neo and then have to launch Chrome to get into Google Classroom.

31:22Tom Merritt: He says, I'm not saying it's not possible, but right now is probably not it. Reviews are out for the MacBook Neo as it's shipping on March 11th. We're recording on March 10th. They are all fairly rave, but they are basically saying this is a great entry-level laptop if you are comparing it to other laptops in that same price range. I didn't see a lot of the reviews calling it a Chromebook killer. So I think you're right, Josh. I don't think it's meant for the classroom Chromebook, although Apple would love it if people used it for that. And I think that's why they have that educational discount.

32:01Tom Merritt: It's more of, are you buying a laptop for a student, not a Chromebook, it's not part of a school thing, or are you buying a new laptop yourself and you want to spend$600, look at an equivalent HP Windows laptop, look at the MacBook Neo, and the reviewers are saying, you know what? The Neo holds up. It's got a better build quality overall. It feels firmer. It's got a full-featured OS, even though it's got an A18 Pro in there. It's got Mac OS running on it. You'd never know you were running on a different chip. It's snappy. It does everything well. And it's decently priced.

32:36Jason Howell: Yeah. I mean, your school is probably going to continue offering to the students who want a free-for-the-year laptop to do all their work the Chromebook, because I think I think what Josh lists out here are very sensible reasons why that would be. They are not – many of these schools are not flush with money to just go throw around and spend twice as much on each laptop for every student. Some school districts might choose to do that. But I would guess that the majority of them are going to stick with the Chromebooks. And honestly, it's kind of like – I liken it to a certain degree, like how it's been in our family with smartphones for both of my kids.

33:13Jason Howell: Because I am an Android guy. I've got shelves filled with Android devices, current and previous. And when it was time for them to get a smartphone, I was just kind of like, well, look, if you want an Android device, I've got one for free. I've got plenty of them. If you feel like you need an iPhone, I don't have those. You're going to have to buy it. You're going to have to spend your money to do it. And what did they both do? They saved up their money and they got it. And so I would imagine that the Neo might be something similar. It's like, well, this is the Apple thing. I want to stay in the universe.

33:44Jason Howell: all the cool kids are doing it you know that sort of thing whether that's right or not but yeah

33:49Tom Merritt: yeah but there's something to that especially because they have the nice pastel colors and all that so totally they're still a little pristine device in that price category uh and then also uh i mean there's reviews for all of the new apple stuff that's out right there but i think also notable is the benchmarks of the m5 max uh and the m5 pro uh out today and really saying like Like, yes, these are benchmarking better, and most people will notice iterative difference. But there's some big changes under the hood that have people saying, look, if you've got a MacBook that's more than three years old, you are going to notice a huge performance difference, even if you don't expect it.

34:29Tom Merritt: So I thought that was worth noting as well.

34:31Jason Howell: So you're saying that I need to get one? Yes. You just called me out, Tom. I did. You just called me out on the show. You did that on purpose, didn't you?

34:38Tom Merritt: I said, Jason, why haven't you bought a new M5-based MacBook? Throw that one away. responsibly. I don't need it anymore.

34:45Jason Howell: Effective immediately. I don't need it anymore.

34:47Tom Merritt: Wait until you buy the new one first, actually. That's a good point. Folks, if you've got great insights like Josh had, thank you, Josh. Send us an email. Our email address is feedback at dailytechnewsshow.com.

35:00Jason Howell: Thank you, Josh, for contributing to today's show. Thank you to everyone for being along for Daily Tech News Show. You can keep us in business. And you know how you do that? You become a patron at patreon.com slash DTNS. We'll see you tomorrow.

35:21Jason Howell: The DTNS family of podcasts, helping each other understand.

35:27Tom Merritt: Diamond Club hopes you have enjoyed this program.

35:33Jason Howell: This episode is brought to you by Capital One. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one. It's called Chat Concierge and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. Youth mental health is a complex challenge that requires comprehensive solutions. We must strengthen after-school programs.

36:10Jason Howell: We must make digital literacy tools available in our schools. We must work with mental health professionals to support children. And we must empower mentors, educators, and parents to keep kids happy. Learn more about our commitment to finding lasting solutions at empowerourfuturecoalition.com slash solutions. Paid for by the Coalition to Empower Our Future.