Limitless Podcast - Zuck's Open Source Playbook: Meta Muse Glimmer and 14-Page Manifesto

Episode Date: August 12, 2026

🔒 Check Out Our Sponsor: LEDGER AGENT STACK 🔒https://developers.ledger.com/docs/ai-tools/overview/?utm_source=Audio&utm_medium=Podcasts&utm_campaign=Limitless------🌌 LIMITLESS HQ ⬇️NEWSLETTER:    https://limitlessft.substack.com/FOLLOW ON X:   https://x.com/LimitlessFTSPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQAPPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890RSS FEED:           https://limitlessft.substack.com/------Let's unpack Meta’s return to open source AI and its latest laptop-compatible model, Muse Glimmer.Alongside its focus on cheaper, private on-device inference, we cover Meta’s broader AI strategy, including personalization, compute spending, competition on price, and its AI glasses and hardware efforts.------TIMESTAMPS0:00 Meta’s Open-Source Pivot1:43 Why Local AI Matters5:02 Meta’s Model Strategy6:39 Zuck’s Open-Weight Thesis10:43 Flooding the AI Market13:00 Ads, Data, and Free AI15:34 Cheaper Models Take Hold18:41 The Earnings and CapEx Bet20:33 Compute Is the Real Moat22:24 What If Meta Fails?23:48 Meta’s Bigger AI Vision------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠Josh works with Anthropic as a contractor. All views expressed are his own and do not represent Anthropic, its leadership, or its affiliates. Nothing in this episode is investment advice.

Meta has pivoted again for the third time and now back to an open source company. The story of meta has been pretty insane. Two years ago, they were totally open source, hellbent on wreaking havoc on the entire marketplace of closed source models. They were doing this through their model called Lama. And for those who weren't familiar, Lama was just kind of like this close to frontier open source model that was expected to change the world in a meaningful way. It did not go as Zuck foresaw.

And therefore they went and they closed off all of their model development and went closed source just like all of the other labs. Now, they're going back. They're transitioning back to open source. And as a result, they've spent a tremendous amount of money on this. So what happened in the last earnings report? After all the Kappex spend came in, Zuck just dropped a 14-page paper explaining exactly why they're making a re-entry into the open source world. That entry is defined by a singular model that we're going to talk about right now called Muse Glimmer. And Muse Glimmer is a really impressive model that fits on your laptop, runs locally, and I think it's going to shift the world of open source AI.

I think this is a great move from Zuck and Meta, and this is coming from probably the Meta's biggest bear in the past on this show. Yeah, definitely. We're professional haters. I should preface this with. I'm a professional hater of Zuck and Meta for the last six months, but I love to see this pivot.

And I'll explain why. Number one, Meta's entire philosophy around AI has always been open source. And his simple reasoning behind that is he believes, like, everyone should get access to this thing. right? If everyone has access to this thing, there'll be no centralized core power for AI. There's no AI Overlord going forwards. He then kind of pivoted on that, like you mentioned, and now he's back with not one, but two new open source models. So what are the models? Number one, the headline

is called Muse Glimmer. It's an open-weight, open-source 30 billion parameter model. Now, if you're not familiar with the sizes of these relative models, this is a pretty tiny model. It can actually fit and run on your MacBook. It's 20 gigabytes of data or RAM that it requires, which is significantly smaller than any other frontier model, which is like hundreds and hundreds of gigabytes. And the way that they were able to achieve this is they took a bigger model and they made it smaller. It's something called quantization. Now, why did they create this model? Well, he believes, or meta believes, that locally run private AI models should be the future of how AI models are dispersed amongst everyone. Right now, we pay subscriptions to get access to a model that lives in the

cloud that lives in a centralized company like Anthropic, like OpenAI. With this model, you can kind of run it locally, and that brings several different advantages that we've mentioned quite a few times on this show, though first one being, it's cheap as hell. The second one being, it can run on your private data, so it can become a lot more smarter and attuned to you specifically. And the third thing is, it's very efficient when it serves you answers and results to your different prompts. So I think it's a good model, but it's certainly not frontier in any way. Yeah, it's far from frontier and I imagine they're probably not really concerned about that. We mentioned this on previous episodes where there is this fight for the frontier, but now there are different frontiers. There is the frontier

of intelligence, which Anthropic and Open AI are kind of working on. Then there's the frontier of cost and efficiency per token. That's kind of where I see metaplacing themselves. In fact, They're fighting on multiple fronts at once, and it seems like Mews Glimmer is sitting at very base of that, because it is such a small model, is so quantized, it's able to actually run on a local machine. I think the fully capable largest version of this takes about 55 gigabytes of memory, which runs on my MacBook that I have sitting on my desk right now, and that's really impressive. This puts them in a race that not many other companies are competing on, which is just like

that edge-in-friens local compute thing that I think we can expect to see out of Apple Fair league soon. Like with the new Siri going to be running on local models, this is very much feels like a response to that where now for the first time you're able to run this free inference that's fairly capable. It has a somewhat large context window. I believe a couple hundred thousand tokens. I'm not sure the exact number, but a couple hundred thousand dollars tokens of context. It has, you know, pretty good outputs in terms of tokens per second. And it runs at a level that you would expect a frontier model would have ran at maybe, say, 18 months ago. So it's not going to do any crazy, difficult, complicated.

tasks, but for something that can run on your machine, you could use anywhere in the world at any given time, and you can trust it to manage all of your data. So, for example, if you have a machine runs a lot of sensitive information, you want to put health information through that, you might not want to give to a public-facing model. This is a really easy way of doing that in a way that we haven't really had before. And I think that's kind of step one in this new strategy that Zuckerberg's going for in, I guess you could call it like scorched Earth 2.0, where why do you open source models? Well, you want everyone to go off. build on them to remove margins away from these frontier labs. And this seems like the first

version of that that people are kind of excited about. And you see him apply this logic to quite a few of his models now. Like he also opened weight or he plans to open weight Mews Spark 1.2, which if you have been following along with the show, was a brand new model, Meta's Frontier model that they released, I think it was like a week and a half ago. And what made Mews Spark 1.2 really special is it's a really good coding model in the way that they've kind of engineered AI agents to kind of like replicate amongst themselves, take your code and go work in parallel.

So it's a really efficient way of working. And that's kind of like Meta's core thesis, I feel, with a lot of these models that they're releasing. It's like they're not trying to get the most intelligent model, but they're trying to produce the most cheapest and affordable model that still gets about 90 to 95% of the work done. And Zucker has said in many times over the last couple of months that this is Meta's core philosophy. They believe that locally private run, effective AI models

will eventually out-compete the frontier. Now, when you kind of compare these models, right, you're probably listening to this and you're thinking, okay, who cares? Meta has a model. I haven't used meta-AI intelligence. In God knows how long. Why should I use this model? Well, if you are using, say, the Anthropic or Open AI API and you're spending a couple thousand bucks a month, even a couple hundred bucks a month, if you're an enterprise that's spending 10 to hundreds of millions of dollars per year, this might be a model that might be more bang for your buck. And listen, you're not going to use it for everything. You're not going to get it to plan your entire company strategy.

You're not going to get it to build a brand new app from scratch and expect it to be the best thing. But you'll use it as a workhorse, maybe in conjunction with some of these more intelligent models. So it's great to see Zuck kind of like reprise Met's role in open source. If we remember, like, in the past it was Lama and then he kind of like pivoted and everyone was like kind of super angry about this. I'm glad that he's doing this. But the question is, like, why is he doing this, right? Like, you know, why would this make sense for META, which is a for-profit publicly traded company,

he sucks, still wants to make a lot of money. How does this make any sense at all? And he kind of dropped his entire thesis yesterday, when he announced that he's open-waiting a bunch of these models. He goes, I believe everyone should have access to super-intelligence. And I wrote a long piece about META's philosophy and values for building that world. And essentially, like, this future that he describes balances on like three, core pillars. Number one is he believes everyone should be individually empowered, aka get access to their

own private AI model and that it shouldn't be centered in specific frontier AI labs. Number two, he believes the thing that pushes humanity going forwards is invention. So humans' ability and intention to invent brand new things should drive who gets access to these different models, and that could be anyone and everyone. And the third thing is he just believes on a core balance of power, which is something ironically that adversaries over in China have been advocating. for a while now. So it's nice to see like a frontier American AI lab advocate for the same. It is, but it feels a little bit like I can't, I can't feel like this is a high conviction post because the sentiment has changed so much. It's not like he's like they started this way,

but then they changed this way when it wasn't working. And now that things are working, they're going to go back and start to believe this. It feels like they're just not really standing strong on morals over time. And this could be the wrong guess. But I mean, he's here saying basically the exact arguments that everyone's been arguing. for a very long time. It's like we don't trust a single monoculture to handle this. We don't believe that the future of humanity is going to be best served by two to three companies. We don't believe that any company who thinks that AI will eliminate most jobs would rush to build it faster than anyone else. And these are the arguments that we've heard for a really long time. And we just get these

echoed over and over again as the company starts to fit more of this mold. I mean, the essay was interesting. What I will be most interested in is just seeing what the actual downstream effects are of these new essays. And like, hopefully it seems like Zuck is getting more confident that they're going to figure this out, that they're going to actually be releasing models that can meaningfully shift the way people engage with AI. And this is probably like the starting gun. He's like, hey, new era, new company, new ethos, we're going to start running with this and see how it goes. And so far, it looks like they're kind of on their way. Like, things are pretty good. I was looking at Muse Glimmer. And I actually, I want to try this one. So my plan is to install this on my local

machine today and give it a go. Because one of the things that is unique about this one is for people who are still using OpenClaw, remember that thing. This actually is a very agendasic model that runs end-to-end for task completion. It uses agents. It has tool use. It has multi-step reasoning. It has multimodal input and output. It seems like a really strong and powerful model. Yeah, we can see here kind of how good it is at creating visuals, how good it is in benchmarks. And it's an actually usable model in a way that I don't think a lot of these others have been. So I'm taking this 14-page paper with a grain of salt. I read some of it. I didn't get my way through all of it. I know you said you read the entire thing. I'm excited for meta. Like, I want them to win. So I'm hoping this is the

new starting point for them to release better models to really get back in the game to compete because, I mean, yeah, duopolys are less than ideal. We want a lot of people competing for a lot of different parts of the pie. And if meta is going to come out here and do it and do so open source, That's great. And you've got to imagine Jensen's sitting over here pretty happen too, Mr. Open Source King, that now there's another company in the running for making really compelling open source models in the United States. I think it's actions over words, though. Zuck has talked this talk for a while, you're right, but he's also the guy that's releasing like the frontier open source models in America. You see all the other AI labs. Look, I'll give you an example here.

Satchi Nadala, a bunch of people signed that they're in favor of open source and open weight movements in general, but they haven't released any frontier open models themselves. So I look at actions over words and Zuck, yeah, he's flailed around quite a bit. And I've been a big critique of that, but he's come back. And I'm going to judge him by his actions going forward and we'll see what he actually uses this for. Now, the bare case for this is this is like an old school business move that he's making, which is let's flood the market with intelligence that is basically free. So we commoditize the paid subscriptions that other frontier labs are paying for.

And that brings down their value. That brings down their profit margins, their revenue. And it allows META more time to catch up. Now, is this a costly and risky endeavor? Yes, META has burned, what is it? Like, $30 billion in acquiring 150 staff to produce a model that is sub-frontier. If you were a shareholder, you're probably looking at this and thinking, Zuck, what are you doing?

But if you're not, if you're going to, you're going, but if you're going, you're you look at what he's been investing in outside of that. I think he's on track to spend, I think, $210 billion this year through the first half of next year on data center CAPEX. He is one of the largest holders of Nvidia GPUs. He's secured like one of the largest arsenals of Nvidia Rubens, which is their next-gen GPU architecture, second only to Elon Musk. You have to question, okay, what's he going to use all of this compute for? Well, he could rent it out, which is what SpaceX is doing and make a bag from that, but he's also probably going to train a bunch of frontier models.

And the biggest signal for whether he's actually going to be able to pull this off is model release caters. And in the last, I think, 90 days he's released two models, but he's going to release another two more over the next, I think it's like one to two months. I think you mentioned in some tweet thread. Yel-Moski is planning to do the same. So I'm seeing this trend where the companies that,

even though they're a little bit behind in Frontier Intelligence, who have the most compute, who have the most capital, are still able to kind of ship models iteratively and maybe respond to the market quicker than some of these other Frontier Labs. Now, I appreciate that that's optimistic thinking, but I do think Zuck is back in the game, and I'm excited to see what these models are like. I want to play around with this Muse Spark 1.2 and this new one that I can run on my laptop and see what fun things we can come up with. Yeah, I was trying to unpack the reasoning. I'm like, okay, why did he make this pivot again? It seems there's so much uncertainty in the strategy of meta so far. And there really does appear to be some

pretty compelling reasons that are kind of underlying this decision. Because when you think about meta, the company, I was like, okay, what does meta actually do? Meta's just an attention salesman. Their biggest part of the business is advertising and sales through that. All of their services are mostly free to use. You can pay for a blue checkmark if you want, but most of their users are free. And they have this tremendous user base. So they can sell a subscription to those people. But those people are not used to and not very probably likely excited to pay $20 a month to use an AI system that can be offered for free because you have to assume the average user doesn't need frontier class intelligence. So if the model is attention and if you can improve your core business through AI, then that's probably what they want to do. And that's probably why open source makes sense.

I mean, yeah, meta already announced it wants to use your AI chats to personalize the ads. and it wants to use the AI to build a large profile on you, to build a Josh.md, an eJazz.md, that has all of our preferences, all the things we care about, that gets updated in real time, and that has an increasing amount of ways in which you can collect context on you. And if you are trying to do this, then you really just want these models to be free and accessible and open

to collect as much data as possible to feel like there is no friction to getting in the game. And then also just to kind of like just chip away at the frontier labs margins, where if there's someone who is deciding between a $20 model or just going into their Instagram app and using this cool new chat bot that they have, like chances are a lot of people are going to choose to use the chat bot. And that puts them in a way, I guess it's kind of unique to them, where they're competing on edge inference in a way somewhat with Apple and what they will be doing with the new Siri, but they're also competing with Google in the ads business. If they're going to have just as much, if not more data than Google has when it comes to search queries and like think

about how custom the advertising can be, and not even advertising, but content creation in general. If meta understands what type of content you like, how you engage with certain types of content, it can serve this very custom profile that allows custom generated content just for you, all in the effort of selling you more ads. And I think this is the reality of meta is, as they've tried to pivot to hardware, they've done all these other exploratory things, it still just comes down to the core business is selling people ads. And it's like, understanding who their user is, pairing them with a seller, and collecting commission. off of those sales. And I think this is probably just a natural extension of that. And they're going to

continue to make these products more compelling and more capable through these very lightweight AI models. At the end of the day, Meta is obviously a for-profit company. They're trying to make money and ads, as you said, is like their main business. So they're definitely going to focus on that. But I also think in the wider market of AI, a bunch of things have shifted recently that we just can't ignore. The first one being, you don't need the most intelligent model to do anything and everything for you. I was catching up with a friend who is pretty high up at a top, I would say, Fortune 100 company. And I asked him, like, you know, how are you using AI these days? Like, do you use private models? Do you use the standard models from Anthropic and Open AI? And he said, I've shifted

completely to some of these open source models as well as just cheaper models in general. And I get it to do the bulk of my work whilst I use a Claude or GPD 5.6 to brainstorm kind of like the architecture of what I want to build. And I'm seeing this pattern reflected across a number of different enterprises. It's certainly been reported across social media like every single day. We're seeing the shift of companies thinking, hmm, I'm not going to just needlessly spend a bunch of money on Frontier intelligence. I certainly will for some things. But I'm, I want to explore some of the cheaper options, right? And you can look at the business models of the Frontier Labs themselves, right? I think over the last 90 days, eight models have been released.

Three of them alone from OpenAI, right? Why didn't they just release GPD 5.6 Sol? Why did they release Luna? Why did they release Terra? Why did they slash the price of Terra? Sorry, of Luna by 80% a month ago? Why did Anthropic yesterday, I think, announced that they're going to keep Claude Sonnet, I believe, at like the cheapest price ever indefinitely.

I think these companies are realizing that cheaper models are dead. definitely applicable and have an advantageous incentive for the customers that they're trying to sell to, whether it's enterprises or consumer customers. And you need more bang for your buck. Zuck realizes this and he's seizing the opportunity, but he's being more aggressive with the open source side of things. And I think Zuck has enough capital at META to be able to keep this train going for a while now. So I'm interested to see how what META and specifically SpaceX looks like over the next couple of months. I think SpaceX is also teasing their next model and they've released like two over the last like three weeks today, which should hopefully

be releasing. But yeah, I think there's something that they're going to pull off something cool. Yeah, I'm looking forward to it. I'm looking forward to all the agents that they're going to spawn from these new models because if you are working with agents, you know who we got to talk about. And that's Ledger, our sponsor of this episode, because they have this infrastructure for agents in which is kind of three parts. Agents propose, then the human approves, and then the ledger signer enforces. The ledger signer is this mostly comes in the form of a hardware wallet in which you can approve or deny. We talk a lot about security and how a lot of these agents can kind of go rogue and do things you may not want. Ledger offers an agent stack, which is a series of open source tools that

allows you to evaluate what these agents are doing across the way. You could tell it to do things like rebalance your wallet or adjust whatever parameters you want to adjust and then it will propose you approve and the ledger signer enforces. It works with cloud code, codex, all of the places that you work with AI on a daily basis. So thank you so much to Ledger for sponsoring this part of the episode. You can get their products at the link in the description below. EJS, you mentioned Kappexpend. Or you actually didn't mention Kappexpend. You mentioned strong earnings, but I'm going to mention Kapp Expend, which is their free cash flow. Like, we have to talk about the earnings quickly because, like, we do want to get a little bit economic here. Like, how viable is this

strategy? Their free cash flow this year, which means basically the money that comes in and out, is down 91% year over year. They're spending. They're spending big on a few key pillars. One of them is this open source bet, which does not have an immediate ROI because, I mean, again, they're offering these weights for free. The other is on their hardware platform.

One of the things that is still shocking to me, and I'm not sure that many people know about this, meta owns 70 to 80% of the entire AI glasses market. And this is not a big market, but this is a market that that is meaningful and that probably shouldn't be overlooked. This is done exclusively with a company named Luxottica. Together, they have more than tripled meta's AI glasses sales just in the year of 2025. We haven't gotten 26 numbers yet, but like that is kind of noteworthy.

It's interesting to see the floor of the KAPX guidance being raised too. In 2026, META is guiding for $145 billion of spend, which is just, this unfathomably large number. And when you think about the places that it's going to land, well, it's certainly open source. It's certainly hardware. But like, what else? Are they really just going to go all in on building out data centers and GPUs to make

larger models? Is that the end game? Because, like, I don't see them competing on the frontier. So do they really need that much compute in order to compete for open source, open weight models? I think so. I think the number one factor that determines how good your AI offering is is compute. at the end of the day. Elon's going to prove that with SpaceX, and I think Zuck is planning to do the

same thing. We saw NVIDIA or Jensen Huang announced yesterday that he secured an additional $500 billion of revenue and capital investment for future compute. We see OpenAI partnering with NVIDIA dollars. You see Anthropic securing deals with Riot yesterday with an old cryptographic mining company to secure even more compute. Compute is the ultimate indicator on whether you can create a better model. It is the one thing that is consistent across all competitors in the AI race. And I think Zuck believes if he amasses enough of it, if he secures enough allocation of frontier Nvidia GPUs, he'll be able to pull off and build better models. Now, whether the open source strategy plays out, again, it's that strategy that I think I mentioned earlier,

which is flood the market with commoditized intelligence, cut down or cut into revenue and margins of your competitors. Yeah, Metas, CapEx takes a beating. in the short term, but hopefully that plays better in the long term when you have per size agents that all run on Meta's infrastructure and consumer apps. Also, remember, meta has like 4 billion consumers right now. I was scrolling on reels yesterday, which I don't like to admit, but a ton of there are now meta glasses. I don't have you seen this, Josh, but like, it's just POV from like a bunch of different people. And I think that they don't have the best hardware, but they have one of the most amazing distributions ever. I think four billion, and then you compare

that to Apple, which has, I think, three point. 5 billion live Apple devices right now. It's probably a lot more. I think that data is outdated. They have a real competing mode here. And I look at Google, who has recently gone into negative cash flow, meta isn't quite there yet.

So you could argue that they're being somewhat conservative. They have other businesses that are earning a ton of money that can support this for a while. But they don't have infinite runway. And maybe we'll see meta raising external capital similar to how Intel and Google has over the last couple of months. I don't know. We'll see.

Yeah, and I was going to ask the question, whether or not this cap-expend is existential. Like, what happens if they fail again? Because this isn't their first time trying to get a frontier model that works really well and the first time did not go as planned. And then I was thinking, well, you know what? At least they have the GPUs. And they could just go resell those and make a great business out of it.

So there's this really weird part of the market now where it's very advantageous to just spend money on building data centers. And even if you can't figure out how to use them yourself, you just sell them off to the next best guy and you can make a tremendous amount of money from that infrastructure build out. So I think at the very least, you could find a lot of solace in that idea. In terms of users, like, that's a great point. I mean, 4 billion users is half of the planet. There's not many companies in the world that. I'm not sure if there's any companies in the world with that quite as many

users as meta or there's a very small handful of them. So if they're able to figure out how to, I mean, it weaponize is like a terrible word for this, but like able to leverage their AI tools to further create compelling offerings on their social media platforms, like that's a pretty big value unlock. So I'm excited. I'm cautiously optimistic on meta. Like, I'm still looking for signals that they are going to figure it out. But like you mentioned, when you're watching reels, a lot of it are coming from the glasses. It's this novel form of creating content. It's this cool new paradigm that every other company is working for. And right now, I mean, say what you want. Like, they are in the lead. It's maybe not the best piece of hardware in the world, but they have the best one available.

And we'll see what happens with the company. But that's kind of the update. Meta is going open sources again. The boys are back. And they're spending a lot of money and they are very clearly hellbent on making the best open source models that they can and implementing them across the entire product stack while continuing to iterate on their hardware products and weird science projects. Like remember the pupil dilation tracker and like the brainwave scanner that they're using to like understand how content affects you. And it's going to be a weird future with meta. But they're working on it nonetheless, full speed ahead. And yeah, we'll just continue to cover it as we go along this journey. Also, I will just say this. If you're listening to this episode and you are a

builder, there has never been a better time to be alive. Like all these tools, all these models, all these products that are being released over the last couple of months that are being increased in frequency are all for you to trial, test out. If it's too expensive for you to buy a Claude Mac subscription, play around with some of these open source models. If you don't think that the American Frontier open source models are good enough, play around with some of these open weight or open source Chinese models. There is no excuse to not be using these things. If you have access to it,

if you have a project or a hobby or an idea in your mind, go out there, try these things that we're talking about and honestly give us feedback because we love to hear from builders to see what you guys are actually doing with the thing. Maybe we're missing something and maybe you can kind of like shine a light

or something that we're completely missing. But we will be covering anything and everything on Limitless going forward. That is the end of this episode. you're all caught up on all of things meta-related. They've done a pivot. They've turned it around, and I'm excited to see where the company is in a couple months' time. That being said, wherever you're listening to us,

if it's on YouTube and you aren't subscribed, if the notifications aren't turned on, please do so. It helps us out massively. If you haven't left us a comment on Spotify, Apple Music, or even YouTube as well, please do so. We have been loving the comments, by the way. I think we've gotten in the realm of like, I don't know, it's like a thousand, two thousand plus comments

over the last like five episodes. Keep them coming. We love hearing from you guys. We respond to anything and everything as soon as we have like free time available. And then yeah, we have a newsletter that goes out to over 100,000 of you. We post twice a week. I think this week we're getting three times because Josh and I can't keep our gobs shot. But Josh, anything else.

Well, and most of all, don't forget to gift this to a friend. If there's someone else who you know that would benefit from watching this every single day, who wakes up in the morning, goes up with a little walk for their dog or like on their morning commute, that wants something to listen to 25 minutes long, that will get, sometimes a little bit longer, but that will get them up speed on everything. It's always much appreciated. It's how we grow the best. So as always, thank you so much for making it to the end of the episode.

Real ones only here at the conclusion. Thank you for watching, and we will see you in the next episode.