Limitless: An AI Podcast - THIS WEEK IN AI: Kimi K3, OpenAI's Alexa, Grok Stealing Code, Thinking Machines

Episode Date: July 17, 2026

Big week this week. Moonshot Labs’ dropped Kimi K3, Thinking Machines Labs’ Inkling, and reporting on OpenAI’s upcoming screen-free hardware device. We also discuss Grok Build, XAI, a new dictation tool, Elon Musk’s energy company purchase, and reports that DeepSeek is preparing an IPO.------🌌 LIMITLESS HQ ⬇️EMAIL US:           [e-mail]:    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/------TIMESTAMPS0:00 Kimi K3 Changes Everything8:11 Thinking Machines Finds Its Niche15:06 OpenAI’s Mystery Speaker Emerges21:15 Elon’s AI Moves and Missteps26:52 Dictation Gets Smarter29:44 DeepSeek Eyes an IPO------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.

Four years, the battle between open source and centralized AI models has been very one-sided. Anthropic open air has always dominated the field until this week where Moonshot Labs and AI lab out of China has finally released their next model, Kimi K3, and it is Fable Five worthy. Now, the craziest part about this is Chinese and open source models in general have been behind American frontier labs. They haven't been able to close the gap. Darry recently said that that gap is roughly around eight months having closed from about a year. Today, with Kimmy K3, that model has closed it to today. It is as good at one-shotting visual prompts as Fable 5 is. It's amazing and reasoning in general, and we're going to get into a bunch of cool demos that we see on our screen today.

Other big news in open source, thinking machine labs, which, funnily enough, raised the largest seed round ever, $2 billion headed by X OpenAI CTO Murati, released their first model called Inkling. It's small yet pretty mighty. We're going to get into that. And then Open AI themselves have revealed some secretive details about their new hardware device. If you've watched a show at all, you know, Josh and I are obsessed with this. We're excited to get into the details and show you what is coming soon. Yeah, so to start, we have a new king on the block.

And that comes in the form of Kimmy K3. I mean, this very much feels like, I wouldn't say the deep seek moment, but something somewhat familiar, in the sense that this is a really novel, leading edge open source model coming out of the Kimmy team. And it was paired with this really lovely launch video that actually, I was shocked that it was coming from this lab because this looks like something Apple would release as a product launch video. It was beautiful. The sound design was amazing. The visuals were great. And it's funny, they launched this before they even officially launched the model. In fact, at the time of recording, they haven't even officially stated that the model is live, but it's available.

It is available in the model picker for you to go and test it. Now, the model comes in the form of about a 2.8 trillion parameter model, which if you listen to the episode yesterday featuring GROC, that's about double what GROC 4.5 was in terms of parameter count. Based on the early reviews that I've seen, a lot of people have been using this model kind of early. It seems like this is pretty incredible. It seems like people have been able to do really impressive long horizon tasks. They've been able to do really impressive coding generation, gaming generation, 3D generation, and a lot of people are comparing this to close to a fable worthy model. And early testers, they put it above GPT 5.6 actually on coding and almost fableworthy at coding, which puts it in a

really solid spot because the pricing of this thing is going to be very low and the quality of it is incredibly high. And to our point yesterday, talking about the two frontiers, there is the intelligence frontier and there is the cost frontier. This is placing a new point on that Pareto curve of cost and intelligence at a part of the curve that I don't think anyone's ever made to before. This is like very much a frontier model in terms of cost per intelligence. It's really impressive. I think this tweet that I have over here summarizes it the best. He goes, the more I test K3, the more it feels like another deep seek R1 moment. It is often Fable level, maybe a little worse, but consistently better than 5.6. And that's a recurring theme across

a lot of different takes from early testers, that it's not quite as good as Fable. It's as good as Fable when it comes to certain things, but not everything. But it is certainly better or on par with GPT 5.6, which is a bold claim, obviously, because Open Air has invested billions and billions of dollars into their thing. Now, enough of us claiming theory. Let's talk about some actual examples. So on your screen right now, you're looking at the promotional video from the Moonshot Labs themselves. Now, what I'm going to shift to right now is Kimmy K3, one-shotting the video entirely from scratch. Whoa. Oh, that's really cool. Exactly. This is one single prompt. It was fed the promotional video

that Moonshot Labs themselves created from scratch using humans and all that kind of stuff, and it only took 25 minutes to do. Now, rumor, the cost to create this thing, Josh, actually, take a guess. How much did it cost to create this video? Oh, my God. Probably nothing, right? Because the cost per million tokens is 95 cents, I think. So what would this take?

Maybe $5 million or say $5 to $10, no more than $10? It costs under $10, which is just insane. Oh, that's so cool. It's insane. Now, like, if you have used products from anthropic or open air that are relevant to artifacts, the visual diagrammatics that they help you create, it, you know, it takes a while. It takes a couple of bits of prompting, sequential prompting. With Kimi, it seems that it could just one shot this completely. And it's not just one example I have. I have quite a few. If you look at this, you're looking at a 3D simulation. Now, you would be excused if you thought that this was Fable 5 worthy, because if you rewind a few, where we covered all of Fable Five's cool demos and stuff, we showed you a Hogwarts demo, a bunch of other things like that. It was of the same quality in par. Kimmy K3 can now compete on the same level. And it's not just for simulations. It's recreated Mario Car from scratch. Now granted, it looks like Minecraft, but, you know, what AI model doesn't do that? And you

could imagine with a few more prompts and like visual graphics added, it could get to, you know, maybe a, not a AAA, but like a double A type game vibe. So the point is, it's a very impressive model and it's physically accurate. It gets the physics of how things work, the mechanics of wings that you're seeing on your screen right now, how it works. It nails it. Now, where it falls or where it's a little lackluster, is general reasoning, writing, and coding in general. Now, coding for visual prompts, it's very good. Coding for hard specifically developer tasks, FABEL5 is still good.

but it matches the likes of GPD 5.6 Terra and Luna. Maybe not quite Sol, but Terra and Luna, and certainly Saul at least on medium effort. So if you're looking for a comparison as to whether you should use this model, it's pretty up there, and you should be surprised that they've been able to close the gap so quickly. The final thing I'll say is on cost. It is super cheap.

It is under a dollar input tokens, and I think it is a couple dollars output tokens. Again, none of this is officially confirmed. Just here we're getting leaks as we're recording this episode, but that would price it at the same value as Sonnet 5 or the Sonnet model from Claude themselves. So we're not even talking about opus level here. We're talking about like Sonnet cost, which is like cheap as chairs. Yeah, very impressive.

I mean, it's funny, as we cover this model, I think about the open claw community and people who use these long running agented tasks. I don't know where they are now. I haven't seen many of the open claw community publicly in a while. But this is such a dream model for them, where it is so cheap, it is so capable of these long running agendic tasks, it has all the visual understanding. It seems like a very good companion. And this seems like a legitimate kind of dent in the moat of these companies that are kind of holding the frontier of that Pareto curve in terms of cost, in terms of intelligence. So it's a really

impressive model coming out of China. I'd love to know how this came to be. What types of models were distilled? If any kind of, whose IP are you taking here? Who's stolen GPUs are you training this on? Listen, Josh, that's actually important context, right? You raise a good point because Moonshot Labs, the guys that created this model, have been accused by not just Anthropate, but also OpenAIA and Google, for distilling their frontier models. So we have no proof

whether they've done the same thing here, but maybe it can be assumed? I don't know. Yeah, I find it hard to believe that it's anything otherwise, then they've just distilled the frontier models and turn them into cheaper models that are more capable.

It seems like this is kind of well-known that they have smuggled GFUs into the country, that they have kind of nudge the numbers on how much the training costs were on a relative basis where they say it's a lot cheaper than it actually is. There's a lot of malpractice going on behind these models, which is why it's a little difficult to support them. But man, on just a pure open weight basis for them to publish a model like this, it's pretty impressive. And I think a lot of people are going to get a lot of use out of this. So I am going to go and try this. I'm going to check my model picker after this just to test.

Maybe we'll have a follow-up episode showing some of the demos, all the interesting things you can do with it. But if you are watching this, chances are it is live. It's available. Go try it out. and let us know what you think. That brings us to our second topic of the day. And this one I am particularly excited about because, EJ, if you remember, 18 months ago, after leaving Open AI, the co-founder of Open AI, Murmurati,

she left and she went to go start a company named Thinking Machines. She is one of many co-founders who have left the company, unfortunately, for better or worse, and would often try to do her own thing. During this process, she raised the largest seed round in venture capital history. Now, seat rounds generally speaking, I remember back in the day a seed round of like a couple million dollars. If you raise like $5 to $10 million, that's a remarkable seed round. This was $2 billion. And for those not familiar, the seed round is the first money into the company.

You have an idea and you say, I'm going to go build this idea, and an investor gives you some money to do it. She raised $2 billion at a $12 billion valuation right off the bat from companies like Andrewson-Harwitz, Nvidia, AMD, and Jane Street. These are the largest ones in the game. So for the last 18 or so months, we have been asking herself. the question, well, what are they actually working on? Two billion dollars is a tremendous amount of money to go off and build something. And Mira and the team are clearly very talented because they were one of the day ones at Open AI who were building the GPT, early GPT days. And as of today,

we finally have an answer. And it comes in the form of this model named ship, Shipling, Inkling. What is this? No, it's, what did they actually release? To be fair, it doesn't roll off the tongue, but it's called Inkling. And it is a model that falls. Yeah, yeah, it's a model that falls just under a trillion parameters in size, so relatively small, given the landscape today. But would it surprise you if I said it's a pretty unremarkable model that is remarkable in a very specific way that not many people see? So let me explain what I mean by that. It's trained on about 45 trillion tokens of text, image, video, and audio. Now, the reason why that is pretty cool is that it's completely omnipodal. It's like a multimodal model that you can.

can send text to, you can send audio to, you can speak to it, doesn't interrupt you. It's that type of model that is very intuitive. It feels very human. And Miramorati released a really interesting blog post earlier this week, which described her vision for thinking machine labs. It's not to be the frontier LLM, it's to be the frontier companion for humans. And so she's building these models in a very specific way. So when we look at the model itself, it was trained in under nine months. It is quite good at reasoning. It is okay at coding. It is nowhere near Fable 5 worthy, but where it gets super powerful is if you take their model and then you fine tune it for your own specific task department or team. So what that looks like is, and there's a live example of this actually

happening right now, where you can take this model and let's say you operate like one of the world's top trading hedge funds, right? You could feed it data, you could run it locally, by the way, feed it data, all encrypted, train the model to understand how your team, company operates, what your investment thesis are, feed it maybe even proprietary information, because, by the way, this model is completely open weight. So, you know, open sources come back to America, well done, Miramarati. And you can hyper-focus this model on your daily trading activities if you are in that hedge fund position. Now, if you don't believe me, there's a live example of this actually happening from none other than thinking machines themselves. They're working reportedly pretty

closely with Bridgewater Capital, which is one of the biggest funds in the entire world, using their Tinker API to fine-tune models. So the long story short is, this model isn't meant to be remarkable on its own. It's meant to be taken by you, whether you're a startup founder or a hobbyist or whatever that might be, and fine-tune it to do the specific task that you can do. If you end up doing that, you end up with a more powerful model that is less generalized and more specific than a classic anthropic or open AI model, and it is way, way cheaper. if you use any of these other generalized models. And as we've spoken about on yesterday's episode and a lot of episodes recently, actually, cheaper models might actually end up being the models

that are used more than the intelligent models themselves. Yeah, I think this is a really interesting angle. And I'm excited to see that this is the way that they're going. They're not trying to compete with the frontier labs. In fact, to use their own words, they said, it's not the strongest overall model available today, open or closed. So they're very much aware that they are not on the frontier. The interesting thing is addressing this common concern that we see talked about a lot. If you remember, there's this viral clip on CNBC a couple weeks ago with Palantir CEO Alex Carp. And he was describing how close frontier tools are too expensive and they have very somewhat murky IP protection where you kind of feed these models your context and then perhaps they go off and use them to train new models. And what thinking machines is doing is basically building this infrastructure for you to build and own the IP custom to you.

And I think this is the interesting thing where when you think about monetization for a company like this, they raise $2 billion. Now they're just giving away their technology for free. That's the reality to an extent because this model is open source. You can actually go and download it and use it for your own, but they have this platform called Tinker that's meant to hyper-optimized these models specific for your use case with a much less amount of data than traditional models would. So traditionally, these models need like terabytes of data in order to go through this reinforcement learning process and to become hyper-specificly good at a certain task.

This model, and using this Tinker product, you're actually able to feed it a lot less context to get a very specific outcome. So if you want to train it to be a financial wizard like Blackwater was perhaps doing, or if you want to train it to be a professional private chef, if that's what you're doing, it requires a lot less context and you could kind of mold it to the form that you want specific to you. And I think that's a really powerful thing. This is mostly focused on consumer and enterprise, but very specific use cases for it. And I think this is a novel angle in the sense that a lot of the frontier labs, their general purpose.

You go to them with any question you want, and they will answer it to the best of their abilities, is generally better than everyone else. But this lab will allow you to do a very specific answer with a very low cost using open weight models without any potential leaking of any of your intellectual property. And I think that's a pretty novel use case and really cool to see out of the thinking machines team. I think an obvious question to come from this as well

is how were they able to achieve this? Well, I'm pleased to say that I think the Chinese are getting a little taste of their own medicine. What I'm showing on the screen here is, In order to train this model, thinking machines used Deepseek's V3 architecture. Now, if you've heard of Deepseek before, it comes from an AI lab in China, funnily enough, a hedge fund that built their Chinese model called Deepseek, and it uses this architectural mixture of experts. I'm not going to bore you with the details, but basically you can have a really large model

and only utilize the necessary parts of the model when you're answering a prompt. It's way more efficient, it's way more cheaper, and it allows you to customize it way more. That's how they were able to pull this off. And it's kind of nice to see this reverse distillation. And I say that, justingly, because obviously all the Chinese models are open source. But moving on, Miramorati's former boss. My favorite topic. Miramorati's former boss has been up to some good and some no good. If you watch our previous episode earlier this week, you'll know that Sam's in a bit of hot water right now.

The big boys at Apple are suing Sam and Open AI for stealing hardware, trade secrets to build their new AI hardware device. And a question on Josh and I's mind was, what earth is this device? Josh, maybe we have some light now. Can you give us some details? Oh my God. How many times have we spoken about this on the show? We got to be approaching like double digits now. This is probably the 10th time we've talked about this freaking device. We still don't know what it looks like. I'm so desperate to know I've never wanted to hold a product in my hand so bad in my entire life. Then I do want to hold this Johnny I've open AI collab device. And it of course has not come without its troubles. They are currently experiencing lawsuits.

I'm wondering if those lawsuits have impacted the trajectory of the suite and family of products that they're going to launch. But we do seem to have somewhat of a close confirmation to what the very first device in this family of devices will look like and when we'll actually be able to get our hands on it. Now, Mark Herman, who some people might know as the Apple leaker, he is kind of the guy who you go to when you want to know about leaks. He has very reliable sources. He is generally right when he publishes these things for his publication. I think he works with Bloomberg. I'm pretty sure that's where he publishes all of these. And we have an idea of what this looks like. And drum roll please, it is coming in a form of a speaker, which is interesting.

It is described as a screen-free smart speaker that will act as a new type of home computer for the AI era. So it's screenless, which is important to note, but it's certainly not senseless. The idea is that it will have a camera plus these environmental sense. It's probably like light. It'll have microphones. It'll have a rechargeable battery, which I found interesting because that means that this will be a portable device. And then I just as I was asking you about this yesterday is, does that mean that it's going to be something that fits in my pocket? Like, is this the puck? Or is this going to be, I think, of the Amazon Echo, like that little block that sits on your desk and you could kind of carry it from room to room? I'm kind of now

hypothesizing of what this is going to look like. Well, the secret is revealed in this tweet from Mark German and none themselves. He says as one of his first descriptors of this device that it's a mobile device. And the idea, if you open up this article and read through it, is he talks about this device being somewhat portable. It's meant to live in your home. It's meant to indressed and listen to everything that is said, the conversations that you have, you can also speak to it. But the goal, the stated goal of what Open Air is trying to build here, is a companion, a companion that can sit by your side that you can speak to, that sounds very human, that understands the work or tasks or goals,

whatever you're trying to achieve for that day and over a set number of days, and it can work productively with you. Now, a big question for me is, how different is this from just opening up the chat GPT app on my phone and speaking to it? And I think I've answered my own question, which is it is quite different.

There is a big delta between having a device that is omnipresent, that is ambient, that listens to everything you say that can, like, chime in when necessary, like a real human does, like a companion or an advisor or a mentor desk, versus having something that is a little more archaic manual that you have to kind of like trigger when you think of it. And I think this kind of like jump going from kind of like manual devices

where you have to click button, swipe and go to the specific app versus having this omnipresent artificial intern. It's funny, it's like I'm describing like the Terminator plot right now and I'm listening to myself and I'm like, oh God, like maybe this isn't good for us. But I get that that's where we are moving to. Now, of course, all of this is really exciting to see, but if there is a massive Apple lawsuit looming over their heads, there are grounds for this potentially being delayed.

Now, the target date is early 27, specifically February. Sam wants to ship around 100 million units initially. Josh and I are probably going to buy up 50 million of those units, but the point is that this is going to be really in demand, but it might get delayed to later on in the year if Apple has grounds for the them stealing secrets. I really hope not. The reports look like it's going to come out at around $200 to $300, and you have to imagine if they're building $100 million of these, a couple hundred million of these, it's going to be a fairly simple design. There won't be a ton of moving parts.

There won't be a lot of moving pieces. This will kind of be a solid machine, is my assumption. If I know anything from my obsession of Johnny A.I. His entire goal is to reverse the damage that has been done through the smartphone era in the sense that smartphones have taken so much of our lives away from reality and they've placed them into this digital world. And I know he is on a mission hell-bent on reversing that. And part of that is removing the screen, but the second part of that is trying to make the experience more human. And I think we got an early release version and look at this with the new chat model from chat GPT on how conversational it is, how it feels very warm and friendly. And I expect this to very much be an extension

of that. In a way, it's kind of converging on the voice models and Miramaradi, Miramirati's like tinkering model. where it will kind of build up this context window about you. It will be able to be your personal assistant and allow you to remove yourself away from the matrix that he has created such a large amount of through his hardware devices almost unintentionally. And to me, that's super exciting.

And I cannot wait to get my hands on it in early 2027. So we will be keeping our eyes peeled, any more rumors, you know we're going to talk about it because we can't help but talk about it. This is like the 10 time. But that is the update on the opening eye device coming soon. I will say that Apple is planning to at least announce their AI hardware devices, plural, by the end of this year. So they might just be doing this whole lawsuit thing just to front run them.

And I know that Apple is working on, I believe, some kind of a disc pendant, airports with cameras, as well as some smart glasses. So I'm excited to see these two honestly go out of it because the end benefit is us. We get to buy a bunch of cool gadgets and hopefully wear them every day. Moving on to the next topic, Elon Musk, surprise, surprise, is in a bit of trouble. Now, he has this product called GROC build, which is basically their ClaudeCodeC competitor. It is GROC used specifically for coding and building different apps. And some people leaked that XAI's GROC build CLI, so their API basically or their front-end interface, was taking people's code, uploading it to private servers for some.

SpaceX AI and then using that data to train new models. Now, if you've ever been in this model game for a while, personal data, privacy and encryption are incredibly important anthropic and I take this very, very seriously because of course you want to make sure that they're not taking your personal information that you're imparting to these AI models and using it for nefarious ways or taking advantage of them. GROC or rather SpaceX AI and Elon were caught doing this and they had to do a lot of damage control, which ended up being a few days later, them open sourcing the entire GROC bill. They apologized and basically said, zero data retention policies are super important where we don't retain any of your data. And if we do, we will be up front. You can

turn this off in settings, but it sounded like it was on by default, which is the warning part. But in order to retaliate, they've said, we've open source this entire product, take it, fine tune it, do whatever you want with it, and see the code yourself and see that we're not taking any of your stuff. So it's a it's a bit of a roller coaster for them. I'm glad they're being able to do damage control, especially as like Elon has like herding insults to his enemy at Sam and open AI, which arguably haven't done this yet. I respect it. It feels like the early days of Facebook, the move fast and break things where they're just going balls to the wall, full speed. If they make mistake, they correct it. And in fact, I feel like they really even over indexed on correcting it.

They open source everything. They had zero data retention policies. They really went like fully locked down almost Apple status where now there is zero data that is going to cross the line. I think this is going to prove to be a very unique advantage. When you think about the Apple Edge, a lot of it has to do with guaranteed privacy of data. And if Grockett's going to be the only person who is going to provide that guarantee, that is a huge benefit. Because remember, they have the cost per token down about as low as it gets. They're on the frontier in terms of cost per intelligence.

And now if they have privacy baked in, that's a very compelling product that I think is noteworthy to watch. There is a second bit of news about Elon, and it is about Elon, not about his company, SpaceX, not about Tesla, but about Elon Musk, the individual who just acquired a company himself for an estimated $1 billion, which I found interesting. I was like, why is Elon the person acquiring this company? And I mean, surprise, surprise, it's an energy company named APR Energy. And they are a Jacksonville fleet of trailer-mounted mobile gas and diesel turbines. Why on Earth would he want to buy this? Well, it's because they have a gigawatt plus of capacity. And to buy a gigawatt of capacity for a billion dollars, turns out that's a pretty good deal. And as we know, GROC has these Memphis data

centers. They have Colossus 1. They have Colossus 2. They need to spin them up. And one of those big constraints has been power. And we've seen how much in-demand power has been in the ability to deploy it quickly through companies like Blume Energy. We've talked about Blume Energy a lot on the show, their stock has gone basically vertical recently because they are one of the few people who can modually deploy energy to these data centers quickly. This company that Elon bought now is able to stand up these projects in about 15 to 30 days, which is really exciting because these grid interconnection cues to tap into the mainline grid normally take a couple of years. So this really shrinks that time to get energy to the data center by like a order of

magnitude. And I think that's a really big deal. So this is exciting. It was kind of on the radar. It didn't get published anywhere. It was just kind of discovered through these filings that Elon had made the purchase. And I think it's a testament to how fast they are moving. They need power. Elon's going out. He said, I'll get you a gigawatt myself. He bought it. They're going to deploy it in 15 to 30 days and they're going to bring way more GPUs online. And that's a testament to the velocity in SpaceX AI. So again, really bullish news. Keep an eye out on them. They're making their way up to the frontier very quickly. The problem Elon's solving with this is he has way too many GPUs to the tune of like one to 1.5 million of all the latest Nvidia GPUs, but he has no means of bringing them

online. The actual wait time to get your GPUs hooked up to the energy grid is around five years right now. And New York City, or rather New York State, this week became the first state to pause or ban data center development. And it's following a trend of a few other states. If this continues to happen, workarounds like this, like acquiring companies like this, to get portable gas turbines and powering your GPUs, maybe the only way to work around it. And the important part about this,

why is this important at all, is you need compute at the end of the day to build the best models. We've seen meta-release Muse Spark 1.1, which isn't quite frontier, but is pretty good. And you've seen GROC 4.5 be able to compete with the big boys unexpectedly. And the way they've been able to do this is because they've acquired so much compute.

but none of your GPUs make sense. They're worth those pieces of metal if you can't power them up in the first place. One thing I will say is we've spoken about Bloom Energy a lot on this show. Leopold, Aschenbrenner, who runs one of the biggest AI funds, 24-year-old kid, has made a big investment in Bloom Energy. This is their direct competitor. And Elon just went out and looked at Blume Energy and thought, I don't want to have to deal with this.

I'll just acquire the company themselves and integrate them into what I'm doing here at SpaceX AI. And I think it's an incredibly smart move. Now, moving on, there is a popular product that I believe both of us like to use quite a lot called Whisperflow. And the way that that works is instead of typing prompts, it takes a lot of time. You can tap a button, hold it down on your laptop or on your computer, and you can speak into a microphone or speak to your computer and all your words are recorded. And it has been a major unlock for me. Voice to AI is just the way forwards and I honestly can't see anywhere else where it. doesn't result in that. Now, this guy, Farza, released a new product from his company that is called

Screen Aware Dictation. Now, one problem that I have when I hold down the button in Whisperflow and I speak into it and it records exactly everything that I say is, I'm also thinking in real time. I'm like, oh, I want to say something in a tone that is kind but not too assertive. And I want to talk about this topic. I don't know the definition of this topic, but like, I need to do some research on that. And if I was using Whisperflow, it would record everything I say verbatim and put it in that email. With this new feature, it listens to you. It understands when you don't understand something. It goes away, does the research, comes back and formulates the reply or the response that you want. And it's this massive jump up where you now have this feature which kind of like simultaneously

listens to what you want to say exactly, but also understands when you don't want to say something or what you don't know, goes away, does the research, pulls all the context from your personal files that it's connected to, and then comes up with the perfect response.

You can literally say, hey, I want to create a Google slide deck that is to do with this document that I have on another screen or another tap. It'll go away, research that, come back, and put that into a very well-presented slide deck. So it's a really amazing product or feature. I've been using it a bit this morning.

I think he got overloaded since he announced this literally under 24 hours ago and it's got like 3 million views. But I don't know. I just want to highlight this feature because I think it's really cool. Cool. Yeah, I'm excited to give it a shot. This feels like something that like Open AI can release tomorrow and kind of knock off the grading call today. I think this is a thematically interesting project though because, again, everyone is really trying to collect the context. If you think about what thinking machines released, they are just collecting context so that they can give you a custom tailored experience.

the Open AI speaker, that is probably going to be coming out. Which job is just to collect context about you and to then customize your experience. Same thing with this. So much of the AI story now is kind of pivoting to context because the more an AI understands about you understands about your preferences, the better it can serve you. And this is another interesting example of it. I saw there's a few other kind of rapid fire points that we can get through. The next one being Deepseek is preparing an IPO, which is kind of crazy.

That's going to be interesting to talk about. Are there any interesting talking points with this IPO? or is this kind of a rumor? You as you know? It's kind of an unofficial confirmed thing at this point. A lot of the main investors have kind of been like referring to this IPO and referring to this new raise. The reason why this is like compelling news is they literally raised, I think,

$15 billion a week and a half ago. So as soon as that round closed, they've raised yet another round that values them at like almost one and a half X higher than they were before. And now they're touting the IPO rumor, which, you know, they're going to join Chippoo and a bunch of other Chinese AI companies in that public race.

Another interesting fact is the founder, Liang Wen Feng of Deep Seek, owns around, I think it's 94% of the company, which means that his net worth is roughly around $26 billion, making him one of the most wealthiest AI CEOs out there. This random guy from China who was a hedge quant fund trader literally about a year and a half ago. It's insane.

Good for him, man. He was doing well. And then with that, I think, that is pretty much all the news for this week. That's it. We got it all covered. It's been an eventful week. This has been, there's a lot going on.

It's amazing. I'm always surprised at how every week we can sit down here and somehow managed to find so many topics that are interesting to talk about. The just frontier and the world of AI and technology is moving so rapidly. So if you've made it this far through the episode, thank you for watching. If you've listened to the three prior episodes this week, which I encourage you to do so, we had some pretty interesting ones. One was about the GPT 5.6 release.

The other one was about Open AI getting super. sued by Apple. And the third being about why the price war actually matters. And we referenced that episode quite a bit about how the new frontier is kind of being battled on price instead of intelligence. So those were all really great episodes. If you enjoy this one, please do not forget to share with your friends. And if you did make it through all of that, congratulations, you're caught up. You can go touch grass, enjoy your weekend. And yeah, any final parting thoughts before we head out for the weekend? I want to know what people are doing on the weekend. Are you using these models? What are we tinkering? What are we tinkering?

with, you know, pun very much intended. We, I think 90% of our content this week has been about new models. There's open air releasing literally three different models of three different modes. We've got all these new open source models and they're super cheap to you. So I'm wondering what you folks listen to this are tinkering with. What are you doing? What are you building? What are you doing with these models? Are they even useful?

Are you still just kind of like ignoring all of this and using Fable 5 or GBT 5.6? Let us know in the comments, DM us. And the final thing I'll say is we are on the lookout. for sponsors. Limeless has spun out. We are officially on our own. We keep the lights on through our own and we would love to partner with someone that has a product or service that they feel incredibly passionate about and want to share with our audience of users who love to use these different things. So if that sounds like you or if there's a friend that sounds like that, please reach out to them, reach out to us. Our email is in the description below. You can DM us on X or wherever we are

and yeah, we'll see you on the next one. You're all caught up. Awesome. Well, have a great weekend, everyone. See you next time.