Limitless Podcast - Apple WWDC 2026: Finally Delivering on AI Promises

Episode Date: June 9, 2026

This year, we got what we came for. Apple’s WWDC 2026 focused on improving AI, especially Siri’s new capabilities, on-device models, and privacy-centered request handling. With new AI features in Photos, Apple Maps’ 3D Gaussian splats, and Apple’s broader position in consumer AI, maybe Apple AI will finally come through.------🌌 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/------TIMESTAMPS0:00 Apple WWDC 20262:15 Siri Gets Context7:45 Developer AI Unlock11:35 New Practical Features14:30 Under the Hood21:29 Photos and Maps25:51 Future Hardware Plans30:21 Apple’s Long-Term Bet31:21 Tim Cook’s Legacy------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

Two years ago, Tim Cook got on stage at Apple's Worldwide Developer Conference and announced what I believe to be the most amazing set of features in the history of Apple, Apple Intelligence. It turned out to be the most disappointing WWDC of all time. It was a complete and total failure so much so that they are still currently dealing with lawsuits around false advertising for how big of a flop Apple Intelligence was. They promised us the world, they delivered absolutely nothing. In fact, the iPhone that they built for that year in particular for Apple,

intelligence is not supported by Apple Intelligence. This sounds like a disaster, but just yesterday Apple had an opportunity to fix this wrong, to write the things that they were not able to do the first time. And I've got to say, I'm very impressed. I think this year's WWDC left a significantly different taste in my mouth than the last one, because this year was all about AI, and for the first time ever, we have features that are actually going to work. I remember filming last year's WWDC episode, and I was just being thwarted. already disappointed with just the lack of AI stuff. Like, they just didn't pay it any kind of attention, and they delayed it yet another year.

Up until yesterday, actually, Siri couldn't even tell you which month we were in. And that was ironically one of the demos that they showed yesterday. Anyway, a slew of new AI updates. The two flagship headline updates from Apple is Siri is now Siri AI, powered kind of by Google's Gemini model and also by Apple's own foundational model, which brings me to the second major headline announcement. We've got four to five brand new Apple foundational models, which can run both locally on your device, as well as through their own private compute system. We're going to get into a lot of the details about how this actually works. But my main takeaway from yesterday's

WWDC is that Apple is finally taking AI seriously. And while some of the feature updates that we're going to go through may seem like the kind of basic. Collectively, they actually make your life way, way better. And you listen to this, if you have any kind of an Apple device, an iPad, an iPhone, a MacBook Pro, whatever it might be, you can use a net new feature once these features actually roll out by the end of the year, which will change the way that you work. And it's done over or across 3.5 billion active Apple devices. So yesterday they kind of oriented this presentation around three key pillars. The first was stability. The second was kind of how you can use it with your children and security and privacy. The third was AI. We all want to talk.

about AI, we want to hear about AI, let's get into the AI first. It's now Siri AI, like you mentioned. And what we're seeing on screen is unique and novel. It's a demo that's in real time that actually works. And this was something that I found uniquely interesting for this presentation in general. Generally, when Apple has these things, they have these very pre-produced demos that look very good, that operate very well. When we saw them initially released two years ago, they looked incredible. The difference now is you'll notice that there's no cuts happening in this demo. They're not doing it for speed, they're not doing it for time, they're doing this for proof that it actually works. So what we're seeing on screen is someone from the Apple team who is actually asking it to

search as text messages and uncover a text message about a specific song and then go off and play that song. And for the first time, ever, Siri has the context. So what we're seeing now is he's showing it a screen from Instagram and Siri can see the content of your screen. It could understand where that location is and then it could route you there on your maps. And this is something unique and novel only to Apple, where it has the full permissions of your phone. It could see what's on your device. It can actually engage with your device. And it does everything that I think we imagined and always hoped it would, which is just have full context of your entire life and be able to turn that into some useful information and useful actions.

So for me, this was exciting because this was kind of definitive proof that it actually worked. And what we'll notice as we go through the demos from yesterday is that this is actually available for developers right now. So people have been testing it. They have shown that it worked. And there's a lot of cool features that they actually released. I think the concept that we're describing here is basically like grounded context. We've spoken about personal AI agents a lot on this show.

And the reason why we've done that is a lot of people think that the best AI features, the best AI models is only achievable if the model is actually larger, the bigger it gets, the more compute that it has. Whereas in reality, you can have a smaller model. that is trained on your own personal data that you don't necessarily want to share with chat GBT or Claude, and it could do this so in such a private way that you can create a much more personalized AI experience.

And so some examples that we're about to run through that is summarized by Marquez in this tweet is basically, you know, you can message Siri and say, hey, do you remember that restaurant that my friend was talking about about a month and a half ago? I think it was somewhere in my neighborhood and it pulls it up. Or you could say, hey, could you look at my recent,

location history and let me know what that place was there, saw that beautiful sunset, and it'll be able to tell you. And it's these little nuances that you can't really achieve in chat chbt or any kind of LLM chatbot, simply because it's not grounded in any particular context. Now, that all sounds great, but let's walk through some examples specifically. So you mentioned that Siri A can react to things that are on the screen.

This is a classic case where I'm updating my iOS, right? And I don't want to stare at the iOS screen to see when it's finished updating. So I say, hey, Siri, can you set a timer for when this finishes? And it knows what you're talking about. It sees the amount of time remaining and it sets that specific timer, which I thought was pretty cool. And then we've got another example here where it's like, when was my flight last month? And it pulls up, your flight manifesto goes through your email app or your calendar app or whatever that might be. And it says, I think that's your flight to Budapest. It was a British Airways flight. You left it this time and you arrived

at this other time. And then there's this other one which I kind of cracked up at. where you get a text from someone and it says, you know, have you heard of this plant called Calithia? And it's basically like a tropical house plant. And Siri summarizes it basically saying the exact same thing that's in the text. Now, the point I'm making here is that it's not exactly perfect. There's some kinks that we need to work out. But I can already see myself using this and it being hugely valuable,

even though it's in this V1 format. It's funny. As I was watching this announcement yesterday, I'm sitting there and I'm like holding these two months in parallel. One is that like, oh my God, this is unbelievable. I can't believe Siri works. The other is, okay, so what you're telling me is all this is just to deliver on the promises that you should have made years ago. So it's like, it is this amazing novel experience, but it's like, okay, yeah, they've been working on this for a very long time. We haven't got it. So while we are enthusiastic,

it's like, finally, thankfully, this actually works. And I think it's important to really understand how big of a deal this is. I think when a lot of people use AI, the thing that they, are missing the most is the context. And we have memory when it comes to anthropic and open AI and you have like a loose memory based on what you've spoken about, but it doesn't have full access to everything. A lot of people, their whole lives live on their devices. It lives on the data on their phone, on their laptop. Apple is the ICloud account that runs all of that and having the full contextual awareness to go off and then make actions on your behalf is a huge thing. And at the core of this is Siri AI? Is Siri that actually works? I mean, for the last, what, five,

six years, I've had Siri disabled. It's been worse than useless. In fact, it just gets in the way whenever it hears his name and it gets summoned. Finally, there's an opportunity to turn it back on and start using it. And I think that's a really big deal for because for those people who don't want to go and subscribe to a frontier model, they don't want to pay $20, $100, $200 a month. They are just interested in personalized AI. This is a huge opportunity, a huge market for Apple. In addition to the developers, I mean, this is a developer conference. Apple developers now have to be a, have access to these really nice quantized local models that run on your phone that now have the ability to edge inference on any application that they want. So if I'm an app developer,

I'm going to the app store right now because I'm getting AI inference tokens for free that I could deliver near instantly because it's run locally to my customers on my Apple devices. And that's a huge unlock for anybody who's building for the platform, anyone who's using the platform. And while it did take them a long time, it's here. And we always talk about Apple has an install base of how many tens of billions of devices that now have. have edge inference ability on device, and that's such a huge unlock. I think this is very much version one now. This is the starting points of this new paradigm of AI-first hardware devices. I think there's also some important context to provide. Like, if you're an investor and you've

been watching Apple over the last couple of years, you just have one clear statement to say about them, which is they just messed up. They completely left the AI race to all the other players. They didn't train their own foundational model, and they didn't burn as much. capital, but also they got left behind in this entire race. And then you might be asking, well, how have they been able to produce these AI features that are seemingly quite good when you've got Google, Microsoft and Amazon spending upwards of $260 billion this year alone on AI CAPEX? Well, there's two answers to this. One of them you just kind of lightly touched on, which is Apple probably has the largest moat of consumer data of any singular company, right? They have 3.5 billion devices,

everyday collecting all these kinds of personal grounded contexts, which they can use to train a model. When you look at Chinese open source models, a big question that is being asked right now is, how have they been able to catch up with the American Western models that are spending so much money, and the Chinese models aren't doing that. The one answer is data.

Dwakash had a really good paper that he released yesterday. Well, he explains that it's primarily data. If you can get access to really personal data, you can train a pretty good model. So that's one way Apple's been able to do it. But the other thing is the elephant in the real. room, they signed a massive partnership with Google to get access to their Gemini model specifically. Now, the way that people frame this partnership in their head is, oh, they're just going to

plug Google's Gemini model into Apple Siri, and that's the brain behind it. And that's partially true. They do use Google's 1.2 trillion parameter model, but it's also not true. Apple has created a series or a slew of different models called Apple foundational models that are versions of models that they've trained by themselves on their own servers, on their own GPUs at home versus a kind of distilled effect from Gemini or Google's Gemini specifically. There was actually an agreement which got leaked, I think, like two weeks ago, which showed that Apple gets access to Google's Gemini model weights, which is just crazy for some kind of proprietary agreement where they pay them a billion dollars a year. So it's this mixed relationship, this mixed partnership,

which has allowed them to kind of achieve this particular type of model that we're seeing on screen today. Yeah, and it has some pretty cool technology, how it works. Instead of forcing the entire model into DRAM, which we've talked about many times on the show, if you have it, go find the episode. The model is actually stored in flash memory, and because the NAND and DRAM bandwidth is too slow to swap weights back and forth, they actually process it on like a per prompt basis, so it's just much faster, it's much more lightweight. It's this kind of novel architecture that they're using.

and I imagine a lot of people are going to slowly start to mimic that. And it's also important to note that these models use a novel architecture. They're using DRAM and S-RAM to pass things off to each other in ways that are faster and efficient than you traditionally could have. And I think that's why the models aren't going to be running on the iPhone 16 Pro, even though they claim that it was built for Apple Intelligence, because it really is pushing the limits of the hardware that they have on-device. Now, what are the types of things that you can do with that on-device intelligence?

they're pretty freaking awesome. One of my favorite demos from the whole event that they snuck in with one sentence, it was like three seconds of the whole event, was the ability to split the bill with Siri. So you could actually take a photo of a bill. Let's say you're at a restaurant and you had a receipt that gets dropped on the table. You take a photo, send it to the group chat. Everyone can be assigned what they bought. They could pay with Apple pay. You're done. You're on your way. It is so clean. It is so cool. There's a second feature that I love, which actually surprised me because it relies on agents to go ahead and do this, is you can figure. your passwords agentically. So inside of Apple, inside of iOS and macOS, there is a passwords

app where it'll safely and securely store your passwords for all the accounts that you have. It'll also go out and see if those passwords have been leaked, if they're being reused, if there is any sort of security compromise that's happened to those passwords. And what this new agentic system will do is within the app, it will go out and reset the passwords for you. So if you have an insecure password, if you have a password that's been reused, it'll go off in the background, it'll go go into the browser, it'll log in using your credentials, it'll change the password, and then it'll populate the new one into the app without you ever having to do that again. So the days of having to reset your password, which you forgot it, those are in the past.

And I think these types of use cases seem silly, they seem small, but they increase the quality of life so much for the day-to-day user. And this is at the core of what they spent almost a third of the show talking about, which is just stability. Like people just want a stable operating system that works and all the features that are advertised worked, and that's what we have. So these two features in particular were very exciting for me. There's something about giving an AI agent access to my passwords that just doesn't sit well with me right now, but maybe I just need to use the feature and let it change my life.

They've been refining this for how many years now that they can figure it out. They've been workshopping this, man. They had all the time. Their timelines are measured in centuries. Oh, that's so funny. Kind of similar on like the idea of like how Apple approaches their kind of like design when it comes to AI, we had to like look at the system prompt itself. So with all of these AI models, with all of these AI agents, including Siri, it's not too different from any of the other AI models. It has a base system prompt which the company, in this case, Apple feeds it in order to kind of like give it its personality and give it understanding of how it's supposed to act. And I'm not going to go through this entire system prompt, but it's just interesting to see that

they use visual richness, like quite a few times in this paragraph where they're basically saying, Siri, like, you aren't just a chatbot LLM, you visually understand the importance of design in this person's life, and you must make visually rich choices. Like, you must see what they see. You must think how they think and care for them in a particular way, which honestly, I haven't seen in any other kind of system prompt when I look at my clod or chat GPT. Now, I want to spin back to something which kind of gets into the infrastructure of how all of this works, because I actually think Apple's made a breakthrough in AI market.

model architecture, which not a lot of other companies have. And we've said on previous shows that Apple has the perfect architecture and kind of device system to run AI models locally and privately, and that gives them a significant advantage in the world or realm of AI agents specifically. Now, I want to briefly go through their foundational models. So it's something called Apple foundational models, and they're about four or five of them, which they release. And there's a voice one. There is an LLM-based specific thing. There's a transcription agent. there's quite a few others. You have it on our screen here. We've got AFM Cloud, AFM Cloud Image, and there's a few different features that kind of collectively feed into their Siri AI system.

But what interested me the most was their 20 billion parameter model. Now, before you get bored, I just want to explain why this is super important. Typically, to run a 20 billion parameter model on your phone, on any kind of Apple device, it would be nigh on impossible because the chips are just not too bleeding edge, and it results in a very slow architecture. Now, what they did was something very smart. You mentioned two memory types earlier on, Josh. You've got the NAND flash memory, and you've got random access memory, RAM. What they've done is they engineered both of these memories in such a way that you can access a 20 billion parameter model at lightning speed for super cheap cost. This is not something that any other device architecture has been able to achieve, and the only reason why they were

able to do it was because of Apple's supply chain. They have one of the biggest and best silicon moats across the world, and they were able to exercise that. And there's like a three-tier routing system. So any kind of simple prompts where it's like, hey, can you check my text messages? Can you like tell me what my friend said last week? That runs completely on device, privately encrypted. It never leaves it. It doesn't even need an internet connection at that point. And then any kind of medium-sized request gets routed to their private cloud server where it's still encrypted, but it kind of like routes to another server that's not on your phone. And then for the heaviest stuff, it gets sent to Google Cloud specifically, run on

Nvidia GPUs, again, still encrypted. So privacy is at the core of Apple's foundational architecture, and that might sound like an unsexy thing to announce, but it's actually crucially important if you're going to start sharing a lot of your personal information, your financial information, your grocery list, your personal text with an AI. It's important that that's kept at a privacy level, and Apple's really focused on that, which I thought was really cool. Yeah, in fact, at the end of the presentation, Craig Federigi went on stage or he was on camera saying it's actually not even rolling out to the EU just yet because the EU will not allow this amount of encryption on the services that are being offered. They don't allow for a complete and total privacy. And that's actually what Apple is doing. Same thing with China. They're not releasing it in China right now. And they said they're working with legislators to get it over there because they will not compromise on the privacy, on the security. And I think that's a seriously big deal. And when I

think about, I recall the bull case or the bear case that we mentioned probably close to a year ago now, which was the ability to have like on-prem compute on the edge inference, like on your local device in a secure and private way with all of your context. And that has not actually been a real consideration because no company's been able to do it. But starting later this year in fall when iOS 27's coming out, when all these new updates come out, it's going to be a real thing where if the average person wants to use AI, all they have to do is hold down the power button on their phone and activate Siri. And that gives them enough AI to get done the things that matter to them, to get done all of the contextual tasks, to have small agentic abilities, like going and change

your passwords, but also being able to plan a trip. And all of that now is going to live within the Siri application, which is now a standalone app on your phone, similar to what you used to with chat chit, with Claude, with Gemini, where you start a conversation, you continue a conversation, except you never have to train this AI on memories about you. It has all of the memories. It's It's totally private. You can trust it with your data. Your data is already there. It's an incredibly compelling offering if you are AI curious and you're not a power user who pays

$20, $200, $200 for these premium AI services. And I think for a lot of people, that's going to be enough. So we're going to see what the implications of that are as this gets rolled out through the end of the year and next year in the sense that is this actually going to make a dent in these subscriptions of the larger AI labs? Because people, turns out they don't actually care too much for frontier intelligence. They don't need to go solve biology problems or chemistry problems. They just want to be able to plan their calendar, schedule their gross free delivery automatically,

and make sure it's all managed internally by their AI. I mean, it's something I think a lot of people are going to be very intrigued to want to use. Yeah, if you're looking at this just from like an investor perspective, this could be a real problem if Apple actually scales this up, right? Because it proves that like small models work and you don't need a large model. And also the cheaper, less frontier models are just sufficient enough for people. Now, with the concept of subscriptions versus paper usage, I believe Apple's just letting people use their inference and AI features for free, but there's a daily usage limit

or like a pay per usage limit that might come into effect after you've hit a particular limit. What those tier packages look like, I have no idea. I think it would be crazy if Apple subsidized 3.5 billion active Apple devices. But I think we're going to start to see some kind of a subscription package where it's focused on giving like 80% of frontier intelligence for free and then like if you want anything after that you can kind of like pay you know on a usage base type thing now I have one quam with all of this Josh which is if you bought a very expensive iPhone 16 a few years ago which they marketed and advertised as built from the ground up for AI you can't run like 75% of the AI features that they announced yesterday you need something.

of the iPhone 17 or 18 and above because of the chip device architecture that I was mentioning earlier. So this is something that I guess is not the greatest thing. It might be another false advertising issue. But if you had a two-year-old device that you bought from Apple and you're thinking, like, hey, I want to run a bunch of these features that are rolling out by the end of the year, you can't actually do that. Yeah, it's a bummer. I mean, Apple messed this up a couple times over. And also, a note on the usage limits. Craig actually during the presentation yesterday, spoke about them briefly, where if you're using local inference on device, you have unlimited infinite. If you're querying the private cloud server, you get a certain amount. The overflow will be directed through

your iCloud account, your iCloud service package. So if you're paying $5 a month for two terabytes of storage or whatever backup, that will be enough to kind of absorb the extra queries that you have. So it's already baked into where you are. There's no separate subscriptions. It's all under one roof. In terms of this, yeah, I mean, it sucks. Last year, everyone was buying the new iPhone 16 built for Apple Intelligence. Apple Intelligence didn't exist. Now that it does, the phone is not compatible. It sucks. It's a bummer. And it's particularly a bummer because of some of these additional features that I do have to share in the Photos app. The Photos app is getting a really amazing upgrade. And as I was watching this, I have like the ghost of siege I was past

perpetually in my head. And I know that he is sitting here cringing because so much of the art form of photography is the real raw and natural way that it comes out. not the digitized version. What they've done is they've stepped it up. They've taken it all the way to the furthest degree. In three features that they announced, clean up, reframe, and extend. These are the three core pillars of your new photo app.

If you take photos, if you own an iPhone, you're going to love this. So the cleanup feature is basically what we're seeing on screen here is an original photo with someone holding a card in front of their face. In iOS 26, it lacked the contextual awareness to understand what your face actually looked like.

So it turned you into a blob. You kind of had like a fudgy, nose with like really mixed up eyes. IOS 27 solves this. It applies real intelligence to these photos. It really makes it look like a pretty big deal. Refram is the second feature. I think reframe is probably the one that I'm most excited about. Reframe will take a photo that you've shot that is framed pretty poorly. If you didn't frame it well, if it didn't look so good on the first version, what it does is it turns it into a splat we're seeing on screen. It turns into a 3D splat which basically adjust and adds layers to the image so there's real depth. It allows you to

zoom in, pan, tilt, and move the image, and then any parts of the images that are lost, it will actually generate in the background afterwards. So you can very quickly reframe it, and then it will generate afterwards in your camera roll. And it leaves you with really compelling results where, I mean, what we're seeing on screen is this is a totally different image. I'm looking at the background, Josh. Can you see it? Like, I'm looking at this person's face over there. I'm wondering whether it gets distorted a bit in the background. But it is a very cool feature. No, that's cool. Yeah, it's amazing. So that's the second one, reframe. The third is something that's really cool that we've, I mean, granted, we've had these for a while.

We haven't had reframed, but we've had Extend for a while in Photoshop. But Extend is now making its way to iOS as well. So if you've taken a photo and you need to fill the edges, if you want to remove something, if you need to kind of zoom out, this is what Extend will do. It will take an image. It will build artificial borders around it that look real, that are artificially generated, but are meant to be hyper-realistic. And this is all leaning on the visual AI local model that runs on your device.

You can see on screen, it just generated an entire coffee shop around a uptight, upclosed shot of coffee. So it's really impressive. I think a lot of people are going to use these. A lot of people are going to love these features. The example they used inside of the demo on the actual presentation was just photos of your loved ones. If you take poor photos that you don't want to lose that moment, you can actually manipulate the moment in such a way that it still maintains the integrity of it, but captures it in a way

that's much more just delightful to see. So this in particular was something I'm really excited to use. As someone who loves the phones for the photo features, having these tools natively built into the photos app, two thumbs up. Okay, so when I was a kid, one of the things I enjoyed doing because I was such a cool guy and I had lots of cool friends to hang out with was play around in the virtual simulator. So I would fly around the world and be able to see the world for what it is.

It looked very much like how Minecraft looked today. And what Apple released yesterday was something pretty cool, which is their 3G Gaussian Splat version for Apple Maps itself. And I mean, so sick. Like, I don't need to do, I don't need to say anything. Just take a look at this video. This looks so hyper-realistic. The trees don't look like broccoli's anymore.

In fact, like, if you weren't, if you didn't know this was a Gaussian spot, you would just assume this was like an aerial shot drone video of like your city, of New York or wherever you might be. So the purpose of this, I guess, is to make travel or navigation way more interactive. and way more intuitive for you. Maybe you're looking for a particular shopfront, but you don't know what it looks like. It can point out, like,

hey, this is where you are and you can kind of recognize your surroundings around you. Just a really cool feature and use of Gaussian spots. I know that we've been wanting to make an episode on this for a while at Limitless,

and some of the examples that we've shown in the past has been primarily focused on Hollywood studio effects and recreating movies from scratch for a couple hundred grand versus millions of dollars. This is a real practical application

that will be live by the end of the year in 3.5 billion Apple active devices, which will be super cool to see. Yeah, and we got to put on our conspiracy hats a little bit here, because why is it that we only talk about Gaussian splats 3D splatting when we mention Apple? Why is it that they're the only company doing this? And I think a lot of what happened at WWDC this year was just returning to their roots, patching all their bugs, building and establishing a new foundation in which they could then launch the new hardware architecture on top of. John Turnerst, the new CEO, this sadly was Tim Cook's last WWDC.

After this, John Turner steps in. He's a hardwork guy. What Tim Cook is leaving him is this really strong foundation to then build this next level of hardware on. So when we look at these splats, how accurate they are. When we look at the photo app, how it's so good at understanding three dimensions and three-de-splatting,

a lot of this is the infrastructure and the foundation for the next suite of hardware devices, for the glasses that they are going to be working on. And you'll note that there was a lot of rumors over the past couple of weeks that Apple has fully scrapped their Vision Pro program. The large goggles that everyone loves, I mean, that I love that were largely considered a flop, but I think we're the most impressive piece of hardware ever. They're no longer on the upgrade path. They're not going to continue to iterate on those. They're going to build

glasses instead. What are those glasses going to rely on? A lot of things like this. Splats. Why? Because splats are very low bandwidth, high fidelity versions of reality that you can interact with. And when you're trying to run these things locally on device, on something as small as glasses, it's really important to have that architecture in place. And what we're seeing here is early tastes of what these future devices are going to look like deployed as really fun products. So we have Google Maps that has really great splats.

Why are they doing that? Because they're improving their splat technology for this next generation of hardware. And I think that's really important to note. Another thing that was really exciting is, or maybe not exciting for some, is that they didn't actually capture any surface area. A lot of times at WWDC, we make the joke of like,

what startup are they going to kill this week? Who are they going to destroy a knock out of business because they replace their feature? And the answer today was nobody. There's really no novel new things that they were trying to replace. They were just going back to the basics. They're making everything that should have worked in the past. Work currently, for example. On my Mac, when I press the search bar and I go to search for something, it takes 10, 15 seconds

to populate. If I'm searching my email in the mail app, that search bar sucks, dude. I can't find any emails that I want. It's so bad. So they spent this WWDC really just fixing everything, building the found for this next frontier of hardware. And I see on screen you're teasing a resizing feature that also infers there's another piece of hardware coming

separate from these glasses that we could expect pretty soon. Yeah, so for a lot of people seeing this demo on screen right now, they might just kind of look past it and be like, okay, whatever, who cares? What we're seeing on a screen is basically the UI or the user interface being resized in real time, flipping between an iPhone orientation to an iPad orientation, to maybe even a full MacBook Pro screen sizing.

The reason for this is they're reformatting the code, Xcode specifically for their rumored new device that is going to be a foldable phone. So you should be able to flit between a regular iPhone-sized screen and an iPad-sized screen. And so Apple is really prepping not just their software, but for their future of hardware, which is going to look very much like the new world of AI

or for a new era of AI. And it's going to involve a lot, not just like screen and visual outputs, but it's going to involve a lot of transcriptive stuff. It's going to involve a lot of voice-based AIs, as we've seen with the Apple foundational model. So I think the way, if I had to summarize WWGC of this year, I would say that Apple is laying the foundation

for taking AI a lot more seriously going forward. And I think with John Turner's at the helm, hardware is going to be at the forefront of their focus. I think they have the foundational elements. They have the Apple Vision Pro, which was just kind of too expensive and not accessible to enough people. They're rumored to be working on a new set of glasses, which has been derived from the hardware that they built using Apple Vision Pro. They're rebuilding their entire iOS system. They've got Siri as their main flagship AI system, which, by the way, doesn't need to use Google Gemini models in the future. It could technically be an orchestrator. That's how they described it on yesterday's event, that it could be an orchestrator of different models.

It could be Apple's own foundational models. It could be chat GPT in the future. It could be clawed in the future. The point is, if Siri AI is the funnel, then Apple owns the entire distribution, and that alone is the most important or rather bullish investment case for Apple alone. Now, despite this, the announcement yesterday or WWGC sent Apple stock cratering around 6%. So, and I think that was a classic case of buy the rumor, sell the news. But I do think if you're taking a long-term perspective on Apple, at least for me specifically, I'm incredibly optimistic about where they're going to take this simply because they have one moat. that Open AI and Anthropic don't have, which is just so much user data and intense and the

hardware to be able to distribute it on. Open AI is working on their own AI device. They haven't released it yet. And once they do, it's going to be really hard to scale. Anthropic, for all I know, isn't working on a hardware device yet. Maybe they will in the future. But in order to ramp up and compete with Apple, the only other company that can do that is Google. And they're kind of sequestering their own AI model to Apple to use for a billion dollars here and giving them access to model weight. I think if you're looking at Apple as an investment case, if you're looking at Apple and thinking, how will they fare in the future for AI?

I think they're going for the consumer AI mode specifically, and I think they're really well positioned to do so. Yeah, I think sentiment for me personally, at least, changed in a meaningful way this time around. There was no flashiness. There was no, like, selling the magic, which is a bummer, but that's not what this needs to be. This needed to be stability,

returning to the roots, building that strong foundation, establishing a place in which they can build their next form factor of hardware on top of. And there was very much a mission of success. And what a gift this is to John Turnus, the new CEO from Tim Cook, to hand off a feature-complete software stack ready for the next generation of hardware, which he specializes in. And we just need to take a moment to appreciate Tim Cook one more time,

because this is the last event that he's going to be doing. The next hardware event is going to be in September that is going to be for the iPhones. that's when John is going to be starting his journey as CEO of Apple. And Tim did the impossible thing of taking over a company from what a lot of people deem as the most incredible entrepreneur of all time Steve Jobs. And he did that 15 years ago in 2011. And over that 15 year run, he managed to take Apple on a, I think, 2,200% growth rate of a stock that was impossibly challenging to scale to follow up the most challenging CEO role

of all time, and he's just, he's just a superstar. He just did a really great job. He managed to steer Apple away from a lot of the drama that could have got them in trouble. Although, like, the company, they did lack magic during this time. They didn't have any incredible products or hardware. They did the basics right. And still, to this day, 15 years later, I am using exclusively Apple products for everything that I do. And outside of the software that they just fixed, I have no complaint. So it was a pretty incredible run by Tim Cook. Just got to give him some props for what he did. And now I'm just really excited for the next generation of Apple. This very much feels like a fresh slate.

Tim Cook handed off a really strong company to John. John is now going to take it forward with hardware. And personally, this is what I'm most excited about. We have Open AI with their AI-first device. Apple has a chance to get in the game. They're going to be doing it. And the next couple of years are going to be awesome. We have the 20th anniversary of the iPhone next year.

We have a foldable this year. We have the glasses coming probably next year, 2028. I mean, this is a huge, fully stacked lineup. And, man, Apple's got a shot at this thing now. I think something coming potentially sooner than that will also be like their earbuds or their airports that have the cameras and they've got some kind of other pendant type device, which are all rumors,

but maybe this fall. And so I'm excited to see that go head to head with whatever open-AIS devices. But that brings us to the end of the episode. That is everything and anything you need to know about WDC, very AI-focused, which is what we like on this show. And thankfully, things that we can actually use today.

If you're a developer, I don't know if you mentioned this earlier, but you can get access to this for about $100. It might be a little tricky too. You need to get like that membership, but it is accessible. If not you, these features will basically roll out by the end of the year. So lots coming by the end of the year. We've got all these IPs, we've got all these new AI features,

all these new Rumed AI devices. It's going to be a big, big rest of the year. But that is it for this episode. Josh, any final thoughts? Yeah, note, if you're a developer, if you have that license to go and download iOS 27, You probably already know this, but don't download this on your personal devices. The beta versions of all of these are notorious battery drainers.

They will destroy your phones. They're not meant for public use. They're meant for developer use. But if you do have it and you do install it on a second device, let us know what you think. I'd be so curious to hear the reviews. I've been on X scrolling, seeing all the features. I'm curious also just to know your thoughts on this. Are we, is Apple back?

Are we forgiving Apple or can we not forgive them for the fact that if your phone is one year old, you can't even run the front-year-models on your device? I understand both sides. I understand both sides. They're both right, but I want to hear which one you stand on. And yeah, that's the update. You are now fully up to date on WWDC. We have a very exciting week of content coming, so please stick around for that.

But without any further ado, I think that wraps up. Thank you guys so much for watching, and we will see you in the next one. See you guys.