Sharp Tech with Ben Thompson - Apple Gets Intelligence, Unpacking the Partnership with OpenAI, What WWDC Can Tell Us About AI Adoption

Episode Date: June 12, 2024

Reviewing all things WWDC in 2024, including the introduction of Apple Intelligence, why Apple's vision for mainstream AI was compelling, the Apple partnership with OpenAI, questions about Private Cloud Compute, and lots more.

Hello and welcome back to another episode of Sharp Tech. I'm Andrew Sharp. And on the other line, Ben Thompson, Ben, how you doing? I'm enthusiastic, Andrew. I'm enthusiastic. VisionOS 2 is out. It is going to rescue the VR headset market, ready to rock. I almost consider just for you leading with the VisionOS news coming out of WWDC on Monday because I know.

I have a Vision Pro take, actually. That has nothing to do with WMDC. So John Gruber, my co-host on Dithering, incredibly enough, is going to be doing the first ever live broadcast to Vision Pro. Yeah. It's going to be like a 3D sort of thing. I'm very excited about it. Sort of wait a waiter tonight.

And so I put the Vision Pro on, which I hadn't had on for a while. Fortunately, I brought it with me back to the States here. And you know what? When that's on, I enjoy it. Like, it's like this is cool. I believe it. I still. think like, I don't know, we've talked about that Apple has dropped the ball. They are going to have

this sort of widescreenmatic experience, which by the way, I think makes a ton of sense. It's so much smarter than two screens. This is why we are pundits and not product designers. And also, I would imagine sort of scales based on sort of your bandwidth or Wi-Fi, so sort of handles that gracefully. But I don't know, I put that on. I'm like, man, I wish there was more stuff on here. Every time I put it on, I feel happy to have it on. I just have no motivation to put it on. That's part of the problem. You know what my Vision Pro take is? I wish I had more disposable income because I heard the announcement about John Gruber.

We're running this on Wednesday. So the event will have already happened. It will have already happened. But I heard that he's going to be broadcasting live on the Vision Pro. And I was like, man, if I had a Vision Pro, I would be all in on the Sandwich Vision experience and, you know, seeing what that's like. But alas. really intrigued though because I've been there obviously I've been there lots of times when John is hosting this and and you know it's enjoyable to be there but you know just this idea of like what's it going to feel like are you going to feel like you're alive like it's it's very intriguing so yeah I'm looking forward to it this is the the dream is really like this live event sort of bit with the Vision Pro where you feel like you're there so I will report back but yes that is not necessarily the topic at hand I did enjoy broadly in WWC actually had a question was Because Apple always organizes WWC by operating system.

And actually this is something John and I have talked about on past dithering's where that organization is kind of starting to fall apart, particularly when they have features that stretch across operating system. They introduce them in different places. And then otherwise, oh yeah, it's also on that one and also on this one. And this Apple intelligence bit was going to be the ultimate pinnacle of that. Like obviously this feature is going to be available across their operating system. It's not operating system dependent necessarily. And so how are they going to do it? And you kind of realize right away, it's like, okay, they're going by operating system

and they are blowing through these. And they're blowing through these because there's very little to talk about. It's very clear that, number one, it was smart to break out all the Apple intelligence stuff. And number two, you definitely got the sense that everyone at Apple has basically been working on nothing else but this for the last year. And so it was very easy to swish everything else into sort of like the initial 45 minutes or so. Yeah, it worked out well because frankly,

I was interested in Apple intelligence and nothing else that they were bringing to the table. You're not looking forward to color your icons? Well, my eyes really started to glaze over during the Ubisoft portion of WWDC. We shifted gears and got into Apple Intelligence, and that's where we're going to live for most of this episode today.

So Apple Intelligence is here. And there were a bunch of different threads I found interesting as the internet was digesting all the Apple announcements on Monday. So we'll go through tweet by tweet on various aspects of WWDC in 2024. And I'll start with this note from Steven Sinovsky. This integration is so exactly right, he says, of Apple intelligence. It shows that a quote unquote chat app was a short term demo UI. Combine that with all the context. on the OS and the privacy.

And this is all how AI can really come to life the right way. Did you pick this first because you're a chat app hater? A little bit. Yeah. Honestly, that's okay. Well, then let's get your response. You are the chat app aider. You pick a tweet that is anti-chat app.

So what was your sort of perspective as a newly minted sort of tech expert watching them to WDC? I'm a tech guy. days locked in for the entire presentation. No, what I found striking from the Apple event was it was a vision of how this will work in normal people's lives and it was a clearly articulated vision. Like if Siri can ingest all your old text messages and tell you what someone's address is or if Siri can work across emails and calendars. That was the single best demo was the woman who needed to pick up her mom.

Yeah. And there was information scattered across text messages and email. It was so relatable. Like we have all been in that situation. What time is the flight? Where was the dinner reservation? Having to sort of go across apps and just being able to find the stuff. Yeah.

Yeah. Like that one demo right there was just because that's sort of Apple at its best. Like super relatable problem. Here, this is how you solve it. It's like, oh, yeah, of course that's how I solve it. Yeah. those are features that everyone will use whereas with the chatbots it sometimes feels like people have to strain to invent use cases at least normal people who don't work in tech and aren't programmers or anything like that but those are features that are so convenient that everyone will use them and watching the presentation one of the things i was reminded of was how we were talking a couple weeks ago about the iPhone and when the iPhone came out initially it was an unbelievable piece of technology, this huge leap forward in capability.

And then as it was updated year over year, the updates became more ho-hum and unremarkable. But you pointed out, if you step back and compare an iPhone 15 to an iPhone 8, the difference in capability and convenience is pretty massive. And so to go back to what Sinovsky was talking about, what? Was there even an iPhone 8? Oh, no, there was no iPhone 9. Okay, you got lucky. Oh, wow. But the iPhone, the iPhone 8 was a weird model because that was the same time they released the iPhone 10,

you know, common referred to as iPhone X. And so that was when they split to the bifurcated sort of models, but they hadn't adjusted the namey yet. But there was no iPhone 9. I thought there might have been no iPhone 8, which would have made that a very, very funny reference. I completely interrupted you. Sorry. The dangers of podcasting with an expert historic. in here. Yeah, no, but...

Well, I was wrong. I'll take the L. Well, L's left and right these few weeks. So, continue. To go back to Sinovsky's tweet, so much of the AI hype cycle started with a chatbot UI and this chat bot that had human-like responses and all that inspired these wild proclamations about what AI might be as a product. But what I saw with Apple is what AI is going to look like. as a feature where there are just a bunch of different tasks that AI can streamline for people. And similar to the progression of the iPhone, there are going to be all kinds of subtle ways that AI improves what we do on a phone. And then that will build year over year and just make for a much better computing experience

while not necessarily radically transforming our relationship to technology or leading to this science fiction. But you won't notice until you sort of look backwards. Yeah. And it's not going to be the science fiction future where we're talking to agents like humans all day. I don't know. This vision, the vision that Apple articulated just feels a lot more plausible to me in the short and medium term. And maybe it will ultimately lead to a future where we're just talking to our phones all day. But I came away very impressed by Apple's integration of AI into products that people,

people use every day, and also impressed by Apple as an institution where they weren't reacting defensively and panicked. This was Apple recognizing what will actually matter to people, what will move the needle, and being able to separate the signal from the noise as far as how AI might actually be adopted by consumers on a day-to-day basis. So I thought they nailed it. So this is something that is both very beautiful and maybe for you, of personally a little bit tragic, which is I know you like to think of yourself as sort of a unique flower, you know, you're, you, you, you're, you, you, you, people say I am. I want to zad. Yeah. Yeah. And unfortunately, I have to say the sort of the path of going from,

you know, for whatever reason, I'm interested in technology. Maybe it's personal interest. Maybe it's my job. Uh, to then sort of a couple years later being like, why does Apple do everything right and everybody does everything wrong. Like, this is the road to Apple fanboyism that everybody goes down. It's not a special journey, Andrew, but congratulations on making it down. I mean,

they just did a really good job. And honestly, one of the other things that I came away wondering is how could they get this so right and get the iPad commercial so wrong? So the iPad commercial is now even more incredible

to me. Maybe all the capable people were working on AI and everybody forgot about the iPad commercial. I mean, no, I think you're exactly right. I think Apple crushed this and like it is, you know, there was an aspect of I think this was coming into focus a fair bit, you know, not to sort of overly focus on myself, but like, you know, usually you don't want to write the article about WWC before WWC happens, but there was sort of, I just had this sense. And it's not a new sense. If you go back to my AI and the big five that I wrote like, you know, back in 2023, right after Chatupy had happened as my first article of the year, sort of weighing how I viewed the industry as a whole.

It was pretty positive about Apple saying, look, they're not going to be disrupted. They're pretty well placed here. Local inference is definitely going to be a thing that's going to be important to them. And so that was sort of the fuzzy outline of a vision that I tried to articulate much more clearly that week, this week where it's not just that Apple has sort of the right place in the value chain to really leverage AI. but also, and I think what came through WWC, maybe I didn't put sufficiently in my article, is that Apple is just really good at this.

Like they are the reason why people fall in love with Apple, the reason why they're so compelling, it goes back to sort of this Steve Jobs sort of truism where everyone in tech is obsessed with sort of feeds and speeds, and they forget about the experience, the look and feel of sort of using a product. And the point is not to sort of do something because you, can. It's that technology will assist you in getting things done that you need to get done. And it's, it's going to be, it's not going to be a burden. It's not going to be something you feel compelled to do. Oh, what can I create to ask chat GPT so I can feel like I'm in with it. Right. It's just, it's there and it does sort of what you want to do. I think that's why you're feeling this positive reaction. I think a lot of people can feel this where in this new arena with AI, you're getting this sensation of it's not, I'm not going to say it's an iPhone moment.

it is a characteristically Apple sort of arrival on the scene where everyone's like, oh yeah, duh, of course that's the way stuff should work. Of course that's the way it should sort of be integrated. And you add that to the fact that I think strategically speaking, owning the device is so powerful. And I put in my update today, it almost feels unfair in that what Apple is doing is the stuff they're proposing to do with Apple intelligence is all. all relatively simple. They're not doing these crazy, wide-ranging, agentic sort of things necessarily.

They're not doing like knowing all the knowledge in the world. What they're saying is, look, there is stuff that is pertinent to you. And opening eye doesn't know what that is. Google might know what that is. We should get to Google made a little bit, but Anthropics not going to know what it is, all these sorts of things. Even an entity like Facebook, like your email is not in Facebook, for example. Your phone is sort of the point. That's the, yeah, that's the hub around which everything organizes. Because

we own the phone, we can know everything about you. And you will trust us because we have spent years and years burnishing our sort of privacy credentials, you know, doing things that are arguably self-serving and hurt other people, but also is

consistent with our brand and our sort of brand promise when it comes to things like, like, you know, ATT, and privacy and sort of all those sorts of things. We fought with the FBI about unlocking a phone. Like you go back to an episode like that, and you realize how impactful that is on what they're doing today because they get the benefit of the doubt to say,

what we're going to do with AI is actually dig into every detail of your life. And we're going to use that. And what we're going to do technically, and this is why technologists get really mad about Apple, what we're going to do technically is going to be relatively simple. We're not going to boil the ocean here. And that's where we're going to focus all our resources, all our models, is we have one thing that nobody else has, knowledge about you. And so we're going to do stuff that is predicated on having that knowledge.

And oh, by the way, if you want to spend billions of dollars and have a gazillion Nvidia GPUs to have knowledge of the world, okay, fine. Here's a little interface. Come plug in. You'll come in on our turn. And so you end up with a situation where the iPhone is proposing to have everything from these personal models to world knowledge. And Apple's investment is actually on the quote unquote easiest part of that. And then they will at Open AI. Oh, come line up.

Oh, by the way, it's not going to be just you. Anyone can come line up. Google, you want to come line up? Anthropic, you want to come line up? Meta, you want to plug into here. It's an open interface. It's there if you want to come and get it. Or open-ish.

is now you can just sign up and do it, but they made very, very clear. Anyone, any big model can sort of plug in there. That's not the core of the offering. What, and so it's, it's like catnip to someone like me, not just because of the product bit, but the strategic sort of placement and leveraging of where they are in the value chain. It's, it's like the analogy I would go is with the iPhone and phone carriers. The actual cost in delivering incredible sort of like internet everywhere experience are the poor saps who are out there actually putting up towers and investing in like fiber and satellites and all this sort of thing. And they all work really, really hard and sort of knock themselves over and barely earn any money in the service of Apple selling iPhones at 40% margins.

Like if you own that user experience by delivering something delightful and nailing it, you get everyone. else to serve you. And that feels like what they're doing with AI. Yeah. Well, and here's a rundown for anybody who didn't catch some of the specific announcements. This is from NGadgett. Apple revealed its plans to incorporate AI into its operating systems at WWDC this year, dubbed Apple Intelligence, this new generative AI system will appear in the form of practical tools that most people can use regularly. Those features include new writing tools that can help you rewrite, proofread, and summarize things like emails and other messages, original emoji and image creation, and more.

Siri. Sorry, can we, can we, can we pour an out for the emoji bit? For my emoji bit? Well, so, no, so emojis, I love emojis. They're delightful. They're great. Actually, you know, the, it's this whole sort of thing where there's like this international consortium that decides on what the next set of emojis are and then everyone sort of

implements them. And Apple is basically sort of blowing that out of the water where, I mean, I do love the fact that you can create these emojis in the style of current emojis. It's not like there is something else. It's like it's going to look like it's a regular emoji. That's what I wanted to mention is I mocked the emoji bit a couple weeks ago and then I watched it and I was like, ah, it does seem like something I might use and should be kind of fun. So really W is across the board for Apple on Monday.

What I'm saying about about pouring one out though is there's a bit about sort of the internet generally that it's like your monoculture sort of bit. Yeah, a universal language. Yeah, a universal language of like, yeah, there's a common. Everyone has the same set of emoji. The same limits. And that is that that period is sort of ending. Like everyone's going to have custom emoji and it's cool and customizes neat and niches. but it's going to be sad when we don't have everyone upgrading their

version of iOS just to get the new sort of, you know, taco emoji or whatever it might be. So an era did sort of end there, sort of silently in the night. Wow, striking a poignant note here. The end of the emoji era, the universal language that we all shared for decades on end. Yeah, I can't argue with you.

Now I was all excited. Apple intelligence is the first day of the rest of our lives. What am I signing? Burst your bubble. Sorry about that. Well, no. So continuing on, Siri is getting an AI infusion. And in addition to asking Siri to delete an email or edit a photo, users will also be able to ask the virtual assistant to summarize articles and web pages in Safari and even extract personal information from a picture of an ID so it can fill out an online form for them. The company emphasized the importance of, quote, personal

context with Apple intelligence, which will enable things like using natural language to search for photos that contain only specific family members or friends. OpenAI's chat GPT is also integrated into Apple intelligence, allowing users to give Apple permission to share their queries with chat GPT, quote, when it might be helpful, end quote. How annoying is this going to be, by the way? I don't anticipate using the chat GPT feature all that. often, but I know that you're pre-annoyed and will be very, very annoyed once they actually make this available. Apple's been softening to the ground with the stupid, do you want to paste in this sort of app sort of thing? What I've never, I think I already just complained about

this lack of last episode. But it came up today. I was putting my wordle into a chat group. I guess I want to paste my word'll win. But, um, it'll. Well, so examples of the open AI functionality include asking for menu ideas that. incorporate specific ingredients or asking for decor advice while providing a photo of a space that needs sprucing up. So the example they had was someone on their back deck asking chat GPT for plant suggestions to spruce up the back deck, which does seem like something that would be useful. But we'll come back to chat GPT momentarily. Ethan Mollick tweeted, Apple is betting on AIs based around lightweight reasoning and very good local

retrieval slash low hallucination rates. Example, what's my dinner plans with mom? Open AI is betting on autonomous agents. Example being set up a meal with my mom somewhere she likes and maybe a show afterwards in my budget. So do you think that's a good synthesis of the overall approach from some of the incumbent AI players and the direction that Apple wants to take it? Well, I mean, no one's going to do that prompt for real with Open AI.

Right. So, but which sort of speaks to the point. I mean, I don't want to shortchange the large scale models. Like, and just to be very clear, what they are doing, and again, this is what drives technology up the wall, is technologically much more impressive than what Apple is doing. But that doesn't mean it's more useful, right? Like, which, which, which is, you know, sort of the thing that can sort of drive people up the wall. But it's, it's just sort of the case, being able to pull out this personal context. I talked about how Apple. models are grounded both figuratively and literally in my update today, literally means like that's actually a term of art, which is like this response is based on this information that is at hand. It's not sort of just like hallucinating stuff, making stuff out of thin air. But the figuratively is the actual things you want to do in your life or you actually care

about is actually a relatively constrained space. I'm happy for you decorating your back porch. But, you know, number one, I don't believe. you that you think that sounds useful. And number two, once you do it once, you're never going to sort of do it again, right? This is the challenge of trying to be a jack of all trades, everything to everyone. Now, that also means there's a ton of places where these chatbots can be really, really useful. I mean, I use them a fair bit, like for doing some like basic research. And of course, I check my citations wherever I be, but just reading up on something sort of understanding broadly,

asking questions, really, really sort of good and useful. I'm glad it exists. And also, also I could probably use Claude or Gemini or sort of whatever it might be. But again, the secret sauce is when it comes to asking when, where am I having dinner with my mom? Only Apple or potentially co-pilot on Microsoft or, you know, Gemini on Android has that context. Yeah. No, exactly. And OpenAI couldn't do what Apple is proposing to do in a lot of these cases because Apple has the context. I will just note for the record, on greatest of all talk, Ben Goliver has been compelling

me to take various oaths over the course of this playoff run. I had to take a Drew Holiday oath on the most recent episode, which you could add by going to your show notes. And I've been swearing these oaths on a copy of Cable Cowboy, John Malone's book. And I will swear on the copy of Cable Cowboy. I do think the functionality for the back deck would be useful to me, personally, only because I'm so ignorant as far as gardening that I really need the help finding out what plants would work where. But in any event, the other note that I wanted to

mention was just the name Apple Intelligence. Ben Bajaran says naming it Apple Intelligence is the most Apple thing to do. Coming soon, all your devices enhanced by Apple Intelligence. Just such a classic case of evil genius from Apple with this one because it will appeal to people who are really excited about AI. And then on the other hand, for the people who are uncomfortable with AI, Apple gets to say, it's really more of an Apple feature and it's defined by our core privacy principles. We're not selling AI like everybody else. This is Apple intelligence. Did you have any takes? This was the sort of most nerve-wracking moment of the keynote for me. Because, again, I had written and posted an article before the keynote came out is that Apple intelligence is right on time.

Like basically they're not late. Like they are extremely well placed. You know, it's going to be great. And then they launch into it. They're like an AI for the rest of us or something along those lines, you know, sort of a classic Apple line. And they were talking about personal intelligence, personal intelligence. I'm like, are they going to call it personal intelligence? I thought it was going to be it.

I trusted you, Mark Kervin. But then they finally said, yes, it is Apple intelligence. intelligence. Great name. It's AI, Apple style. Yeah, evil genius for sure. Apple does have a much more mixed bag when it comes to naming things. But I think this was a good one. And they stuck with Siri, which was another question. They did stick with Siri. Yeah. Which is interesting. I guess it is established. You know, certainly what they're showing and promising looks so much more capable. You know, and there's a bit, you know, and there's a bit. here too where a lot of this stuff is going to take time to roll out. Like it's not just going to be available in the fall. It's going to take time. Like the betas don't even have a ton of the stuff that they showed. You know, they've fallen into this pattern more recently, which is they show stuff at WWC. Some of it comes in the fall. Some of it comes in what they call like a spring update with a huge sort of sort of update sort of way or that year. I think that's okay. You know, Apple, I think,

is by and large, earned the benefit of the doubt. But it is worth noting all this rapturous praise and all these sorts of things does depend on this stuff actually working. And, you know, they did release these like open source models that were terrible. Like they were like Microsoft at the same time released their Phi models that like wiped the fourth them as far as super small models goes that could be embedded on device. Apple did release a post, not really a white paper with a bunch of scores that said, you know, look, our models on device actually outperform Microsoft Phi are better than GP3.5. So TBD, like, but they have to actually actually.

deliver. There is a bit where I think what should make people nervous is if Apple delivers, this is supposed to be a sort of technology and capability that Apple is inherently not that good at. If they deliver and it's good enough, it might be sort of the nail in the coffin for, look, this is a commoditized technology. Yes, you might on edge cases be good at certain things. But look, even Apple can deliver something that is sort of good enough. So interesting read on that. Yeah. Yeah, I mean, we'll see how any of this actually works in practice. More than anything else, I thought it was just a really clear articulation of what AI can do today as opposed to what AI could mean in 20 years. Well, that is the other, that is the other big question, right? The way in which this strategy is wrong is that AI actually gets to be as good as its biggest proponent, sort of argue it's going to be, and that happens quickly.

Yeah. And so Apple is sitting here, you know, sort of like, oh, we're not, it's a feature, not a product. And actually, no, AI is so good that it actually is the product. It's going to be a product. Yeah. Well, and Apple's screwed in that scenario regardless, correct? Well, I mean, to an extent, I mean, like, like, like, like, like devices are always going to matter. They're all, like, so Apple makes devices. But if the AI matters more, than the device is sort of commoditized to a certain extent, you can use sort of any device. I mean, look, if you want to get to a her sort of example, Apple's going to figure out how to make that hardware before anyone else.

Like they have real capabilities and strengths in that area. But to the extent that AI is fairly commoditized and they just need to deliver something that's good enough. And they can do it sort of locally and have the sort of trump card of we have your data so we can make it actually useful to you. Then they're in phenomenal shape like better than ever. I mean, like the, you know, they can drive like there is a reason to buy new. Apple hardware now. So right now it's shaping up. If you're relatively like, look, this AI stuff is

great. It's not AGI. It's not like, you know, there's fundamental constraints or limitations. Apple is phenomenally well placed. If you're on the, look, you don't understand how good this stuff is where the exponential curve is going. Apple is putting themselves in a corner in a certain respect. Then, I mean, that is sort of the bare case here. Right. Okay. So yeah. And just to be clear, if you want to use Apple intelligence, you would need an iPhone 15 Pro or an iPhone 16. Well, it doesn't exist yet to be clear. But yeah, the, yeah, well, so this is interesting. I mean, it goes back to the M1 Mac or the iPhone 15 Pro, which is, so the Macs go back like

four or five years or whenever they first introduced that. The iPhone only goes back one year. And even then, it's only, it's not the regular iPhone 15, that watched last year. What that says to me, and this is totally understandable, is that the constraint is RAM. We've talked about, you know, these models have to fit in memory. You have to have the entire model in memory to use it. You can't sort of page in or out. And so the, you know, it's a relatively small model, but it still takes a few gigabytes of memory.

And the iPhone 15 Pro is the first iPhone that had eight gigabytes of memory. The M1 was also, number one, I had the neural chip, which is the Wimiter and the Mac is the neural chip. but the M1 had 8 gigabytes of RAM. And so it's a totally understandable and logical cutoff point. I do feel like Apple doesn't like to talk about specs. They never talk about RAM in particular. I feel like this is one that would have been useful to talk about. Number one, people are going to whine and complain that Apple is, you know, being mean by cutting people off or whatever.

Actually, no, there's very legitimate technical reasons why the cutoff is drawn where it is. And number two, if you really want to kick that sort of upgrade cycle in gear where people feel like they should get a new phone or get a new Mac, or they should get a higher end phone or Mac, like if they have an iPhone that is 16 gigabytes of RAM, that is actually a good reason to upgrade. My MacBook sitting here has an absurd amount of RAM, and the reason I did that is it's more than I need, but I knew this model stuff, you know, if I wanted to run larger local models, just need a lot of RAM. Right. And so, and RAM makes a lot of money for Apple, but they don't talk about it.

I had never thought about RAM until you started making this point a couple months ago. And I agree. It would have been a natural inclusion to the presentation on Apple intelligence and how all of this is actually going to be implemented. Right. It's very geeky and weird. But I think,

again, it sort of insulates you from criticism about why you're drawing the line where you are. Yeah. And it's a real spur to buy a new device. and make sure you get a high-end device when you buy. Yes, absolutely. Okay, well, so shifting gears, the open AI aspect of this entire arrangement. There are two tweets in tandem that I'll read here.

Rachel Metz says, I'm still at WWDC. And during an onstage chat just now, Apple's SVP of software engineering, Craig Federigi, said the company plans to integrate with AI models such as Google's Gemini in the future. and that different people will be able to use different models. Interesting. And then Jason Calacanis says, Apple can hot swap OpenAI at any time. They made that point multiple times. This means the default LLM is going to be an auction to the highest bidder, all things being equal, in parentheses like the search box on iPhone that Google won.

So what do you think? Do you agree with that prediction in terms of how this is going to play out for Apple? I do. And this is the thing that I think I shifted the most on after watching WWC from what I wrote before. I was still in the, okay, in this case, Apple's probably paying because it's expensive, blah, blah, blah, blah, blah, the huge user base. They need this capability. That's not the impression you got from WWDC. They, you know, no, actually all the capabilities that matter are going to be compelling and are going to sell our devices. We are delivering. We feel confident our ability to deliver it. And sure, okay, chat,

bots are nice. You might actually want to buy some plants. Okay, we don't believe you, but, but okay, you might want to buy some plants. And so look, oh, here's an opportunity. Who wants it? And that who wants it is, uh, my now assumption, and I don't know this to be clear. My assumption now is Apple is not paying Open AI. It is just a look, here's an opportunity. And from an open eye's perspective, I think it makes sense to seize this opportunity. And this is by the way, I think a fairly bearish indicator on the sort of potential. of this AGI that does everything. If opening I was confident they were coming out with sort of a capability soon that just

obviates sort of everything before it. Yeah. Yeah. Then you don't need to be doing deals like this. People will come to you. If you're going to be Google in like, you know, 1998, you don't need to be doing these sorts of deals because you're just so much better than everybody else. I obviously Google did do those deals over time, but yeah, you get my point.

Right. It does feel like though, like this is what I, you know, one of my. somewhat controversial calls, but I think it's more and more sort of, it seems to be more and more the case with Open AI is being a consumer tech company. Like, look, our biggest asset, at the end the day, sure, there's some things we're good at, quads good at, Gemini's good at. Actually, our biggest asset is ChatGPT, as in the name, as in the brand. You'll notice it's no longer openaI.com slash chat. It now goes to chat gptt.com. Like they are, that is the brand. That is the value. that is what was at risk of being destroyed

with the whole sort of controversy last year. Sure, they could get wrapped into Microsoft, fine for Microsoft, but there would have been massive value destruction, which is the fact that when people, like chat GPT is the Kleenex of AI. Like when people think about it, that is sort of what it is.

And this feels like a real aggressive play towards it. Like right now their website model is, you can use GPT 40 for free. If you subscribe, you can use ChatGPT4, which is still better than 401, by the way. But so they have the freemium model on their website. This is just the exact same model with a massive sort of ability to acquire a gazillion new customers that have never actually really tried this stuff. Yeah.

And you don't need to sign up for it to work. It would just use it. I assuming there's some sort of upsell model, you can sign in with your account and then you'll get access to sort of the models you have access to. So you can imagine they have a chat GPT5. You sign in on iPhone. It uses, you know, chat GPT5. I hope if you sign in, that's when Apple should let you get rid of those stupid warnings, right? Right. But the implication of this broadly is, number one, it suggests Open AI does not, at least the leadership, does not believe they're actually going to sweep the entire market with an AGI that does everything in the near term.

Number two, the way Apple presented it, and in this context, it makes a lot of sense, is, you know, Apple's sort of the big boy in this relationship. Yeah. That's what it feels like. That was definitely my takeaway. Sam Albin did not appear on the video. He was at the event. He did not appear in the video. There was a real sort of we're the ones in charge.

And this, again, I'm tempted to assume this is the case because it fits my sort of aggregation sort of concept. Apple has the users. Apple has the interface. Because of that, suppliers who spend billions of dollars on compute are forced to line up and take Apple's terms or not. It feels like Google did not want to take Apple's terms. Apple's like, fine, we'll take open AI. Just be aware, we're going to funnel hundreds of millions of people in that direction. So if you want to change your mind, it's available at any time. I don't think Apple is getting paid like they are for search. But if I'm right, and again, I'm not sure I'm right, but if I'm right

that no money is changing hands, if we ever get to a state where money does change hands, I think that money is going to go in Apple's direction, not the other way. Yeah, well, and that's why I found the event so impressive is I had sort of calibrated my expectations based on some of the early speculation surrounding the Open AI partnership. And I was bracing for Apple to have Sam Altman in a video and just generally introduce a really clumsy OpenAI and or chat GPT integration where Apple is seating like, 95% of the AI tasks to chat GPT, when in fact, what's actually happening is Apple will execute 95% of the AI features that we saw on Monday, and then 5% will be handled by OpenAI.

I mean, the Apple functionality, this tweet from Ben Bajaran, another Ben Bajaran tweet, Apple confirmed they built a host of their own foundational models and used a host of licensed content to train and fine tune their models. Essentially, 99% of Apple intelligence is powered by Apple's own proprietary foundational models. Is any of this surprising to you? Because all I've heard for the last year and a half is, man, Apple is really behind on AI. I mean, part of the reason I was assuming that there would be more of an equal partnership with Open AI is that Apple just was said to be just totally lost in this space.

But they didn't seem all that lost on Monday. Well, again, we need to see how well it all works. So this is all sort of a caveat hangover all this. But I think that was definitely the perception, I think, a year ago. If you've been paying close attention and I think the big thing is you saw sort of the llama models come out. Then you saw Mistraw models come out. Then you saw the DBX models from Databricks come out. and there has been sort of this general march of,

look, this isn't that hard. Like, especially to get GPT 3.5 sort of level. Yeah, to be on the cutting edge is hard, but to get sort of a good enough model that does, particularly when the use case is very constrained. Like, again, this is the whole bit why it feels unfair. Apple's not trying to do crazy complicated stuff. Their leverage is not that we have the best model,

is that we have a good enough model with the best data. And data is sort of what matters. And so not in terms of the license data to train, but the data that actually goes into the question, they have your email of your mom's arrival. The data that people have your text messages. Yeah. That's right.

And they have that layer at the operating system to sort of get access to all that data, sort of what it might be. And so I'm not too surprised at that. I'm actually sort of beating myself up because I think this was, I could have been stronger in my statement that Apple is well. You could have seen this coming that they're actually going to do almost all this.

And the other, if there is a mistake they made, Ben here says, you know, 99% of Apple intelligence is powered by Apple. I would have said 100%. I would have set the chatbot integration outside of Apple intelligence. In fact, I would have led with that, set that up, say, we're partnering with Open AI, anyone who wants to plug in, we have an interface. But let's set that aside. What we're actually here to talk about is Apple intelligence.

intelligence, no outside models. Your data does not go anywhere. It's just you and Apple. And it was pretty, it's been pretty amazing to see a lot of tweets. Elon Musk is sort of leaning the way on this, but it's not just him, a lot of folks. That seems to be willfully misunderstanding what Apple intelligence is and what the relationship to chat GPT is. And I think Apple, they could have been even clear.

I thought they were pretty clear. They could have been even clear, but there is a bit where people do have this assumption that Apple would be behind. So they must be relying on Open AI, number one. Number two, maybe there's some willful ignorance here because if Apple pulled this off, if the models are good. And if we're not approaching a stage where models can be independently massive sort of products on their own, but need to be tied into features, then it's kind of scary, but we're looking at. looking at Apple in many respects, re-semening their dominant position. And the more that that happens over time, the harder and harder is displaced them.

And as we've talked about, Apple's not the most popular company, particularly amongst technologists, because of the way they do leverage their position. But, you know, you can still sort of put your, you know, at some point you have to let go of your personal animosity and look at the situation objectively and they look stronger than ever. Tip the cap. Yeah, absolutely. speaking of technologists, yeah, just a really rough day on Twitter for Elon Musk. Clearly did not see WWDC. Maybe he did see it and didn't care. But a very clear delineation

between Apple intelligence and the discrete tasks that are going to be outsourced to chat GPT. And this is probably a smart thing with doing that warning now. You know, Gruber said they're in like sort of a Q&A session. Someone asked Apple if they could. turn it off. And they didn't say absolutely not. They're like not at this time. We should be clear for anybody who didn't watch the presentation, anytime you want to use chat GPT, Apple will hit you with a prompt saying, do you

want to, or Apple will suggest chat GPT and say, do you want to share this data? If it's a photo of my back deck, do you want to share this with chat GPT? And that will happen every time. Yeah, like what's what's a dinner I can make from this ingredients? Do you want to share this data with chat GPT? Yes, of course. It's just what Ben is pre-annoyed about, just to be clear for everybody.

But also, it underscores that the data is being shared with Apple as a rule and not chat GPT as a rule. So Elon's concerns and talk of putting every Apple phone in a Faraday cage if you're on campus, it's like, relax. That's not actually what's happening. But that doesn't necessarily matter to Elon. So here we are. Well, the funny thing is, you know, a lot of people are motivated by anti-Apple sort of animus. Elon is anti-open-A-I animus. So, like, Apple is a bit of a passenger in this car wreck or whatever.

Yeah. Well, car wreck, maybe bad choice of words. So video. There you go. All part of the fun here. So Rav says, speaking of other non-Elon technologists, and you mentioned the developers, I'm curious about the developers. So this is an email from Rav.

he says Apple completely nailed this event. Apple presented products which consumers will utilize and care about, not a bunch of technology jargon. After the event and seeing some of the examples, I'm actually a bit more concerned for Google search and utility consumer internet apps in general. Assuming the revamped Siri performs to the demoed capabilities, a user can simply ask Siri for miscellaneous information

or schedule an Uber without ever opening those apps. Clearly, this will depend on the API capabilities Apple offers developers, but developers also have an interesting choice here. If Apple's integration is so great that users open the app less frequently, it's going to hurt. So, Ben, stepping back, what do you think this means for third-party developers? And how do you think that relationship will look as we move forward into the Apple intelligence era? Well, this is why it's useful to look at structure, right? If you have competitors in your space that are going to implement this functionality, then you're going to feel pressure to do it. If you have a high degree of integration, that means you don't, then you won't.

Like go back to when Apple did the Apple TV app sort of in 2016 or something like that, where they wanted to say, look, we'll surface all your video in one place. It'll make it so delightful. And Netflix is like, nope, not going to do it. You know, if people want to access Netflix content, they will do so via the Netflix app. and people were very upset about it, but number one, they didn't need to do it because they had different,

you know, the beautiful thing about video is it is differentiated. Unlike, say, music, wherever it has the same sort of stuff. And number two, how much in a worse position when Netflix be today

had they done that, right? They didn't own the interface. Like if you want to do advertising, you need to know the interface. The power, aggregation power comes from recommendation engines, from like helping people discover content that they want. And so in that case,

Netflix could push back and say, no, there's going to be a lot of, developers and a lot of things that can't say no, right? If you're like a weather app or something like that, I mean, that's a bad example because Apple has weather. But if you're, you know, there's lots of things where you're going to have no choice but to do it because your customers will leave you. Like this is why Apple has the entire world telecoms industry working for them. Because you can, you know, this is one of the, there has to be some sort of Harvard business case

study about that. I wrote about the very early history. It remains one of the most remarkable things in like the history of business, which is because Apple could bring to bear its customer base, who for the first time would switch carriers to get the iPhone, Apple would enter market after market with the second place carrier or the third place carrier. That's why they started with singular and not Verizon. They go to Japan.

SoftBank is a distant third. NTT is like no way, you know, we're never going to do that. We've been investing in our ecosystem for years and years. We have the most innovative phones in the world, which have been the case before the iPhone. And in every case, Apple, because what happened? People switched. They wanted an iPhone. And that meant that every telecom carrier at the end of the day, they were reduced to being a commoditized provider of bandwidth.

And sure, I have opinions on which sort of carrier has the best sort of service. But that's not, but how do you compete in that? You compete in that by spending more money to build more cell towers, which is really expensive. It's way cheaper and more capital efficient to differentiate based on, oh, we have this neat, we sell these unique phones that access our e-mom service or whatever it was sort of called back in the day. And Apple took that all away. And that's, you know, this is just sort of value chains and sort of business 101. Like to the extent people use Siri and TBD, Siri definitely has to prove itself. But if people are using this and they're like, I want to call a car this way, I'm going to, you know,

And with supports it and Uber doesn't, Uber's going to have to get on board. Now, which use cases this manifests in to be determined, right? The whole like call a car sort of thing, for me, I can't imagine ever not wanting to go in Uber and specify and see where it is. Knowing that I'm being wide to. It's going to be 10 minutes later than it says. But whatever, I still want to have that aspect. It's like people doing the travel down. I had that exact thought.

And like, even if I could do this with Uber, I don't think I would do this with Uber. But that speaks to what Apple is doing here, though, is so many of the chatbot demos, to your point, are these farcical scenarios that actually no normal person is going to trust AI to do for a very long time to come. Yeah. What you need to do is do stuff that is genuinely useful and saves time. That's why I love the, what time does my mom's flight arrive demo? Where was the reservation? How do we get there? And it's pulling from messages and it's incorporating data from like the flight aware or whatever.

be and it's pulling from email. And what that's doing is I could do that. But it is pure busy work to figure out, was that a text message? Was that an email going through and searching, do his XYZ? That is such a crystal clear manifestation of solving a problem I could do, but just making it a million times better. It's not doing stuff on my behalf. Yeah, it takes three minutes to do and people procrastinate it as long as they possibly can. And it's something that everyone experiences. She would, I would probably just end up calling my mom because it's too much of a pain to find the information, right? And it's such a great, great example. Yeah, and they will streamline it. No, Rav's point, assuming the revamped Siri performs in the demo capabilities, we've come back to that a couple of times. We should be

clear. We'll see whether Apple could actually execute on all this functionality. But I think what I'm hearing from you is that, yeah, this could make life more complicated for companies that, don't have like a dominant market position or enough confidence and and differentiation that they can say no it's in theory 101 like if you own the interface where you where users do stuff suppliers have to sort of fall into line like you have to be very highly differentiated for people to go directly to you and that's what you have to be focused on because that is the only way to sort of sustainable margin is people going to you and you control the pricing as opposed to you have to slot in to where Apple Danes to sort of give you space.

Right. If Apple nails this, they will have even more leverage. So, Rav, your concerns are well-founded. One technical aspect of the discussion as we try to wrap our arms around everything that happened on Monday, I'm going to read a thread of tweets from Matthew D. Green, and this is longer, so just get comfortable. Apple has introduced a new system called Private Cloud compute that allows your phone to offload complex, typically AI tasks to specialized secure devices in the cloud. Apple, unlike most other mobile providers, has traditionally done a lot of processing on device. For example, all of the machine

learning and OCR text recognition on photos is done right on your device. The problem is that while modern phone neural hardware is improving, it's not improving fast enough to take advantage of all the crazy features Silicon Valley wants from modern AI, including generative AI and its ilk. The fundamental limitation here is not the processor, it's memory to if the larger the model, the more memory you need. You wouldn't put that much memory on an iPhone size device. Or the battery life would be basically zero.

So that is, the constraint is memory just to sort of reemphasize that point. Okay. And he says, this fundamentally requires. servers. The solution Apple has come up with is to try to build secure and trustworthy hardware in their own data centers. Your phone can then outsource heavy tasks to this hardware. Indeed, if you gave an excellent team a huge pile of money and told them to build the best private cloud in the world, it would probably look like this. But now the tough questions, is it a good idea and is it as secure as what Apple does today? And most importantly, can users opt out

entirely from this. I admit that as I learned about this feature, it made me kind of sad. The thought that was going through my head was this is going to be too much of a temptation. Once you can safely outsource tasks to the cloud, why bother doing them locally? Outsource everything. As best I can tell, Apple does not have explicit plans to announce when your data is going off device. You won't opt into this. You won't necessarily even be told it's happening. it will just happen magically. I don't love that part. So Ben, you and I both found this Twitter thread interesting,

but you're the tech expert here, and this is getting a little bit more arcane. So I'll cede the floor to you. What do you think about private cloud compute and what questions do you have about how this is going to be implemented? Oh, well, I mean, I think it's great. I mean, I think it's a very sort of clever solution.

I'm still very curious about what this architecture actually look like, looks like, is this computation actually happening on Apple chips or is there sort of Apple chips sit at the front end and do the sort of handshake and assurance of security so that your phone says, okay, yeah, this is a valid to do some sort of key exchange. We're sending this information completely de-enotomized. It goes through a third party, routing providers. You can't even trace IP gets on their computes and then the system is flushed and restarted and someone else can sort of come along.

But that actual computation is happening on Nvidia chips or AMD chips. probably not in video, Apple and Nvidia do not get along. AMD chips maybe, right? I think that probably is how it works. It would be my guess is there's some sort of front end system based on Apple Silicon with a fleet of like AMD chips behind it. Maybe it is just being done on Apple Silicon. I would love to know more.

The details are not that important, though. I think this concept of, you know, the idealized computer experience, experience is the reality is basically everything we do. I mean, I appreciate sort of Matthew Green sort of idealism here. But the entirety of mobile is actually mobile plus cloud, right? Yeah. The reason why Intel stock could go up for a decade in the mobile revolution is not because they were selling mobile chips. They famously missed mobile, but because all the cloud providers were buying astronomical amount of chips to run the services that were accessed via mobile. Turn off internet on your phone and try using it.

It's like it's debilitating, right? It's like every you just keep doing so. Oh, yeah. Oh, no connection. Like almost every app that you have beyond sort of like maybe some games or something like that are predicated on having sort of a consistent connection. Right. From my perspective, this is a natural extension of that. But Apple's whole point is look, where are our differentiation here is we are going to make AI better by.

delving into your personal sort of stuff. And so because of that, we're going to make sure it's not going to open AI. And it's not going to be interrupted in transit. And we, even if we were served a warrant, we can't access this stuff because the way it's architected is nothing is preserved or is tied to sort of an individual or whatever it might be. And I think that makes sense. Like this is, if this is what your point of difference is going to be is we're going to use your personal data, the number one, you need the brand to back it up. Apple does.

and it's worth it to spend the money however much it costs. I'm sure this inference cloud Apple's building is much more expensive than a bog standard, get a bunch of Nvidia's sort of GPUs and set it up, even if to the extent they're saving money by using their own chips or however it might be. You don't save money just by getting cheaper chips if you lose it all because they're less efficient and they don't work very well, XYZ. This is worth investing in. And I think the fact that Apple intelligence is going to be free as sort of shown on stage.

And I know you have a tweet here. Maybe they'll have a premium service at some point. That's possible. But by and large, everything they showed is going to be free. These servers, this infrastructure to service all of the iPhones in the world is not going to be cheap. And they're doing it in a more expensive way than they need to do. But that's worth it for all the strategic and product benefits we talked about before. This places, this positions them so well going forward and is going to drive upgrades,

is going to drive people into their ecosystem to an even greater degree. And so it's worth it. You know, Apple, yeah. So, I mean, I'm not, like, sure, you probably should be able to turn it off. That's sort of a reasonable sort of thing. But there is a bit where, you know, I too miss the days of truly open computing hardware. You can run your own sort of software and sort of the Richard Stallman's sort of view of the world of like every bit of software. My computer is open source and I know what it is.

I appreciate that. To expect that from Apple is completely unrealistic. They deliver products to people, not idealists. And the idealists have, you know, that's so that that's the long and short of it. Right. Well, and I would say, I mean, the point on trusting Apple to make the cloud computing decisions is ultimately meaningless to me as a customer. Because as I've said in the past on this podcast, I'm just not precious about data access. Yeah, we might be the wrong set of hosts.

Yeah, I don't know. The horses so far are the barn. What do you sort of concern about? Exactly. How am I supposed to retain my sanity in modern life if I'm worried about who has my data? But the reason I found that particular thread interesting is as a juxtaposition to the persistent prompts about whether you want to share data with chat GPT, the one point that's easy to forget in the midst of all the privacy talks. and data security from Apple is it's all predicated on the reality that you will be trusting Apple with all your data and Apple's power as a centralized entity that just knows everything about everyone is only going to be further entrenched as we go here. And again, I don't have a huge

problem with that. I'm not worried about it. But it's similar to the ATT conversation where Apple can track all sorts of different things you do on your phone. It's the third parties that they're concerned with. just a fun little reminder. The AI stuff that I am most perturbed about is stuff like I mentioned like I like rewind. That's sort of everything on my screen, right? They have their changes name to limit list and they're going to have a pendant where you can wear this pendant around and records everything that you say so you can access stuff later.

Like, hey, it's great on the computer. Why don't I have in the real world? And I kind of want to give them the middle finger. I'm like, no, the last thing on earth that I want is to feel that people around me all the time are recording. I'm very concerned about, yeah. But it's going to happen. I would like my sort of idealized world is there's a very clear bright line between the online world and the offline world. On the island world, track away, do X, Y, Z.

It's just the reality of computers. They spew out data constantly. That's why it's just a fight to stop tracking because it's like fighting against the type. Well, and it's directly correlated to functionality. That's what they do. Yeah. Like you need to share the data. And so things like facial recognition in the real world, speeding cameras, like all this.

That's the stuff that I'm like, I wish there was a bright line there. The stuff that gets me sad and disappointed is not that Apple's going to deliver better digital functionality and they're going to do their absolute best to sort of lock it down. But while we're really losing something, it's like now once you decided to use computers, you were kind of signing up for this reality in the fullness of time. I don't like the idea of talking to someone and finding out was being recorded. Like that's pretty irritating. Unfortunately, that is coming. Well, I will note that Christina Warren tweet here on the rundown, prediction Apple is going to make Apple intelligence a service Apple intelligence plus

that they'll charge $20 a month for enhanced features and whatnot. That was another light bulb moment. The presentation itself was a light bulb moment. Oh, this is what AI looks like. And then when I saw that tweet later in the afternoon, I was like, they probably will charge for that. And, you know, $20 a month. Well, not just that, but like the more apps that have sort of inference at their core,

which is an ongoing cost, they're going, like the subscription model, until we get to world, GPU cost being zero. Right. It's expensive. Or could be at support. Yeah. So you're going to have subscriptions. And guess what?

You sell, you know, or Open AI, for example, right? Oh, this is so great. I want to upgrade to an Open AI subscription. if you do that on the iPhone where they acquire this customer, you're going to be doing it through in-app purchase, I'm sure. And Apple is going to be taking 30 to 15% forever. And so that's the other bit about this whole AI revolution. It's very good for Apple from an app store perspective.

Yes, the services revenue will never die for Apple. All right, one final question from Nick. And this is a bit of a haters take, I would say. He says, hey, guys, love the pod. So that's nice. As an Apple fanboy, I have three questions for you. One, this seems like a change for Apple to announce a major feature at WWDC that isn't even compatible with its most recent base model phone. I feel like in years past, they would have waited to announce this alongside the next generation iPhone in the fall.

Was Apple motivated by Wall Street expectations to demo their AI strategy early at WWDC? If so, that would be a departure. This is the most obvious WWDC content ever. WWC is about software. It's when they announce this sort of stuff. They don't do major software announcements at iPhone announcements. Those are about hardware. This is, sorry, by apologies, Matthew, but this is, or Nick, Nick, Matthew is getting strays.

Sorry, Nick, this is a bad take. Yeah, apologies to Matthews also. And then he says, you guys rightly ragged on Google for presenting vaporware in their keynote. and to be quite honest, I felt a similar twinge with this keynote. Lots of down the line and in the future talk. Heck, the feature isn't even shipping until this fall, parentheses in beta nonetheless. Oh, boy, what a sneer. Another thing that felt unlike Apple, did you get similar vibes?

Did you get any of those similar vibes, Ben? Or were you too pleased with yourself for nailing the article on Monday morning? As I mentioned, Apple's been doing this more and more. I actually think that's a good thing. Software and services are difficult in holding yourself to a fall deadline leads to buggy releases that drives everyone crazy. So I don't have a – WC has shifted more over the last few years to this is what's coming down the pipe this year as opposed to a singular release. And that's sort of the reality of software. The other reality is that Apple generally ships, they say they're going to ship.

And so they get grace for doing this, whereas Google does not. And there was the bit about the recent Google one in particular was the project Astro or Astro or whatever it was, which was basically what Chat Chb-T had showed the day before. The contrast was sort of pretty stark. Now, to be fair, opening eyes not yet shipped the voice broadly speaking. But that was, you know, it's not a hard and fast rule on what is announced. Microsoft and Google, Microsoft less so now, but particularly in a decade ago, they did this stuff all the time. and it would not ship or look a lot different when it did. If all this stuff that Apple showed today does not ship in the next year or it's dramatically diminished, like it was a huge deal when Apple talked about that charging pad.

Right. And then they didn't ship it because that's not what Apple does. Apple ships the stuff they say they're going to ship. And if they don't, we'll be critical here on the podcast. We've been very critical of the Vision Pro over the last six months because they haven't really shipped anything new on there in terms of content that anyone might find. compelling other than the gruber talks show. Yeah. Yeah.

Yeah. Yeah. Yeah. Yeah. Well, it's funny because John Gruber is saving the vision. Pro. Putting it on his back. The question does hit on a sensation that I had where it occurred to me that just

comparing my own reaction to Google I.O. and WWDC is sort of like the flirting versus harassment meme where they're leveraging control of the OS. The other thing that Apple. is leveraging is that it has a huge user base that loves interacting with Apple products. And so when I see AI being integrated on a screen that I use every day, it's just a lot more compelling to me than what I saw from Google. And maybe that's unfair.

That's what Apple does. They ship products. They don't ship sort of just the technology. And again, this is an Apple fanboy's keynote. Like, in that, and I don't mean that derogatorily. No, this is how Apple gets fanboys because they do stuff like this. And so, again, they have to deliver on it. But yeah, but in short, you didn't read it.

But in Nick's note, you question the latest model bit sort of thing. No, it is RAM. Like we've been talking about RAM for a long time. I think if Apple could go back, they wish they would have shipped eight gigabyte phones before. But this is stuff that's specked out like years in advance. Like, like, and so, you know, if anything, they're fortunate that they were able to get 8 gigabytes into the pro, it would be worse if they

had no phones with it. That's also their best selling phones. Most people do buy pros. I mean, a beautiful position to be an Apple. Like, people will just buy the highest in one. There's reportedly a super high-end phone, or that's one of the rumors coming this year, which Apple fans will just default to. It's good. It's good to be Apple. Breaking news here on Sharp Tech. No, they did a really nice job after me, sort of making fun of them a couple weeks ago. I was very, very impressed. But, and I guess that now makes me one of a million Apple fanboys. We've given a lot of praise. Yeah, we've given a lot of praise and hype to Microsoft, but I mean, they made Microsoft look pretty bad, I think. Like Microsoft looks like this dependency on

opening eye feels much more extreme than Apple. And I think that is a bit of a branding problem. Like Microsoft's on-device stuff is using Microsoft's models. They're five models. And, you know, but their perception is certainly much more of this sort of being tied to opening eye. But there's also just this bit of, you know, and a lot of the co-pilot stuff, Microsoft has shown, does this theoretically at looking your emails and finding stuff. It's just like Microsoft is like a developer company that they're great. Like everyone in the app store wishes Microsoft running the app store because they're really good at that. And Apple's a product company.

And they're really good at making products and selling products. And that reminder pops up sort of again and again. again, and I think it felt like that sort of popped up the last couple weeks. I don't know, man. I'm not going to write off the possibility that Microsoft is able to draft off the momentum from Microsoft Paint 2024 and some of the TikToks. Hey, it was awesome. I still want it.

We'll see what happens. But Ben, for now, I hope you have a great rest of the week. People can send us questions about Apple and anything else to email at sharptech.fm. I enjoyed this recap. and I'm sure that we will be discussing it more next week. But until then, enjoy the weekend, and I will talk to you soon. Talk to you later.