Big Technology Podcast - Who Wins The AI Superapp Battle?, Apple’s Consumer AI Victory, World Cup Automation Mistake

Episode Date: July 6, 2026

M.G. Siegler is the author of Spyglass.org. Siegler joins Big Technology to discuss the race to build the AI super app and which companies are best positioned to win. Tune in to hear why OpenAI, Anthropic, Google, Microsoft, Meta, and Apple are all converging on the same idea: an AI interface that can handle more and more of your computing life. We also cover Apple’s new Siri, whether consumer AI will be won by default on the iPhone, and what World Cup automation says about our growing reliance on machines. Hit play for a sharp, wide-ranging conversation about where AI products are headed next. --- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. Want a discount for Big Technology on Substack + Discord? Here’s 25% off for the first year: https://www.bigtechnology.com/subscribe?coupon=0843016b Learn more about your ad choices. Visit megaphone.fm/adchoices

Everyone's making a super app. Who's in the best position? Has Apple already won consumer AI? And does automation ruining the World Cup show that even if AI makes things worse, we still can't help ourselves. That's coming up with M.G Siegler right after this. In the face of ongoing disruption and opportunity, TMT leaders need to deliver tangible results, not just ideas.

When pace and performance matter most, PWC combines market insights and deep sector experience with AI, cloud, and emerging tech to accelerate your transformation and drive measurable ROI from strategy to execution. PWC can help you anticipate what's next, outpace disruption, and compete. For more information, visit pwc.com. Welcome to Big Technology Podcast. It's the first Monday of the month, and so we are joined, as always, by M.G. Siegler and his first Monday of the month appearance here on Big Technology podcast. We have so much to speak about with you.

Today, including the latest on the super app battle, whether Apple's victory in consumer AI is complete total and unimpeachable, and whether the World Cup is ruining soccer with automation in its replays. So, MG, it's great to see you again. Welcome back. Thank you, Alex. Excited to talk sports. Never rarely get to do that. That'll be exciting. Well, it's interesting because Marty Swant, who writes for Big Technology on Mondays, wrote to me, and he's like, I wish we could write about the World Cup. I was like, of course we should write about it. It is like AI or automation, whatever the technology label you want to apply to is destroying these games. And so we'll have a piece out on that. And I think you and I will have some interesting conversations about it, especially since you have Norwegian blood in you, which I just learned.

And I live in England and oh boy, this week is going to be interesting. At 10 p.m. it's on this time. So last night's game was on, you know, in England time at ended up being what, 2 a.m. because it was delayed. But this coming one against Norway will be 10 p.m. The pubs will be crazy. It'll be awesome. So if Norway wins, given that you have Norwegian blood and you live in England, do you row with one hand or how does the celebration work? Okay, so folks, that's just a preview of what's to come. But let's talk about the summer of super apps, as you're calling it, MG. So first of all,

A little news that we covered a little bit on the Friday show, but I think we should spend some more attention on is that Microsoft is now emerging its consumer and enterprise co-pilot apps into a new unified app that will feature AI coding tools and new AI agents that customers would need to pay extra for. That is according to the information. I'll just read this brief note from Spyglass that I think will kick us off into it. You wrote, I am shocked, shocked that the strategy of having 17 different versions of co-pilot across 17. different surfaces, including, yes, services with a capital S, hasn't worked out for Microsoft. Delivering one co-pilot is obviously the right approach and messaging, and yes, that puts Microsoft in the camp with MetaX, Coinbase, Airbnb, Uber, Snap, Spotify, even Disney is going all in after the elusive super app concept and assuming the timeline above is right.

They'll be in a foot race to get their take out the door with who else Open AI. So, I mean, let's just start broad here, MG. I mean, it seems like if you think about it, we talked about this a little bit on Friday, but Microsoft, Google, OpenAI, Anthropic, meta, anyone with an AI play is trying to make their own super app, which is basically an AI interface for all computing that you would do, whether that's on your computer or on the web. I don't know which question to ask you to begin with. Is this the right strategy? is there a company that's going to stand out and its ability to deliver this?

So I'll just throw both at you and get your take. Yeah, the original genesis of this sort of thought was, of course, the realization that all of them are working on sort of, quote unquote, super apps. But it's interesting because they all view it slightly differently. I think you talk to Greg Brockman about like Open AI's viewpoint on it. But they're all coming at it from a different angle, right? because OpenAI obviously has chatGBT right now. And we've talked about previously, they run a real risk of sort of changing the dynamic

of how people have used a billion people now using it, the fact that it's been sort of this thing that popularized the notion of a chat bot. And of course, over time, they've brought more stuff into it, like images and a few other things here and there. But obviously, they're sort of now in the position where they need to chase a little bit anthropic.

and Anthropics model has been different, sort of led by coding, cloud code. And so, you know, obviously Open AI is codex, but they're going to merge them together. You know, imminently, it seems like it should be any day now that they actually do this. And then, of course, their web browser, which is the different element. So they all have sort of this unique sort of element. Like maybe Open AI has the web browser part, which Anthropic doesn't have,

you know, Microsoft, as you noted, like had all these myriad different plays that they were going for with co-pilot. It's easy to say in hindsight that it never made any sense. But come on, everyone knew that Microsoft was not going to be able to pull off 400 different versions of co-pilot. It was so confusing when they launched their original batch of AI PCs, which were called Surface Plus, something like that. Whatever plus PCs, awful branding, as always for Microsoft.

But it's like people just wanted one place to go to. And again, ChatGBT, GBT had already made the notion of sort of this being built around a chat bot popular. And that, by the way, of course, is, you know, and we'll get to this, is what Apple didn't want to do either. They thought, like, there should never be an app, you know, and there shouldn't be a centralized place for this. But again, chat CBD popularizes this. They're now moving beyond that to try to match Anthropic, as mentioned. Microsoft trying to come in to match, you know, the major players and actually have a cohesive offering.

Apple trying to come in and leverage Google's technology to build the best consumer-facing version of this, and so on and so forth. There's so many different, again, people going after this notion. And as noted, it's not necessarily the same thing as what they're going after in the Asian markets, right, where super apps, you know, I think primarily have been built around like commerce and sort of everything in your life. This is all just the super app to be able to try to win AI. And you would think, again, that that's mainly from a consumer standpoint. But again, Open AI is there with ChatGBT.

They've been winning that space. And now it seems like, you know, they feel like for business purposes and, you know, the potential of eventually IPOing that they need to go after Anthropic on that business site. And then, of course, there's Google and everyone else who's all in there too. Yeah. And I think like the way that I read it from Open AI and all the others is that they effectively see it more as.

like a laptop versus like an application. So, you know, in a laptop, you would use it for personal use cases, like the same laptop. You would watch Netflix on. You could also use Excel and PowerPoint on. The same thing, you know, is going to be when it comes to these super apps, you will be able to like instruct it to do business things for you, but also instruct it to plan your life and coordinate your kids' soccer schedule and send text messages to the person that's supposed to pick the kid up from the game and then, you know, realize, you know, have it suggest to you that there's actually, you know, a two-day gap in your schedule when the kids have off. And so therefore, do you want it to like buy you some movie tickets and it'll have your credit card and go out

and buy the tickets? So it's like it does. And you're sort of getting at the one thing that I left out, which is, of course, the notion of open claw, which of course we've talked about previously, took the world by storm for at least a moment in time there. I mean, it's still out and about, but everyone's going after them, right, even if they're not explicitly saying it, because they think that that model is interesting, though they don't like the way that OpenClaw itself did it by its open nature and all the security risk and things that we've talked about previously. But yeah, and so that's another layer of this, right, and why it has to be a standalone application, because it's not just about the chatbot anymore, right? It's not just something you can necessarily do within a web browser. Right. So exactly right. So the OpenClaw use case is now every use case. right for for AI effectively it all seems like it's moving to that it doesn't mean it won't answer

your queries in chat it will but all all AI is going to this open claw use case which is again like you basically delegate stuff to the bot and it takes care of it for you yeah so so let's before we just go into so we've set the table now before we go into you know um who might be in the best position to win here can we just talk about whether this is a good idea because again, this is, you know, trillions of dollars, effectively of investment heading towards or, yeah, well, yeah, definitely over time trillions of dollars of investment heading towards this use case, because all AI seems to be converging on this use case. Is it going to, is it wise?

Is it going to pay off? Ultimately, it comes down to how useful sort of agents and agentic use cases end up being, right? Because, again, right now you can do most everything through a web browser. And again, the notion is that once you move over to a native app or on a phone, a phone app, like that this is all about sort of unleashing those technologies to be able to do things in the real world, like you said, sort of utilizing a computer to the fullest extent that a human can. And to date, you know, I think that the jury is still out, right? Like how useful that actually is.

Certainly there's some use cases that have popped up and certain power users have their things that they do and you and I have our sort of different agentic things that we do, but is the general consumer going to buy into this notion? And again, Open AI has done an amazing job, one of the fastest, if not the fastest growing, you know, consumer product of all time with chatGBT. But that was one very small sliver of what we're talking about now. And I don't think that that was by accident, right? It's like it was a simple thing that everyone could understand. Everyone understands the chat interface. We all use text services. And so what the risk is now is that like it's the boil the ocean problem, right? Like you can do anything with these things. So what are you going to do?

And how these companies build a product to guide individual users, especially those that are less savvy and less at the, you know, forefront of using all these things, how they guide them to use this. Like you can imagine, you know, our parents loading up, you know, the new version of even if they don't, maybe they don't even call it chat, GBT. I assume they will because the brand. brand is so strong, but say they load up this new super app version of chat GPT. And like, okay, so they've heard like you could ask questions, sort of like what you used to ask queries and web search. But do they know that you could get it to like, yeah, organize your email and organize your

calendar and organize your photos and folders and things like that? And, you know, again, maybe they've heard of these things individually. But like, are they actually going to be okay doing that? And should they be okay doing that again? So to me, your question it just boils down to you. how useful and how fast these agentic capabilities sort of come into existence and become like normalized. Yeah. So I think my take on this is yes, this is a good idea and it will happen. But the luxury that the company's building this don't have is the time frame.

Right. So for me, the idea of us using the internet and using our computers in the same way that we are today, 10 years or five years even down the road, it's not. going to happen. I do think that most of it will happen by these like smart AIs that just understand everything about us and that is a lot of trust like you talked about. It's a lot of trust and it's a lot of context to give over to big tech. But ultimately it's going to happen. It's just, you know, even in its clunky version today, it's just much better than these, the way that we interface with computing right now or we use computers right now because these computers or software programs that we built are built for the masses, which means inevitably there's going to be stuff you're never going to want to use. There are going to be things that are not built

for you. And it will be sort of, it's sometimes, you know, you feel like you're slamming your head against the wall to get anything done within something. Like think about just Excel, for instance. How many functions and formats are there within Excel where you, you know, most users, I would imagine use 5% or 10% of it, but it's a different 5 or 10%. So what these things will do is it'll make personal, using computers personalized for you. But you're right. The question is,

how long is it going to take and how user-friendly is it going to be? Now, maybe what these bots end up doing is, they end up saying, okay, let's say to the parent example, you know, they write a simple query and what they can infer is, it sounds like you might be interested in this.

By the way, do you know, if you just click allow all, we're going to just take care of it for you, and that might speed its adoption. Yeah, I think that that, That's right. That's my sense of how it will sort of, at least that they'll try to productize those use cases. Yeah, that they'll have little, you know, intuitive pop-ups, not in your face, intrusive, but above the chat bar, below the chat bar, sort of they do it to some degree right now.

And they can help guide a user along, right, through a flow of what they can actually do to let them know. The other thing, when hearing you talk about it, I'm just sort of thinking on the fly here. But, you know, it does in some ways, it's very different, but in some ways it reminds me of the move to when web browsers took off, right? Because those of us who are old enough remember, you know, using computers before web browsers. And it's like, it's weird to think about now, right? Because basically everything you do is through a web browser. Obviously, there's certain apps like Excel and things like that. But even that, you can use a web browser, right? And so these days, pretty much, I think almost everyone uses web browsers for the vast majority of what

they do on a computer. But before, you know, the era before that, there were all these individual applications and, you know, people would use it for very specific purposes. And so I do wonder if it's, you know, if the thing that ultimately drives it is like one very specific use case, sort of like chat chbtee had chat. If there's a new specific use case that sort of takes the world by storm, I don't know exactly what that would be. But if it's like a viral moment, I don't know if that's good enough, right? I don't know that it's good enough to be like the, you know, studio jibbley stuff or or any of the other things that Open AI has, has leveraged in the past to get people using them,

I think it has to actually be like a very useful thing that you will keep going back to in a recurring use case. And so that's what I, that's where sort of my instinct is going. And then I wonder if, if one of them finds it, obviously all the rest will try to copy it. But is that the right method?

Or should they go for something else and try to, you know, battle on those fronts of like having this great use case? And coding has been that use case, right? like for a subset of the crowd. That's obviously what got Claude code, you know, out there and codex now to some degree. And it's interesting because obviously they say, I think Brockman again was saying with you, that like while most people view codex as a coding tool, they view it as the sort of super app, right? That's what it's going to be more like codex than it is like chat chbt,

is what I gather from what they're saying, basically. And so it's interesting that coding was that initial use case. But is, again, is that good enough sort of for the masses? Most people, even though it's easy, most people won't have a use case for sort of coding unless you move coding into like the more nebulous like recipe idea, you know, sort of like Apple shortcuts or something where it's like it is coding, but it's not really the way you would think about coding, right? You're putting together an automation for your home.

And yeah, it involves coding. But you don't have to think about that. Don't worry about that. Yeah, yeah. That's where I think it's going to go. To me, it's going to be coding. coding unlocks the use cases, but most people won't see any of the coding happen.

It just enables like when you code, you basically root into the sort of core of what your computer can do. And so it will be just a means to an end to accomplish things for you. So, okay, so I agree with you. It seems like, you know, it's hard to imagine there being this like one viral moment because the potential use cases are so vast, right? So it's going to be slow.

But that being said, you know, Again, we think about the companies going after this use case. We talked about Microsoft, OpenAI. I mean, I'm just going to list all the companies we cover on the show. They're all going for it. So I'm curious to hear. Even Apple. And even Apple, and we're going to get to Apple in a big way in a moment.

So let's just put Apple aside for a minute. Who do you think out of the non-Apple companies has the best chance of delivering on this? And then what happens to those that fail? Because, I mean, if you're low on the outside, looking in here, it's going to be a pretty rough existence for you, I would imagine. Yeah. So my instinct there would be, I do think that Anthropics still, you know, holds the lead with regard to capabilities, thanks to, you know, mythos and now fable. And obviously, it's back out there in the world after a hiccup. And we'll see if that,

if that stands. But like, if they maintain the lead in terms of, like, cutting edge capabilities and specifically for a Gentic use cases, I think that that gives them. an obvious inherent advantage. Now, there's OpenAI and Google are both said to be preparing, but they've been a while for preparing their new models, and they're not out there for whatever reason. Maybe it's governmental. Maybe it's something else that's sort of holding them back. But if they can catch up there, assume that they can catch up at least to some degree, then it comes back to the products again. And then I go back to the original notion that we were talking about, where ChatDBT has a huge built-in user advantage.

and presumably open ad is going to leverage that when they roll out super app. And so say they have roughly a billion users, Anthropic for all the great stuff that they've done with Claude, like they don't have nearly that scale. And then Google, what do they do there? So right now they obviously have their own standalone Gemini app that they felt like they needed to launch after it didn't, you know, just trying to do it in the web didn't work.

And for these agenetic use cases. But it doesn't seem like they have a huge uptake in terms of like those, you know, native app. usage, at least on a Mac, which is what they have right now. They say that they have the active user usage, but I assume that a lot of that is on the web and or maybe mobile, which is obviously important and good. But I wouldn't be shocked. And I'm sort of surprised they're not doing this, though I do understand, I think, why they're not. I wouldn't be shocked if Google eventually leverages Chrome itself to try to take over, you know, and actually win. Like, why is it that Chrome

itself isn't the super app, right? Everyone has Chrome. It's, it might be two billion. I don't even know what the current number is, two, three billion people, users around the world. And everyone downloads it. Why is that not the payload to get Gemini in front, the Gemini super app in front of everyone? Regulation and regulators, obviously is one answer to that. And they've sort of been tiptoeing into it. Gemini's baked into Chrome, of course, now. But actually, it's not live in the UK. It's live in the U.S. I think they're very delicate about how they deal with Europe, as many of these folks are, for obvious reasons. But again, so if I had to answer right now, I think the battle, the main battle remains between sort of the users of ChatsyBT and then the capabilities of

Claude and what Anthropic can put out there. And I think Open AI, everything that I've seen over the past couple years and we've talked about many different aspects, like they do a great job on the product side, right? Like they've always done a good job productizing this. And so I have some faith in them that they'll be able to, even though they've sort of been a little bit, you know, passed by Anthropic in many regards, I have faith that they'll be able to come up with a pretty compelling product, I think, product experience to do this. Now, the teams have, there's been a lot of turnover across all these groups. So who knows like who's building what these days, but I feel like that. I have less faith that Google can do it because they have all sorts

of sort of product issues when they roll something out. Great technical capabilities, but product issues. Microsoft, I have even less faith in for obvious reasons regarding product. And again, this is talking about from a consumer standpoint. I do think that there's a, there's an interesting argument and compelling sort of back and forth to be had about. And I wrote a little bit about this. Like, what if it's sort of similar, the way that this plays out is similar to sort of the quote unquote bring your own device strategy at work right so are we going to have like a bring your own AI strategy at work like basically are you going to bring your consumer iAI that you use whichever one because it's probably not going to be multiple for regular users you know us crazy

people use multiple ones but it's probably going to be one that you pick and do you end up using that at work too because that's the one you use at home or is it bifurcated and you basically don't want to use the one that you use at work because you want to separate work and play yeah i'm Smiling because that is going to be a fascinating discussion over the next couple of years about whether you bring your own or whether you use the work one. I cannot see big companies, most companies, allowing you to take your own AI and use it because these AIs, the memory is going to build up. I know, but that's the flip side of it.

Like, if you don't have your memory with you, you'll be frustrated, right? That you've taught it all this stuff to do and then you go to work and it's like stupid. It doesn't remember anything of how you operate. I know, but it's going to be, I agree. It'll be, there will be a case to be able to bring your personal AI into your work setting. But again, like, let's say I'm at company A and my competitor company B is breathing down my neck. And now my star employee brings their AI to me. Yeah, that's interesting.

You know, they're in my systems. And then, you know, company B, now company B is really trying to figure out what's the best way to compete. What company B could then do is make the godfather offer to that employee. employee who not only will they get them. Literally, literal institutional knowledge in their... Absolutely. They're going to get them and their AI with the memory of how to do what things work and what things don't work. There's going to be all sorts of new IP law and things like that about like what can transfer

and what you can keep. That's a good argument against why you can't cross those streams basically. But then it will make... It'll be very interesting also because it will separate capital and labor in a very interesting way when you have companies that, for instance, I mean, Met is doing a very interesting version of this, right? You have your employees come in, they use the business AIs, and I'm an employee. Like when I, well, part of what I do at a job is I, you know, obviously give my time in exchange for money, but I also give my time in exchange for knowledge accumulation,

training, understanding things. And now the corporations, if they're using company specific AI, will actually own that part of it. Like, of course I'll get better at myself, but I won't be, If I'm reliant on AIs, I won't get that much better because a lot of that knowledge and training and up-leveling will happen within the AI. And so then if I go, you know, to another company without that, I'm now worth less than whether I would have transferred previously in a non-AI world. Yeah, I'm having the Don Draper. That's what the money's for me. I'm going in my head. Like there's going to be new versions of that going forward. Yeah. So that's all super fascinating. And again, like you said, it's going to be a major debate that's had over the next several years, it feels like. Because again, like there was some level of debate about that, though, with bring your own device, right?

Because obviously corporations are worried about security. They're worried about people having their own devices on premise and using them for, you know, confidential information and proprietary information. But at the end of the day, at a lot of places, it won out because of convenience. And so is it also lead to a world that's bifurcated by big companies that are savvy in this stuff? No, to lock down the employees, you know, AI usage to, you know, be only their own proprietary stuff. Whereas smaller companies, especially if people are bringing their own and paying for their own AI, they're fine with people bringing their own AI to the workplace, right? And what if that leads to these weird competitive dynamics, you know, in big company versus small company, too?

Yeah, might enable small companies to catch up faster. Yeah. But then if they have a technology and, yeah, more nimble, yeah. Yes. They can catch up faster, but then big companies can sort of rebound more quickly when they when they hire the person over and their and their little AI memory with them. Yeah. Yeah, yeah, yeah.

This is fascinating. So just to go. Yeah, go ahead. Just to riff off of the last thing you said, because I feel bad. We haven't even really brought up meta, but they are in the equation. You mentioned them. And can they do? something interesting. Obviously, they're the kings when it comes to making wide usage consumer

products. And so it would seem like it would be silly folly to write them off in terms of what they can do. And, you know, you saw all the recent interviews with Alexander Wang and stuff talking about how it sounds like they're closing in on what they believe to be Frontier-like versions of their models. I believe that when I see it. Yeah, exactly. but what do they do from a product standpoint? How do they roll it out? Like meta AI right now has an app. It's very weird because it's tied to the Rayban glasses thing

because it's obviously their sort of toehold in the general space and a differentiator that they have right now. But what do they end up rolling out? Like they've tried something that was sort of like SORA, but they did all fake stuff, Sora. And then they've tried a few other angles at AI before and making it work. But it hasn't taken off to that degree.

So they're in the conversation because they have to be because they have billions and billions of users across their products. But, you know, they're not anywhere right now. Yeah, just to put a bow on this, it's going to be a very, and I spoke with Sam Altman about this at the end of the year. And for whatever reason, it's stuck with me. It's going to be a very interesting competition between the AI native companies like the Open AIs and Anthropics and the AI bolt on companies like the Google, where like Google has shown it can catch up in terms of capability. abilities of models. It will inevitably catch up again to whatever Fable is and GPD 5.6 is. The question is, and this is the point you made, whether it will have the courage to, you know,

it did it with AI mode and search. Will it have the courage to take Chrome and turn that into an AI first type of experience? And I think it inevitably, it will be pushed into that. It will have to. It will. It will. Because if Open Eye, in fact, does roll out Atlas, their browser as part of the super app like then i think google says like look uh regulators won't like this but we're we're trying to compete here and they've got a browser that's a i you know first enabled and whatnot and and so we're going to leverage chrome and try to do that and you know that's it's good there's going to be fireworks about that because meta will be behind the scenes you know whispering that they shouldn't be allowed to do that what will apple say about that you know behind the scenes and and all that sort

of stuff yeah definitely all right big battle coming up ahead and now um we're going to go to break, but when we come back, we can talk about, I don't even know if you can call them a dark horse anymore. We could talk about the horse, the stallion, Apple, and its chances to factor in this battle. We'll do that right after this. If you want a $3,000 a month payday for life, what would you feel free to do? Maybe take a long weekend, every weekend, or try a bunch of new hobbies. Would you feel free to upgrade and listen ad free? Don't worry. We get it. Every $20 ticket could win you $3,000 a month for life and supports life-saving cancer research at the Princess Margaret.

Feel free to buy your payday for life ticket today. Raffle number 155-2194. Please play responsibly. Visit BetMGM Casino and check out the newest exclusive. The Price is Right Fortune Pick. BetMGM and Game Sense remind you to play responsibly. 19 plus to wager. Ontario only. Please play responsibly.

If you have questions or concerns about your gambling or someone close to you, Peace contact connects Ontario at 1-866-531-2600 to speak to an advisor, free of charge. BetMGM operates pursuant to an operating agreement with Eye Gaming, Ontario. And we're back here on Big Technology Podcast with M.G. Seagler or Spyglass. You could find Spyglass at Spyglass.org. Highly recommend it. Obviously, when MG comes on, we talk through all the Spyglass stories. And one of my favorite over the past month from Spyglass has been about whether Apple will win consumer AI, default. In fact, I'm hedging as I read it, but M.G. was not hedging as he wrote it. There's no question mark there. Here's what you wrote, MG. Apple is set to win AI, at least from a consumer

perspective. Apple seems on the verge of doing what they always do, watching new products and services come about, and then jumping in later with a better user experience to win the day. And that, of course, is going to be with Siri. So not only did Apple present a more realistic and compelling version for Syria at WWDC, which was, you know, sort of, I think it happened the week we spoke. We gave a nice preview the last time. They've debuted it. Not only did the vision make more sense, but the demos have looked good. I'll admit, I personally have not gone through the process to get it on my phone because the last time I put Apple Intelligence on my phone, it became unusable, like multiple apps just stopped working. So I've decided to wait a little bit. But the videos that you see and the

reports that you see of what the new series can do are actually quite compelling. So before we get into why Apple may end up winning consumer AI by default, can you talk a little bit, MG, about what this new Siri will do when people get it on their phones and whether you've experimented with it, what you're looking forward to? Yeah. So this was, you know, my take after watching WWC. And yeah, it was a bit controversial simply because of the past, right, that Apple has promised, you know, this dating back two years ago and then famously whiffed on it. And then last

year sort of had nothing to say while they were, you know, behind the scenes trying to gather, you know, regather themselves and come back at it. And from everything I saw in this presentation and just my gut instinct having watched Apple for, you know, so many years, a couple decades, and covering them and now, you know, just being deeply ingrained in that ecosystem, I feel like I saw what I needed to see from them, which is that, just as you noticed, like they're running their playbook. They watch to see how others have had success in a certain field. And they say, like, we think we can do this better from a product perspective.

And so they zoom in there. And just from those product demos, which were still, we should say, they were still recorded. But by all accounts, they were actually done in real time because they made an emphasis to have pauses. And, you know, it wasn't just like, you know, these canned demos like they were two years ago. So when I was watching those demos, during WWDC, it was like, I see what they're going to do now because it's basically,

they obviously leverage the iPhone in terms of it being pre-installed. They have a button on the iPhone, of course, that you can use to invoke the Siri itself. And so you just saw Mike Rockwell, the executive who took over that group and sort of rebuilt it, just doing these demos where it's as simple as holding down the button and speaking to your iPhone and it being talking to a chat bot just as you would with chat GBT or just as you would with Claude, except that again, this is built into every single iPhone and iPad and MacOS and every other device, Siri in your AirPods. It's all going to be there.

And so, you know, you hate to call it sort of winning by default. But I think it's, it's almost like this is a negative, also sort of a negative framing of it, like being table stakes. But that's all they needed to do because, again, Not you and I, not a lot of people who are probably out there listening right now, but for the masses who are going to be introduced to these things, this is going to be their introduction to it for the first time to the parents of the world and to the kids of the world.

And I do think that Apple has done enough. And you ask, like, have I downloaded it? Is there something else I can help with? I just triggered. Have I done enough to, or sorry, have I actually played around with it? Yeah. So I have it installed on a backup iPad right now. And I'm actually using very dangerously a computer that's running the Mac OS beta right now.

But they're pretty solid. Like because they've said it's like a quote-unquote snow leopard release. Yeah, just give some examples of what it can do and what you've done with it. So I've just been putting it through the paces in that I now bake it into my routine. You know, previously I was using chat DBT, Gemini and Claude. I would just rotate all, almost for every query I do. I would like, you know, just sort of bake off against them all. Now I rope in Siri in there as well just to see how well it can stand up.

And so it's everything from simple queries, you know, factual-based queries, which Siri famously could not do previously, right? Like there's the famous example of like who won the Super Bowls and it just didn't know it, even though that's like one of the most obvious, easy sort of things to pull from a database. And they just couldn't do it previously. Everything that I've tried for the most part, there's certain things where it breaks sort of phrase on the edges. But everything that I've tried has worked from a query perspective. Now, the more interesting stuff will be as it starts to get into the agenic workflows. A lot of that stuff is not live yet. But that's the other major, major advantage that Apple has because, again, baked into the device,

they're going to be able to, for example, the big one, see what's on your screen and see what you're doing and being able to say like, hey, you know, what is that up in the corner of, you know, this website that I'm browsing or whatnot. And it will be able to look at that, whereas the other services and applications either will require like very, you know, explicit permissions and pop-ups to do that beyond sort of the download of their service. And we'll see what Apple restricts and what they don't. To me, one of the things that I've actually really enjoyed using chat chippy T for or with or to do recently has been, it's kind of a weird use case or a basic use case, but just Gmail search. Yeah.

It's amazing. And the nice thing about it is you can just say, like, who is that PR guy on that email chain? And it will be like, I think it's this person. and then give you a link to that email. That is so useful, especially because, I mean, sorry, unfortunately going back to Gmail, but like a lot of these default searches are not good. No.

And so it's coming from a search company. They should be good, but that's beside the point. But the cool thing to me is with these Siri demos, being able to put that type of semantic search on top of all your apps, right? Because it will have email. It will have text messages and have that work. That's crazy. And that's really where the ad,

value comes in to me. And that's why I'm excited about this. And so a lot of that stuff has started to work now. You talk about like the pain it was when you first installed the original versions of Apple Intelligence. It takes a long time to index again. It's like basically going, when you first install it in this new version, it's basically has to reindex your entire, your entire phone. And if you have a lot of photos as I do, like 80,000 or something, maybe even more, it takes a long, long time for it to be ready. You can use it while it's sort of doing that, but it's, it's, you know, there's stuff going on in the background. It's going to be a little bit more slow. Sometimes the phone gets hot or the device gets hot. But it, and it takes just a long time to do that.

I think it took several days for it to be totally complete, you know, by the time it was all done. But now, of a sudden, you can query, yeah, old photos that you have in your library. And it's all this individual stuff, which, again, is the stuff that was promised two years ago. They had the ad with Bella Ramsey that was like exactly your use case of, like, who is that person? that I met at this thing and so and so. And you can do that now. And Google, of course, is promising the same thing because they have Gmail, as noted, and, you know, several other things.

And so that's like a battle between the two of them. The difference to me there is that pixel, while, you know, impressive, I have a pixel fold. I like it. But it's just not at the same scale as iPhone. And so there's no real competition there. You could say like all the Android ecosystem together, but they're so split between Samsung has their own AI and, you know, all these other services, you know, they partner with perplexity. And so it's all going to be split, whereas Apple's is all going to be dialed into their product,

Siri AI, so it's got a slightly new name. But it's, you know, it's solid so far. And I think that every indication you hear from people who are also using it out and about is that it's solid. And it's only going to get better from here. And yeah, and so then it just becomes a question of if there is some sort of new breakthrough in AI or if there's some new way of doing things, if super apps unlock some sort of new capabilities, does that move the goalpost out and then put Apple behind again? I wouldn't be shocked if that happens, right? And so we then, then it just becomes like the thing that we've talked about previously where it's like, does Apple, are they able to get in line and have the right cadence for the AI age to be able to release things, not waiting for the big marquee event a year from now

if they need to like, you know, step up and sort of do something new to meet the, capabilities out there in the field. Or conversely, do they feel like the frontier is still the Wild West and we don't need to necessarily be playing in that field, let the open AIs and let the Anthropics. And by the way, you can use those products on an iPhone still. So if you want that cutting edge stuff, you can go and use those products. But for the real useful AI, trying to think of how they would frame it, we will have it on our device, basically.

Yeah, I mean, I really liked how you put it. I really agree with this completely. Just as we learned about cameras, when the iPhone launched nearly 20 years ago, the best AI device is going to be the one that you have on you. And at least for the foreseeable future, that's the iPhone, right? So it's sort of like we're talking through all these different use cases. And I totally agree with you. All Apple really needed to do was build that table stakes version,

which is still impressive technology. Attach it to, you know, attach it to the iPhone and a way. you go. Like that that example I gave with chat CPT of like I'm now very happily typing into chat cheptie to find me stuff about my emails, which is then using a Gmail connector to find Gmail and populating it back into chat, which I'm on and either an app on my phone or on the browser, it doesn't need to be that difficult. It shouldn't be that difficult. What should happen is I just literally press an action button on my phone and say, who is that PR person on that email thread? And then it just spits it out and I'm good. And so two other things I would mention.

One, we sort of already talked about the demo where Mike Rockwell is holding down the button. I think that that becomes like a normalized thing. It seems weird now, but I wouldn't be shocked that we see people walking down the street sort of hitting that button to talk to Siri AI and just querying all sorts of things on the fly, probably using AirPods, obviously, to talk out and about. And you could say, hey, Siri, I'm sure I triggered it again, just as you do right now. But hitting that button, it's just like, it's such a simple sort of, you know, mechanism that. I think will become like a thing that you see people out there in the world doing. And then it looks weird at first, but then all of a sudden you realize, oh, what they're doing. The other thing that I think will be big, oh, my little daughter's smiling and waving at me through the window, sorry.

And the other thing that I think will be big is visual intelligence. So this has been around since Apple launched Apple intelligence itself two years ago. And it's actually fairly good. It was good because it was actually using Open AI and Google previously. and so now it's being fully baked into the camera. And I do think that this could be something that, again, all these services sort of have levels of this right now, but because it's baked into the iPhone by default,

I think that you're going to see a lot of people out there either holding up their phones looking like they're taking a picture and what they're doing is just querying and using AI to look at the world around them and figure out different things. And I think that that will be key for future devices, obviously glasses and then AirPods potentially. with cameras and whatnot. And so all of that stuff, again, Apple is so well positioned to be able to take advantage of that all because of the iPhone.

Right. And so one note that I made as I read your piece was this is the biggest indictment to me on the value of foundational models that Apple was able to get Google reportedly for what was it. It's a billion dollars a year reportedly to help them build a functional app. or a functional operating system effectively with this technology built in. And why would Google do that? It's not in Google's interest necessarily to help like its number one handset competitor

to build a better experience, AI experience than they have, which is what this might be. But as you point out in your piece, it's becoming commoditized, right? It's, if Google doesn't do it, open AI would do it or Anthropic would do it or who knows would do it. And so,

So like you get to this point where you're like, well, what is this? What is the value of foundational models? And before the break, I said they want it to be like a laptop for you where you can use business and personal uses. Great. What's the one parallel they don't want with the laptop? The fact that the laptop effectively commoditized and up until this recent memory crunch, which is, you know, ironically is being spurred by AI, the prices drop through the floor.

And so why wouldn't the same happen here? Yeah. Yeah, and so Apple was so adamant talking about how, you know, they mentioned Google. Some people thought they wouldn't mention Google at all, right? But they did mention Google and they were fairly straightforward about it. But they mentioned them in like that they were partnering with them, but so explicit that they did all of the heavy lifting themselves. They basically got, you know, the underlying Gemini technology and didn't use the word. I don't think they used the word distillation because I think that it's like something that they would assume the audience wouldn't care about or know about necessarily.

but basically it sounds like what they did. Or they'd be embarrassed to admit. And embarrassed, yeah, sure, right, to admit. Like Elon had to on, you know, in the courtroom, right? That they basically were using other models to distill to get grok to work. And so Apple, you know, is doing that with Google. And but also it's part of their just whole mantra in messaging to say like, this is not just Gemini rescind where we built this from the ground up with Google technology

to build like effectively a new version of Gemini that we're not calling Gemini that we're calling Siri AI and Apple Intelligence. And it will have new different capabilities than what sort of Gemini focuses on and what Google is focused on. And so they went over this. They hit this over over and over again, including after the keynote, they had a separate event to just really talk through these details. And so, yeah, Google, I think, you know, obviously they have the existing search partnership

in place, which is super, you know, loose. lucrative, I think, for both sides. And I think, you know, once everything sort of played out over those past two years, you know, were they willing to just go down this path again and sort of re-up the partnership and sort of do it for this new age, sounds like it. And then it becomes a question of, does Apple ever get to the point where they would sort of, yeah, have all the technology in-house that they need? And to your point, do they feel like, do they ever want to do that, right? Because it's so expensive to train these frontier models. Does Apple ever want to, want to play in that world. I think that's the open question. I mean, I think there's a lot of

variables in there, including like with their own, you know, silicon and chip expertise. Like, if they felt like they could build, you know, specific chips where they could do it much cheaper and all of that, like maybe there's a world in which they do that. And obviously, they're working on their own models. But again, do they want to play in the frontier? And certainly not right now. And I think it's an open-ended question of if they ever do. But that's not a very Apple historical Apple stance, right? because that's him cook doctrine. They want to own and operate everything, you know, that they want to be in control of. And this dates before a cook to the jobs days when he felt like Apple was getting, you know,

screwed over by Adobe and whatnot. And so anyway, where does Apple go from here? And where do they? And I think the answer is like the honest sort of boring, funny in ways answer is like they're just sitting back and waiting and seeing where it plays out. Right. And right now they feel, I think they feel pretty good that they didn't spend. They haven't spent hundreds of billions of dollars, and they're going to be out there in the world with Siri AI. Well, let me put it this way.

If you could spend $1 billion and get a product that works this way, and again, caveats, we haven't fully seen it rolled out to the way that it's supposed to. So TBD, but it looks like it's on that track. If you could spend $1 billion on this product, doesn't it make you question the $700 billion that are going to be spent on this technology? this year? Yeah, I mean, it's almost like the Dr. Evil meme, right? It's like you're spending $1 billion and it's like everyone else would laugh at you. You're spending a billion, everyone else is spending hundreds of billions of dollars. What do you mean $1 billion? And that's all they needed to

spend to get, you know, a year to get to get basically Google's Gemini technology in the door. And again, we'll see how it plays out from here. But yeah, you would definitely take that bet. And I think, you know, in some ways it sounds like Microsoft is is sort of, you know, backing into some of this as well, like with their most recent talking points of Satchi and Adela just talking about like, yeah, you know, they want to be sort of more agnostic and they'll have, they have their frontier models, but they're almost downplaying them even before they're fully up to snuff, where it's just like, I think, you know, everyone's starting to recognize that there might be a world in which, like, you don't want to have the frontier models because you're in charge of

the cost of that and the returns are going to be more and more diminishing over time. Now, we famously have talked about this for multiple years at this point, and that hasn't really been the case is the frontier is still the frontier, but it's ultimately going to come to into what we kicked off talking about. Like, what are the use cases of these things? And, you know, what, what used to people get out of them? And, and what are they willing to pay for them? You know, as cost is becoming more and more of an issue. Yeah, I don't want to kick a horse wallets down, so to speak, but I personally cannot understand. I've said this on the Friday show. I can't get what Microsoft is doing right now.

They have gone, I've never seen a company go from confident and leading to like behind and flailing in what was it, two years. Like Satya Nadella was making Sundar dance like five minutes ago. And now Satya Dendella is like, what is he doing? It's an interesting,

you know, foil against what, um, uh, basically Apple is done, right? Because Apple has, basically sat back and everyone ragged on them that they were going to they were falling behind an

AI that they weren't there Microsoft was the opposite right they were the geniuses they had the open AI relationship like the one of the best investments of all time and look at them and so you know they were in the sort of opposite position and fast forward those two years and the tables really have turned and again Microsoft to me it all it just feels like they've shifted strategy so many times now that I'm not sure that they know exactly what they should be doing like they currently Again, to go to like Sachi Nadell's current talking points, he's writing his own little blog post about these notions. I think what he's doing is basically trying to read the room and, you know, take the pulse of where things are going and saying like, again, enterprises. He sees all the headlines and knows certainly from Azure and everything that Microsoft itself is selling, that companies are wary about the spend getting out of control, right?

And so what does he want to do? If he can get more open source models, including deep seek, which if you remember, like at one point, Microsoft was going to after them for potentially distilling their big investment in Open AI, right? And how that played out. And so now they're talking about, you know, bringing them on board and, you know, that being, that being one of the key models that they offer up, you know, on their services. And again, bringing down costs. And I think that they want to, they want to be the place that's one-stop shop and they'll have their frontier models and they have everything on down to the open source and let users decide, let enterprises decide what they want to pay for.

And if you believe that, you know, also that things are going to start to move more on device. And obviously they have their current new strategy with NVIDO with the new PCs that they're building, these newfangled second go-round of AI PCs, like are things going to move more towards the edge? And also for cost reasons, but also for speed reasons and whatnot. And so, again, I'm with you. I think that their strategy is still being formed. I think, again, because I think that they've shifted it, many times right now, but that would be my read of what they're trying to do.

Yeah. I mean, one thing I've learned is this stuff moves so fast, don't crown anyone and don't bury anyone pretty much. Yeah. You could shift again quickly. Okay. Before we go, I want to talk about the World Cup. Like we previewed at the beginning of our conversation.

So I have, you know, coming off the summit, I will admit, I have definitely needed some downtime to unwind. And the way I've done that is I've just watched, you know, know, basically as many of these knockout rounds as I can. And I was watching the Croatia, Portugal match. And of course, the World Cup has all this different technology involved, technology that I didn't even know existed. So, for instance, to see if somebody's off sides, you know, they've pinpointed it now and they'll show it on the screen. So even if they get like a pinky, you know, in front of a defender, they'll call it back. And they have this new technology or this technology. I'll admit, I'm not like, you know,

as up to speed as I should be on it. It's called video-assisted referee, and they're employing it for that type of stuff. So here, let me read this little section from The Honest Broker, and we can discuss it because there is a downside here. So the story is VAR has gone too far. Last night, a thrilling end to the Croatia versus Portugal match was interrupted by the video-assisted referee.

Croatia's seeming miracle goal in the dying seconds of the match was disallowed after a VAR review. A cross into the box bounced around before it poked home in dramatic fashion, offering all of us neutrals a chance for 30 more minutes of this engrossing match. I'll just jump in. What happened was Portugal was up by a goal, and this Croatia goal went in. Okay, so then it would have gone into extra time in like the last second. So it would have gone into maybe 30 more minutes of extra time.

Instead, VAR concluded that the ball on its way into the box had slightly touched the hair of a Croatian defender, putting his teammate, as you can see in the top box above, I guess. I guess they added the illustration offside. The touch was so light that the on-field refs did not spot it, and even with high-deaf 4K-75-inch TV in my DC Airbnb, I could not detect any change in the trajectory of the ball. The evidence for the touch came from a sensor in the ball, with the evidence of the grays in the graph inset in the image above.

Technically, to be correct, sure, good for the game, I think not. So basically what happened here is they called the... goal score offside because the ball wasn't redirected by one of his teammates, wasn't hit by one of its teammates, grazed the hair of the teammate in a way that the naked eye couldn't see, and the way that you can't even see on the video review, you can only see by a sensor in the ball that says it apparently touched this defender's hair. The goal was called back, Croatia lost the match, and the World Cup was robbed of what would have been a classic finish. And so what I think this tells us is that, you know, we have become in our society so reliant on AI, automation,

and the robots that even when it's bad for us, we can't help ourselves. And we still want to employ it. And we still want to trust it. And it's time for a reckoning where we say, maybe we've gone too far. So I appreciate the technology angle you've got here for this. And it is good. I mean, because it's not just, it's not just soccer slash football, as they call it here. in the UK. It is, you know, baseball has this issue. Tennis has it to some degree now. You know, all of all of the different things are getting sort of automated away. But this specifically, so VAR has also been, you know, live and controversial in the Premier League here in the UK for a while. And yeah, like people get annoyed because like it just takes

some of the majesty out of the game, right? Because it take, it basically takes it up to the cloud and, you know, looking at where, you know, little things that the human eye cannot detect. And so I do wonder, like, hearing you talk through it, it's like maybe there should be a way, you know, like within a World Cup, within like, say, the final, I don't know, five minutes, maybe two, maybe as little as two minutes where it's like, VAR is off, right? And it's just basically going down to the, to the refs having to do their jobs and sort of everyone in the stadium, you know, like the vibe, like sort of, you know, playing out. like, do we believe that this was actually a penalty? I'm sure they've debated that and things

like that before. Baseball has some degree of that debate always, like how much balls and strikes should be, you know, relied upon the umpires versus the systems, the automated systems. And, you know, baseball had the issue because like games were just getting way too long. So they needed to like come up with the pitch clock and ways to basically speed things up. But that's not the situation here. This is, yeah, just using our the over leveraging technology and taking, again, the majesty, some of the fun, and just the competitiveness out of the game because it's, it's, you know, it's, it's something that they can pick up because it's basically the, the old, like, we, we can do it. We have the technology. And yes, we have the technology. Yes, we can do it,

but should we do it. Jeff Goldblum, you know, Jurassic Park style, like, they didn't stop to think if they should. I keep mixing all my pop culture analogies here, but you get it. Like, it is one of those things, like, we didn't stop to think if we should because it's just like, at the end of the day, this is entertainment, right? Like, and is it more entertainment? to have a game tied and, you know, going into extra time and, you know, and going into potential shootout versus having it called on like this depressingly, you know, again, human-eyed not being able to seem technical issue. And I think, you know, obviously I'm leading it in a direction that I would answer there,

but, and you would too, of course, as well. But they feel like it's probably too slippery of a slope to only do it, like to have these, you know, rules in place to use it sometimes and not other times. Right. And I just wonder if there's a parallel for us in like, if we think about the business world, where, you know, we have similar options. Do we leave it with the people or do we go to robot judgment? And I mean, everybody right now is racing to deploy this technology. And I think my main takeaway watching this in action is, man, we just can't help ourselves. And, you know, and I guess that's good for the bottom lines of the open AIs in Anthropic and the cloud providers. But it is going to

lead to a pretty interesting world where we're going to over-rely on this stuff. And I guess in business, it's less about your feelings in a way. So maybe it will lead to better businesses. But it is, I think, a sad commentary a little bit on the state of humanity that we are rushing to hand the things that we do over to the robots at such a swift pace. I'll give you one other example that's been in the news lately of in law where it's the second order to affect stuff. It's like, yeah, you want obviously you want like the law to work and everyone to be able to take advantage of it. But so many courts now are getting flooded with, you know, lawsuits, basically, that they weren't previously because of little technicalities that these services are able to offer up.

as like, well, can you technically sue for that? Yeah, you can. Here's what you would say. And here's the way to file your lawsuit. And it's like, again, would you begrudge anyone who's filing a lawsuit that they previously wouldn't have because they didn't realize it? No. But there's second order effects of it, right?

And it's going to bog down the court system. And maybe the court can't hear like a very important case, you know, in some other angle that they would have heard it previously because they can't scale as they have. So again, to your point, like, yeah, well, this is sort of an example in entertainment and sports that seems silly. Like it does, all of these things have second order effects that are playing out in the real world in real time that we're just seeing now. So I guess my solution for football or soccer, as our listeners may call it, is I think they should revert the offside rule back to where instead of if any part of your body was in front of the defender,

your offside, I think as long as you have some part of your body behind the defender, you should be good. And that would sort of make, if you're going to do VAR, you can live with that as opposed to the other way around, because that feels more fair in terms of like what's left for humanity. And I think they have to end this sensor on the ball that can feel the hair grays. I don't think hair should count as a body part. That's just me. Sorry to all the, Maybe that's, I'm saying this from, you know, biased from my experience with follicles, but that's sort of my perspective here. I like this.

We're going from, is AI alive to is hair alive? Is the new debate? Technically. Well, I don't know. You don't feel any pain when they cut it off. True. It's true. At least not physical, maybe emotional.

Who knows? Okay. Why don't we leave it there? We got the U.S., Norway, and England still in it. So I'm very conflicted, but I'm excited. I'm excited to watch these. Love it. Great games. Great speaking with you as always. MG. Folks, go find MGs writing at spyglass.org. Sign up for the newsletter. Become a member. You won't regret it. So M.G., thanks again. Great to have you on as always. Thanks, hello.

Hi, everybody. Thank you so much for listening and watching. And we'll see you next time on Big Technology Podcast.