How On-Device Intelligence Is Redefining AI’s Future | Utilizing AI Ep. 4

Most of the AI we've been using to date has been processed in the cloud, but this is starting to change. Inferencing work for applications is increasingly being handled on device thanks to advances in processor technology. And the connection from device to cloud is increasingly transparent as Apple and Google and others roll out private cloud compute to seamlessly direct AI tasks to either join Olivier Blanchard, Jon Swartz, and me Stephen Foskett as we discuss connected AI applications.

Welcome to utilizing ai, the podcast focused on practical applications of artificial intelligence from the Futurum group. Each episode brings together diverse perspectives to explore news and use cases of the ways in which AI is transforming enterprise IT and the industries it serves. I'm your host, Stephen Foskett, president of the Tech Field Day Business Unit here at the Futurum Group.

Before we dive into today's discussion, let's meet who's on the panel today. I'm Olivier Blanchard. I am a research director at the Futurum Group, and I focus on AI devices, uh, which is basically everything from the little wearables all the way to vehicles.

Hey, I'm John Sports. I'm senior content writer at Techstrong. I write primarily about ai, although I do dabble into other areas, but primarily I write about AI because everything is ai, it seems every day.

It sure does seem like that, that way. Uh, this is, again, I'm Steven FoST. Uh, uh, I run the AI Field Day event for futurum as well as contribute to podcasts like this.

And, and this is one of the areas of focus for me, uh, both here on our utilizing tech podcast and as well as the Tech Peel Day podcast. This week, uh, you know, we've seen some, uh, interesting announcements. We've seen some, uh, sort of concerns over, uh, AI spending and data center buildouts, but we've also seen some, uh, announcements, uh, for example, from Google, uh, that they're rolling out their private cloud compute for AI workloads on personal devices, uh, and for end users.

Uh, of course Apple has talked about that as well. And this leads me to sort of a, a, a core, um, issue when it comes to AI that we've spoken about previously, but I think we should dive in here. And that is this question of sort of the, the widespread notion that AI is this all consuming monster that's, you know, consuming, you know, gigawatt data centers and all this, and yet we're all using it every day on devices that are a lot smaller.

There's a big difference between Twain training, uh, inference and just sort of AI as a, as an application component. So, Olivia, I wanna throw it to you first. Help us understand the ways that AI is used in the real world by real people and as opposed to what we hear about these big data centers.

Yeah. Um, so there's, there's sort of like a, you, you have to imagine that you're on a timeline, which we are. There is, what we've been seeing in the last two years, there is what we're seeing today, and there is what we'll be seeing in two years and two years after that, and two years after that.

And so the, the first things we understand is that initially most of the AI that we've been using has been in the cloud, right? We're accessing it through a browser, through an internet connection. And even though the inputs and the, the sort of, like the touch points are the devices that we use, whether it's a phone or a PC or something else, um, really most of the work, the training, uh, of, of the AI models and the inference, which is sort of like the, the processing of your query and then the response, all those things have been happening in the cloud.

Um, more and more though, what we're seeing are, are, are two different things. One is we're starting to see a lot more of that inference work moving to devices. So for, for instance, instant translation is the type of thing that could happen in the cloud, but works better on device because you don't have as much lag between the input and the output.

Um, and we're getting some, some much better software and much better hardware processors mainly being, uh, sort of like the engines of these new AI capable devices. And so what we're seeing is a lot of the inference work, a lot of that, that sort of compute heavy work that was a year ago still being done in the cloud can now be moved to devices. Um, so that's, that's the first thing.

The, the other thing is that the, the hardware, so the processors themselves and the models are becoming more efficient, which means that even in the cloud, even if you're running a hundred percent of your AI workloads in the cloud, which you don't have to anymore, what would still happen is over time, fewer resources would be needed to accomplish the same goals. So you don't need as much processing power, you don't need as much energy. Uh, the models can be run on much smaller chips or much better system, or smaller systems rather.

Um, and what you see is sort of, you know, as AI expands the needs for the infrastructure around that AI becoming more efficient sort of shrink, which means that you don't necessarily need the type of massive, uh, power hungry data center, uh, infrastructure 10 years from now than, than most of the industry thinks they need to build today. If that makes any sense. So we're, we're, we're moving towards better efficiency, um, and also a more distributed sort of federated AI processing ecosystem where some of it is done in the cloud, some of it is done more locally, sort of like on the edge of a network, and a lot of it moves towards, uh, actually being done on the device itself.

And so that, that sort of takes a load away from, uh, from these big power hungry data centers that we're seeing today. So, Steven, can I, so, so one of the things, Olivia, I totally agree with what Olivier said, but I think what might be happening right now, and I think it's a lot to do with Silicon Valley, I think that Silicon Valley perhaps is inadvertently feeding this, this narrative of energy and environmental anxiety. And the reason why I say that is that, uh, uh, Sundar Bahai, who's the, uh, CEO of Google, gave an interview to the BBC, and he basically just warned that not only other tech companies but his own is in danger of this, uh, this AI investment bubble bursting.

And, um, he alludes to spending, but for the most part, he's talking about elements of irrationality in the current market and the buildup to, to, to get ahead. And, and, and I think this is compounded by the JP Morgan Chase study that founded more than $5 trillion. That's trillion with a t we'll be invested in data centers and AI infrastructure worldwide over the next five years.

So for instance, the, the figure next year would be about 700 billion. 4 trillion. So this, in a sense, creates a lot of headlines, creates a lot of buzz, creates a lot of talk, which leads to folks like not just, uh, pka, but Jamie Diamond warning that this z investment would ultimately pay off.

But some of the money being poured into the industry would quote, probably be lost. That's what Diamond said. So it, it's this kind of, this monster they've created this Frankenstein monster and Silicon Valley has created in their rush and their exuberance to tell you how, how much infrastructure they're building out to achieve the training or whatever they want to achieve.

So I think they're probably, you know, in part to blame for creating this kind of, this kind of fear narrative. But, you know, I think it's important to think that, you know, at the same time, Google is actively rolling out AI applications to consumers. Yes.

And consumers are paying for those. So, back, you know, that JP Morgan report that you mentioned, um, they talked about, uh, the fact that they're looking for $650 billion of revenue coming from ai, right. Uh, in order to make this CapEx spend.

Makes sense. Um, I think that it's really interesting that JP Morgan was not necessarily negative on ai. They're just basically stating fact like, you know, Hey, we spent this money, where's the revenue?

And, and furthermore, they specifically call out that that equates to about $35 for every iPhone user, or $185 for every net for Netflix subscribers. So they're trying to give us a sense of the scale that's gonna be required in terms of spending to offset this investment. And when you think about it as $35 per month per iPhone user, you start saying, wait a second, that might actually be achievable if it is end users, end user applications that are kind of bearing the brunt, because of course there are way many Android, you know, Google users, uh, Google Ecosystem users.

I could see a future in which, uh, Google is absolutely making revenue from end users to all offset the incredible hardware costs they're making. I'm quite so sure about some of the other companies in the industry, but I think Google's got a pretty good plan. Yeah, Google's a good example.

Google. Yeah, Google's well, well positioned. I totally agree on that.

I, I'm also thinking about like an OpenAI, and they're tied into like more than a trillion dollars in these deals around infrastructure. And you, when you look at their revenue, it's one 1000 of what they're spending long term. So I think that's, that's kind of the, I think that kind of gets people's attention.

Or some folks dwell on it, like our friend Michael Berry, which is another story in and of itself. So, um, our, our big shorts, uh, hedge fund investor who warned about this, uh, AI boom and reaching a, a, a, the crater or reaching the edge, he's since kind of disappeared from the scene. But I, um, just going back to what you said about JP Morgan, I think it was a very cogent, like, sober look at what's happening, very reasoned approach, and it is achievable, but I think it's also something that is, um, weighing on people's minds Well, and to the point that you just made about, you know, so it Barry bet against, um, both OpenAI and Nvidia, if I recall correctly, we also just got news that Peter Teel, um, has, has, um, liquidated his Nvidia investments after many years, but I think it's interesting that he reinvested that money in Apple and Microsoft, who are both part of this com part of this world.

But like we said about Google as well, those are companies that have a plan to make revenue off of this technology. Uh, and I think OpenAI does as well. They're building a platform.

You know, you look at Apple and Google and their private cloud compute, they're built not on Nvidia GPUs, but as, uh, Olivier was talking about based on, you know, based basically inferencing hardware that is much lower, um, you know, power consumption costs, et cetera. And of course, apple and Google are also deeply interested in and as is Microsoft in using mobile devices to do on-device processing. Um, I think this shows us Olivier, is that the path that you're seeing?

Yeah, so it's, it's interesting. I wrote a, a, a piece a few months ago about sort of like the binary model of ai, right? Which was kind of like cloud and device.

And actually I was wrong. I I made a mistake. Well, I was right at the time, but it was a failure of imagination on my part.

Um, because now we're starting to see more of a stack, right? More of a tiered approach to how AI is being sort of, um, ecosystems, uh, let's just Verbiage unified binary model. Yeah.

Well, by, yeah, it's, it's, it's, it's more so what we're seeing with the, with the Google private AI compute solution or offering that they just announced, and Apple's private cloud compute, PCC, which is very similar, is that you have sort of like a, a middle layer, right? So you have like the cloud, which is the, the open cloud, the open, the open web with, uh, with ai like OpenAI and other solutions like those Gemini. You have the AI that lives on device that actually works on your pc in your car, on your, uh, on your phone, in your smart glasses.

And then you have this middle layer, which is this private cloud, right? Um, which in a way sort of bridges the gap between this sort of very distant laggy, uh, expensive, um, uh, very high performance, but sort of like low user experience of, of cloud. That's also not particularly secure because you're sending your queries up somewhere.

It could be intercepted, it can be, uh, captured, it can be sort of mind for information. Um, but then a lot of the, the processing still can't, that you want from, from these AI solutions can't necessarily be, um, be handled by most devices. Even if you have the latest device, it might be able to do fairly well compared to devices two, three years down the road.

Uh, and if you have a device that's not really designed for ai, then you're kind of, you know, left out in the cold, you have to use cloud. So what this does is it brings a lot of that compute a lot closer and in a much more secure environment, uh, which is, I think is, is the most important thing. So, um, the, the, the sort of elements of Apple and Google's private AI cloud offerings that, that really struck me as as being important are privacy.

Uh, right? They're, they're, uh, um, uh, cryptographically secured it's hardware to say, um, uh, a lot of it is, is post quantum encryption. Uh, apple staff can't access any of your queries, for instance, they can't access any of your history.

Um, there's much more security around it, uh, because they're running their own, like Apple is using its own silicon for this, so they have control over, uh, over the hardware. Um, there's better functionality. Uh, it's, it, I, I think what we're going to see is this sort of multi-tiered system or ecosystem of AI processing with some of it going to the cloud.

Some of it's going to sort of private cloud, some of it going to the edge of the network, some of it just staying on device or even federating between devices. So you have smart glasses and a phone and a ring and a smartwatch. Um, all of these have processing power that can be used together, uh, to, to, uh, run workloads locally, right?

And they're all constantly pulling data from yourself, from your environments. Um, and so I think that's what we're moving into this much more organic sort of ambiance, uh, type of AI processing that again, won't require quite as much, uh, input from data centers. And to, uh, Peter Thiel's point, and I don't generally agree with Peter Thiel on, on a lot of things, uh, but what we're seeing this withdrawal from Nvidia, to me, signals two things.

One, what we just talked about, the fact that, um, AI is moving into other areas that don't necessarily require Nvidia chips, uh, and giant Blackwell processors to, uh, to, to run training or inference, uh, and the other that essentially, um, and, and because Nvidia doesn't necessarily, or because we don't need those giant processors, the need for NVIDIA processors moving forward, as we shift from training to inference and inference more and more moving to the edge, Nvidia sort of loses its momentum. There's, there's not, I don't see a demand for NVIDIA chips. Um, the, the, the demand that we've seen for Nvidia chips rather in the last couple of years, really sort of progressing into 2028 to 2029 or 2030.

And I think that's a threat for Nvidia. And I think possibly I'm interpreting these sort of pull aways from, um, from the Nvidia stock as, uh, an anticipation of that moment when Nvidia chips become less of a necessity to run ai. Forgive me for this.

I might be going down a rabbit hole, but what you were just talking about, Olivier and I, and, and also Steven, um, in terms of Apple, I'm just curious, and this is an open-ended question, and perhaps it's a little bit off track, but in terms of Apple and its AI strategy or where it's going, and you've laid it out, laid out where it's, where it's going, it, there's always been a concern out here that Apple is in a sense, kind of a late player, which is what they do in, in most markets. But I think in this market, which moves so quickly, there was a definite fear, and, and that was compounded by folks leaving Apple. There's a rampant rumors that Tim Cook's going to retire in January.

I, I'm just wondering, is that an overblown argument that ai, that Apple is not supremely positioned in terms of ai? No. It Okay, if you, I, sorry, I gotta jump in on that one.

Uh, I, I feel passionate about this. Um, good. I feel I love Apple's AI strategy, and I know not everybody does.

And hopefully, Olivia, maybe you'll disagree with me on that would be great. Um, I love Apple's AI strategy because Apple so far has not invested big in AI model training, building, you know, foundational models. They have not invested big in, uh, GPU hardware that we know, uh, they're not listed.

Um, you know, none of the publicly available sources suggest, in fact, um, most of the sources suggest that Apple is leveraging companies like, um, Google and perhaps, uh, I believe, was it Microsoft or, or, uh, may, I think it was Microsoft anyway, that they were leveraging, um, as a service, um, GPUs to do their AI work. Um, they were the first to basically say, AI is a feature, not an application. Uh, that's kind of their strategy.

Uh, apple is what they're calling Apple Intelligence is basically just a collection of random features all over the place. And some of them are little, and some of them are big, and, uh, frankly, none of them, all that impressive. But ultimately what they're focused on is privacy and serving the needs, the sort of daily needs of end users, which to me is just such a supremely practical approach to ai.

I love it. And also, as you know, as we've said here, um, one of the things that they announced actually almost two years ago now, was this private cloud that would basically extend the capabilities of your device. If a query was too big for your device to handle, it would run it in the cloud.

Um, and that cloud is literally the same device. In fact, from what I'm hearing, Apple's private cloud so far is just a bunch of Mac Minis. Uh, now they've supposedly just this quarter, uh, built their first, uh, actual servers with Apple, silicon hardware.

At least that's what the reports say. Um, but until now, I guess it's kind of been like MacStadium or something up there. You just got a whole bunch of Mac minis up there that are running stuff.

I'd love this approach because it's not, you know, let's have a chat bot run the world. Let's build an artificial general intelligence. It's let's figure out how we can use this in ways that benefit users.

But so far, I don't think anyone has given Apple a dime for any of this stuff, apart from not leaving the ecosystem, which is maybe not the best strategy. Uh, Olivier, uh, tell me I'm wrong. No, you're not, you're not wrong.

I'm wrong, man. However, We're relying on you, you Caveat, caveat. Um, so, and, and circling back to, to John's original question, I think it's a, it's a mixed bag, right?

I think Apple really did miss, uh, AI initially. They didn't really see it coming. They weren't prepared.

They were focused on other things. And to be fair, I think Apple's missed on a lot of things. Um, they used to be, they used to be a very innovative company, right?

They were first out of the gate Yeah. With features, uh, with, with form factors, with all sorts of things that were really cool and that made technology be exciting, and everybody was following them. And no, no, dig on Tim Cook.

I think he's done wonderful things for the company, and he's, he's an incredible CEO. Um, but I think that, you know, there's definitely a Steve Jobs Apple, and there's a Tim Cook Apple and the Tim Cook. Apple is much more, you know, he was CEO and you could tell it's very, um, it's very efficient.

Uh, it's very thoughtful, but it's not necessarily innovation first. And so what the, The joke supply chain focused, because that's what Tim Cook was all about. You know, I'm gonna make stuff, I'm gonna make it at scale.

I'm gonna have it, you know, quality and, and, yeah. Yeah. Right.

And so he's, it's, it's a very efficient machine. It makes a lot of money. It's, it's extremely successful.

And, and really Apple hasn't really paid a price for coming into the market two, three years behind, uh, you know, Android, for instance, for the phones with features. Um, I thought that it would, but it hasn't caught up yet. And so I don't think that that Apple will ever pay the price for this.

But there's, you know, there's a joke for us Android users when whenever Apple introduces an exciting new feature, um, which is always my real reaction. I don't know if I'm being punked or if Apple really is excited about releasing a feature that I've had for three, four or five years on my Android. Right?

It's like, you guys didn't have this yet. That's, That's fun. That's funny.

You, you say that, Olivier, I was just gonna just interject really quickly. So when I used to, I used to work at USA today, so we would all go, they would, they would, they invite like seven of us to go to these Apple events. And we, this was the era of a lot of tweeting.

So we were, part of our job was to tweet as much as, as right about the event. And almost every response we got from the new feature from Apple was from an Android user saying, now you, you've caught up with this from three years ago. So anyway, Yeah, no, it's true.

This is an Joke. I would refute the whole Apple is a leader thing. Apple is never led.

I mean, I'm sorry, but I, I know at least John is my age, um, eight bit hardware, e era. Um, I remember when the Macintosh was launched saying, oh, come on. Like, we've seen graphical environ.

I mean, what about Gem? Right? What about my Atari St you know, I mean, you know, apple, you know what Apple does well, and it's the same with smartphones.

I mean, honestly, they do deserve credit. Steve Jobs, you know, when he, when he held up that iPhone, it was all screen Yeah. Except for the buttons.

Um, you know, he, I think that that was revolutionary. But for the most part, like, you know, those of us, you know, I remember I was a, I had a, you know, a, a, a creative nomad MP three player, like three years before Apple came out with the iPod, um, maybe two years, but for a long time, you know, I was using MP threes on the go. You know, what Apple does is they integrate these features in a way that's very user friendly.

Yes. They've never been a leader. They've always been a follower.

And, and I think they remain Refine, they refine the technology. They make it so easy to use. They understand the customer better than anyone, I think.

Yes. Which of our users thought we would talk about MP three players and Ataris St. Here on No, no.

All, all, all these things are true, but, um, but, but the, but the fact is that, you know, we would still be using Blackberries were it not for, for the iPhone. Um mm-hmm. It, it, it's the smartwatch, even just the Apple Watch is also another example of a revolutionary product that yes, fine, you know, There, the smartwatch Wasn't Invented there, start watching before they just sucked, But they sucked.

But Apple does it better and Apple does lead, right? But, but, but in recent years, Can I ask a Quick question, though? I haven't really seen that from Apple.

Yeah, sure. So, so getting back to ai, we're getting into it here, getting back to ai, um, with, with Apple. Yeah.

Is it conceivable, do you think it's conceivable, I know they don't do big acquisitions where they rarely do them, but Perplexity and Apple in search, is that something that you could see possibly happening? Or is that, am I totally crazy, Libya? No, I mean, it's, it's possible.

Uh, but I, I want to, I want to kinda like, like get back to the thread, uh, real quick because the answer might be at the end of it. Um, so you were asking me like, did, did Apple Miss? And I think that Apple has had a habit in recent years in, in at least the last two decades of, of coming late to the market, essentially sort of watching what everybody else does, taking the time to figure out what's gonna win, what's gonna miss integrating, uh, uh, features that, that have already been in the ecosystem, and so on.

On one hand, it's, it's not great that it takes them that long to do it. On the other hand, it's sort of on par with Apple's strategy. And I think that when it comes to ai, that's sort of where a Apple was.

They weren't really trying to be on the bleeding edge of what was coming next. They were just sitting and watching and analyzing and figuring out, okay, what do we need to integrate in the next two, three years into our ecosystem? And so, initially Apple did Miss ai, um, and, and I think that it, it rattled them because usually they don't pay a price this time.

They did, and they're, they're well, So, but I think strategy difference, I, but there's a big difference when they, when they waited and then they, then they jump into a market and they were successful. That was under Steve Jobs. So this is Cook's Hunters Cook Watch.

No, still still With Tim Cook. Yeah. Yeah.

No. Um, but I, I, I think initially it rattled them because it, the impact of, of not being on the forefront of this did hurt them. They, they realized that they made a mistake.

This was different from just features. Um, and so they didn't really have a plan. They didn't have infrastructure ready, they didn't have partnerships ready.

They didn't really know what to do with it. And so Apple Intelligence initially felt like something that was cobbled together, together because, uh, apple needed to have an AI story and, and have it really quickly. Now what we're seeing is, is a shift in, in the last year, we've seen Apple adapt and figure out sort of what they wanna do and where they wanna go.

And even though they're not pointing the way and saying, okay, we're gonna do this and it's going to, we're gonna create these unique differentiated Apple only features, what they are doing is, is testing the waters of what the infrastructure build out needs to be, who the, the partners need to be. Is it OpenAI? Is it perplexity?

Is it Google? Who are we working with? Um, the flavor of the moment is Gemini.

Obviously, they've, they've selected Gemini to be part of that Siri or broader Siri adjacent ecosystem. Um, they're sort of following Google's model as well with private cloud compute, uh, and some of the, the infrastructure and integration into the larger ecosystem play. And I see a lot of parallels between what Google's doing with AI from devices all the way to software and what Apple is doing.

The difference is that Apple doesn't have giant data centers everywhere. They're, they're not going to spend tens of billions of dollars building data centers, um, that they know will be obsolete in 10 years. That's not how Apple operates.

And so, to your point, and, and to answer your initial question, um, I think that Apple's AI strategy is a mixed bag of, uh, oops, we missed, we need to adapt and adjust and, and change course and be nimble. Um, but also it's, it's incredibly smart because they're not, uh, following the herd and spending a, a ton of money on, on infrastructure Yeah. That they might regret having later, right?

Yeah. That's, uh, they're gonna be more nimble, much more agile, and they're, they're Apple, so their brand can carry whatever mistakes they make because they can just move their business wherever they want. Yeah.

I, I always think of their, their, uh, their, uh, uh, rival or their, their foil. What, so it'd be, um, meta. So Apple always, they, they had this history of, of comparing themselves to Facebook and Facebook not being a privacy friendly company.

And I think about, about Meta and they've, what they've announced is $600 billion investment in US infrastructure and jobs over the next three years. So it's the total opposite. It'd be interesting to see that kind of, that plays out between the two companies overspending versus not spending at all.

Well, I mean, what I, what I look at, yeah, what I think is interesting is, um, the, the, the cloud companies, right? So you have, um, uh, you have Microsoft, you have Google, and you have AWS and Amazon is, uh, is is also in the hardware business, right? They have, uh, they have Alexa devices, they have ring devices, the speakers, the fire TVs.

So they, they have sort of like this baseline of physical interfaces in the home that I assume will start expanding outside of the home in the next, you know, 12 to 24 months. Uh, where we start seeing Alexa sort of become more portable. Um, so they have the cloud solution, they can do the private cloud thing, and they have the devices Google, same thing.

They have the cloud, they have the private cloud, they have the devices, Microsoft less so, uh, but they do have the operating system. They have the pc, uh, uh, ecosystem sort of on, on, not complete lock, but, but essentially in the enterprise, uh, and in education, they're, they're there, um, alongside Google with, with Chrome and Chromebooks. Um, so what's interesting about these three companies is they really do have those, all of the elements of the stack, even if they're not, um, completely fleshed out or, uh, or, or you don't really have like full parody between, you know, Google and, and, uh, and Amazon.

My worry about meta is that, um, Meta's much more of a sort of cobble together company. They have obviously the, uh, um, the Meta XR play, right? With smart glasses all the way to, to virtual reality.

Uh, and they do have, um, social and, uh, um, gaming or mobile gaming and, uh, and some retail, but Meta, I think is much more exposed with the spends than, you know, Google, Amazon, uh, and Microsoft. And so it, it, I'm, I'm a little bit concerned about where Meta's going with this. Um, I'm also concerned that Mark Zuckerberg a few years ago thought that everything was gonna be about the Metaverse, and we were all gonna be in virtual reality ecosystem that changed their name or environments name among Other things.

Yeah. Right? Yeah.

Um, and now he's latched onto this thing. And, and it, it feels to me to come back to original point that a lot of the spend that we're seeing, first of all, isn't really spent there. There are promises.

It's right, it's a pitch. We're gonna spend X amount of money, right? That has not been spent yet, and that is not being spent yet.

Um, but there's this sort of like weird reflex to say, well, we're gonna, if if our competitor, someone we perceive as a competitor is spending $10 billion, we're gonna spend $15 billion. And then it's, it's just kind of like this, this one upping of everybody else just to project that their company is going to be the leader in the AI space by virtue of the fact that you're going to outspend everybody else. Um, and I think that, and it Goes beyond AI Sense that word in a bubble.

Yeah. Yeah. But I, and, and also, I mean, apple has made its quote unquote promise, although I don't think they're going to spend like an inordinate amount of money on, on infrastructure in the US to, that was more of a political play.

But I think you're right. I think these are all promises that we're gonna, we're gonna get like one 10th of, or one 50th of what they say they're gonna spend. That's right.

Well, and, and that, that is why I think people are getting nervous about Nvidia, Right? And so, so to to yeah. To look to kind of bring this all back down.

I'm sorry to interrupt, but we're, we're almost at time here to kind of bring this back to the start. Um, you know, you make a very interesting point. Olivia and I, I actually amazingly a hundred percent agree with you even on the Apple part, that Apple kind of missed the boat on AI and was, you know, playing catch up and so on.

Um, but Apple Cogently stated, I think what the future of ai, at least in personal, is gonna look like, this extension from, you know, your personal devices up to the cloud, privacy control of data, um, you know, Google is absolutely right there. Microsoft is right there. Um, you know, even Amazon as you, as you point out, I always forget the fact that the other Alexa ecosystem is incredibly powerful.

And I also agree that, you know, meta is the one that's really standing on the sidelines here. OpenAI is standing on the sidelines. OpenAI is trying to build an infrastructure, basically the new windows for ai.

But ultimately, you know, if I had to bet which strategy is gonna work, I would bet on companies like Apple, Google, you know, Amazon, Microsoft, and not on companies. And, and it seems like that's where the, uh, where the rest of these are, are headed, at least from the consumer point standpoint. Now, OpenAI may end up being, you know, dominating, uh, business use of AI or something like that for all we know.

But, um, but I think that the, you know, companies not named Apple are also gonna be fighting for the business market for AI applications as well. Um, but we do have to, uh, wrap up. Uh, I think that we've got a pretty good stance here.

Before we go, one last, you, you wanna make one more point, uh, John or Olivier? No, no. I, I, I found this really, uh, fun and interesting.

Thanks for your, thanks for, uh, doing this. Well, I really appreciate that. Yeah, no, I, I, yeah, I think, I think I've made all the points I needed to make today.

Um, but if I'm, if I'm allowed back, I'll, I'm sure there'll be more. Absolutely. Well, thank you both so much for joining us.

Uh, and thank you listeners for joining us for this, uh, episode of utilizing ai. This is really what we're focused on here. We're trying to figure out where this gets real, where it gets productive, where it gets profitable for, for enterprise, for, uh, the providers of these services.

And, um, what doesn't make sense, because we're hearing an awful lot as well about what's going on in the industry. And, and, and a lot of us are scratching our heads over this. Uh, before we go, uh, Olivier and John, um, where can we continue the conversation with you, and where can we find your reporting on ai?

Uh, well, one good place to start is the, uh, future website where most of my insights are published. Uh, another good place is X, formerly Twitter, uh, OA Blanchard. So it's really, I'm pretty easy to find.

I'm not the IMF economist, I'm the other Olivier Blanchard, and occasionally, twice a year, maybe I'll be on LinkedIn, but I know that a lot of my contents, uh, lives there as well, even if I don't. So, um, those are three good places to find me. Well, I'm just the opposite of you, Olivier.

I gave up on X, although I still use it. I don't, I don't post on it anymore. But I do use it as a news source.

I, I am, uh, I spend an inordinate amount of time on LinkedIn reaching out to people or interacting with people and posting stories. But for the most part, I'm on, uh, Textron, uh, Textron group and, uh, throw afu 'em a little bit. But, uh, you know, we're gonna, we're gonna get closer and closer and meld the both sides together, so you'll see it everywhere.

Yep. And as for me, uh, same, same thing. Uh, you know, I'm, I'm on LinkedIn a lot.

Um, I'm on Textron's, uh, media sites a lot, including Textron Gang most every Tuesday. Uh, I know that, John, you're there a little bit more often than me, uh, just a couple times today. ai.

So thanks for listening to this episode of utilizing ai. If you enjoyed the discussion, you can subscribe, uh, on either on YouTube if you like to see what we look like or in your favorite podcast application. If you don't, uh, please do give us a comment, a rating, a review, uh, that always does help all podcasts.

I know everybody says that, but it really does help. This podcast is brought to you by the analysts and experts from the Futurum Group, where insights meet ai. ai, the utilizing AI YouTube channel.

Or, uh, you can also find us in the Textron TV app. Thanks for listening, and we will catch you next Wednesday.