Apple Siri AI Gets Major Upgrade: Google Partnership Powers Revolutionary Voice Assistant

Craig: so we've talked before how far siri and apple is behind in the ai world. so today, we're diving into what could finally make siri smart. apple has just struck a massive 1,000,000,000 per year deal with its biggest rival, google, to completely overhaul its voice assistant. this is a huge shift in strategy and could permanently change how you use your iphone. apple is paying google's 1,000,000,000 annually to use its advanced gemini ai to power siri's most complex functions. this is a massive deal because it's apple publicly admitting its own ai has fallen behind. it means siri is about to get a revolutionary upgrade in intelligence, finally catching up to the competition and handling multi step complex tasks. what do you think about this, john?

Jon: i am happy about this. this just goes to prove that apple is no longer an innovative company, and they haven't been for ages. if you look at their phones, like, their their apple like, nothing is innovative that comes out of there, and this just goes to show that they cannot innovate anymore. and i'm happy about that because, hopefully, like, hopefully, this changes things with them and they become innovative again. but, yeah. so, the thing with this is google's custom gemini model, it has 1,200,000,000,000 parameters. and you compare that to apple's current model, and that has a 150,000,000,000 parameters. so, think of parameters as, like, the brain cells of ai. so more parameters generally means smarter responses. so this is like upgrading from a bicycle to a ferrari. okay? so is, this is this is interesting. i'm excited about this. i'm glad that siri is finally getting an upgrade. i'm surprised. i'm surprised, but i'm also not surprised that they went with google gemini over, anthropic or openai. but think? do you think are you surprised by their move as far as selecting gemini for this, craig?

Craig: i'm surprised, but i assume, and i could be wrong, that they had a number of conversations with different competitors for different reasons. i assume they looked at the buy option for some of these and just thought it, you know, it wasn't gonna happen. even though some of them are only, you know, valued at $4,000,000,000, which seems like a deal based on their $1,000,000,000 a year, situation here. but i guess the $1,000,000,000 a year situation includes power, that google supplying all the gpu, cpu, data center power to power all of this.

Jon: yeah. and and, yeah, again, they apple could have easily bought one of these, but maybe they just didn't wanna sell. i honestly thought that they were gonna go with something like perplexity because i think that aligns more with them.

Craig: yeah. i figured the price was just too high. whoever was just like, yeah. we're not doing it unless it's, you know $100,000,000,000.

Jon: yeah. so, it's it's funny, though. so the thing is just to put this into perspective, that $1,000,000,000 annual payment, guess how much it represents of apple's revenue? point nothing. 6%. not even a percent of the revenue, but it represents 1.1% of google's, which creates an asymmetric strategy value. so for google, the deal provides validation for its enterprise ai strategy. it also creates a substantial new revenue stream, and it has minimal marginal cost. right? so for apple, it's insignificant compared to iphone revenue, which is at risk because of the inferior ai experiences. so serious poor performance could theoretically cost apple billions in reduced device sales and ecosystem stickiness. so, they're smart to do this. they should have done this sooner, but, yeah, this is, a smart move for apple for sure, finally.

Craig: and i'd like to clarify, apple's private cloud compute infrastructure will run the entire system. this setup ensures user data stays completely separate from google's own system. so i misspoke. apple's gonna run the software on their own hardware in their own data centers, so they have total control over the data, and it and it doesn't go to the third party google.

Jon: yeah.

Craig: yeah. this looks like, you know, apple seems their, teams are developing a 1,000,000,000,000 parameter model internally. so i guess they're still trying to work on this, and maybe that's why the rent versus buy solution. hey, boss. we're only five years away or three years away. you know? yeah. and somebody decided to do this. i'll believe it when i see it. they just seem so far behind. every time i've used apple ai on, like, my mac, it's just disappointing. yeah. and even though most times it's just chat g t b t or whatever, like, the images it creates is bad. there's like, it's just not good for anything. other people might give me crap for that because they'll use it for this or that. but like, i try it and, like, yeah. forget it. i'm back to my gemini or claude or chat gtp or what have you. and i think that's part of the problem that apple realizes that the number one apps on their phone are other people's ai apps, and no one's using their ai. so no. they really need to change the game. if ai is the future, then they are way behind.

Jon: yeah. yeah. and you know what, though? even as of right now, what do you think would it be accurate to say that google is behind anthropic and openai, or do you think that gemini is on par or ahead?

Craig: i like, today

Jon: because i don't know. i don't know what that answer is. i would say that they're behind, but i don't use gemini, so i wouldn't know.

Craig: yeah. today, i'm just saying that, openai just released 5.1. yeah. i don't have that it is. gemini is faster. i use gemini. i like it better for some things. like, i use it more when i'm preparing for the podcast now than i do for chat ttp, just because i like the output. yeah. and so i would think that it has come out quite along from the embarrassing start it had. it still has issues. it's still heavily biased in lots of areas.

Jon: yeah.

Craig: but that's why we we have i can't remember. i complained to you about it a fair bit. it won't even do this. yeah. it won't even do that. like, really? so and i just jumped to another ai, and it does it no problem. but it is so much better. like, they've cut their image generation is so much faster. the responses are so much faster. openai is slow. right? it might be say, everything else around what you want, but it's slow. if you're using it through the api, slow. it's one of the slowest even, like, even interacting with it, it's one of the slowest out there. so it may get you good results, but sometimes i'm getting more and more impatient. i can generate five, six good looking images in google gemini before openai is done producing one.

Jon: yeah.

Craig: like, okay. you know? so what do you do in that situation? so besides that, besides the performance, i like openai result in some things that, like, gemini results. and i think they are leapfrogging each other now, and google has really come a long way and has taken a lot of the stuff out that was causing some problems before. so, like, the stuff they pulled from reddit, they're starting to find. i've noticed when i pointed, posted stuff about google, on or google ai on the internet, like, what the heck is this? they have responded, and they have said, hey. can you send me this and send me that? because they wanna know why this is happening. because it's you know? like, when i posted that, welcome to the jungle okay. in the middle of it's it's doing a list for me, and it just decided to print out welcome to the jungle. we've got fun and games or something like that in the middle of a report it was doing.

Jon: that's crazy.

Craig: and i'm like, what the heck is this? so there's nothing there was no anyway, there's some stuff it randomly does, but i don't know if that's any worse or better than chatgit2p. so, i use gemini a fair bit. i mean, ai studio, google, everyone should be using it while it's still free. it's pretty good until it's not when you're building apps and whatever, but that's most live coding platforms. everything's great until it decides to take a left turn, and then things go to crap. when it decides, hey. i'm gonna change react models in the middle of your thing, and all of a sudden you're getting errors like nobody's business or whatever it decides to do. but back to this. like, i think it's going to take this is what it's gonna take for apple or for people to take ai serious on apple products.

Jon: yeah. so

Craig: yeah. and this is what it's gonna take. how long it's gonna take them to come back out with their own, 1,000,000,000,000 plus platform that they're plot they're trying to, build. i don't know.

Jon: yeah.

Craig: and they're using the gemini ultra model, and i'm not quite sure what, what model that is. so i'd have to do a little more digging because there's an ultra image model, but i don't see an ultra model period, like, when i'm using gemini. so that might be something i don't have access to, or that might be a new back end thing. yeah. so we'll have to see. so is apple swallowing its pride to give us a smarter siri all while planning to rip out and replace its rivals tech in just a year? we'll be watching this one very closely, coming, the 2026 to see if they can pull it off. what do you think? is this a brilliant move or a desperate one? let us know your thoughts. is this gonna make, ai on apple products actually usable or not?