[well_ackshually]

Assuming the rate of progress on AI stays the same:

1/ No, you don't get Opus 4.6 level on devices with 12Gb of RAM, 7B quantised models just don't get that good. Still quite good mind you, and I believe that the biggest advance to come from mobile AI would be apps providing tools and the device providing a discovery service (see Android's AppFunctions, if it was ever documented well): output quality doesn't matter on device, really efficient and good tool calling is a game changer.

2/ Opus 4.6 is now Opus 4.6+5years and has new capabilities that make people want to keep sending everything to someone else's cloud server instead of burning their battery life

[tgrowazay]

I think the claim is that in 5 years an iPhone will have enough ultra-fast RAM to run 300B-1T models on-device.

[well_ackshually]

I'll eat my left nut if Tim "I'd rather die than give good amounts of RAM" Cook bumps the top end iPhone's RAM any higher than 16GB by 2035, especially with the current shortages. They already use relatively cheap LPDDR5X-9600 RAM in there, and are being slowly bumped off order lists on high end fabs to make room for AI hardware. Notwithstanding the fact that there's no hardware improvement in the upcoming years that either makes RAM ultra fast, or with higher capacities easily.

A claim like that is at best naive, and in any case ridiculous

[bitpush]

It isnt speed you want. It is storage. Faster CPU doesnt mean you can store a TB model. It needs raw storage, which famously is through the roof.

So unless iPhone 20 Pro Max has 100GB of unifieid memory all of this is just pipe-dream. I mean, it wont even have 32GB of unified memory.