Will macs have any more capabilities on Apple Intelligence than iPhone or iPad?

[mactinkerlover]

Title. With the limitation for devices older than the 15 pro clearly most likely being ram, will apple silicon macs with more ram ever get more advanced features? Like what about my M1 Max with 64gb ram.

What do y'all think?

 

[coffeemilktea]

They're certainly more capable of it, but I think the real question: will Apple ever actually add more Apple Intelligence features (and if so, when would that happen?).

I will say though, it looks like Xcode is getting some vastly improved code completion features ("powered by a machine learning model," the release notes say), so that's at least one example of something on the Mac getting AI features that the iPhone and iPad aren't getting.

 

[mactinkerlover]

They're certainly more capable of it, but I think the real question: will Apple ever actually add more Apple Intelligence features (and if so, when would that happen?).

I will say though, it looks like Xcode is getting some vastly improved code completion features ("powered by a machine learning model," the release notes say), so that's at least one example of something on the Mac getting AI features that the iPhone and iPad aren't getting.

Question, do those xcode features require beyond M1 for the more powerful neural engine?

 

[coffeemilktea]

Question, do those xcode features require beyond M1 for the more powerful neural engine?

No, just any Apple Silicon chip and at least 16 GB of RAM.

 

[startergo]

With the limitation for devices older than the 15 pro clearly most likely being ram,

How did you come up to this conclusion?

 

[NT1440]

Title. With the limitation for devices older than the 15 pro clearly most likely being ram, will apple silicon macs with more ram ever get more advanced features? Like what about my M1 Max with 64gb ram.

What do y'all think?

I’m curious as to what exactly you’re envisioning?

 

[senttoschool]

I’m curious as to what exactly you’re envisioning?

The simplest is much larger local models. LLM intelligence scales with the number of parameters.

 

[dmccloud]

Question, do those xcode features require beyond M1 for the more powerful neural engine?

Based on Apple setting the floor at an M1 with 16GB RAM, there's plenty of head room to run larger LLMs as system RAM increases. It's not a question of more power, but the number of variables which can be loaded into RAM at once. The larger the model, the higher the RAM requirements go out of necessity.