this post was submitted on 09 Nov 2023
1 points (100.0% liked)
Apple
93 readers
2 users here now
A place for Apple news, rumors, and discussions.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
The most Apple-way I see it being implemented is using local processing, which requires non existing (yet) technology. Mobile chips aren't capable enough to do serious AI work, if latest Google Pixel AI-features is any indication (they are using cloud compute).
The Snapdragon 8 Gen 3 has already shown to be capable of this.
Well I don't know about phone chips but using the metal api the cores in the m1 and up chips are fast enough to run 7B models nearly as fast as gpt4 runs.