Word. It cannot be emphasized enough that we're still in heavy AI hypesummer. That doesn't mean there's no value to be had in the new tech, but nobody seems to be thinking hard enough about what it actually means for their project beyond marketing... what place are these things going to have when the market finally breaks under the oversaturation? And they all want to go with closed, gross, and heavily logged OpenAI nonsense.I am extremely content with Apple taking their time incorporating generative AI and LLMs into the ecosystem.
The competition seems to be throwing things at the wall to see what sticks. I’d much rather have a polished, highly-functional, useful generative AI implementation than a half-baked one.
I’m particularly concerned about the data and privacy implications of generative AI and LLMs. As the article alludes to, I believe this is the primary reason Apple is taking much longer than Google, Microsoft, OpenAI, etc. to roll out the headlining AI features.
In theory, the 15 Pro/Max is already capable of comfortably running some more lightweight LLMs on-device (7B@4bpw => 3.5GB RAM, more than enough for a Siri-supercharging integration) so Apple has ground to stand on if they wanted to go into the space cloud-free.