|
|
|
|
|
|
by lucaslazarus
207 days ago
|
|
|
Not to mention you would need an order of magnitude improvement in on-device inference speed to make this feasible at current smartphone costs. Or they could offload it and sell an insecure overpriced-subscription laggy texting device that bricks when you don’t have cell service… |
|
|
But it is clearly the direction. Apple will try to stave off this move by turning iOS more into an LLM as well.