The point here is to replace iPhone 15 which doesn't support Apple Intelligence without raising the lowest price point for an iPhone. Apple has made a bit of a blunder by selling hardware with such limited memory for so long, which gives them a large install base that doesn't support gen-ai tech. Of course, the urgency of upgrading depends on how bullish you are on gen-ai.
I'm not sure if running any local model on smartphones make sense right now for most of the people. I played with many tiny llm models on ollama and only 14b models can feel to be usable - otherwise why not just run any app that do inference in cloud and run SOTA models? Even o4-mini feels more usable than those tiny models on smartphones - apple would have to ship iphones with 24GB RAM but inference would still be s…
It probably makes sense for simpler tasks like summarizing text messages/email that you might not necessarily want to send off to a third party that has a "move fast and break things" approach to data privacy