On-Device AI Models Might Be the Next Reason to Upgrade Your iPhone

4 hours ago

The iPhone 17 runs a 3 billion parameter language model on-device at 30 tokens per second.
Apple now has access to Google’s Gemini model, allowing distillation into smaller, more capable models for iPhones and iPads.
Knowledge distillation involves using a large model's reasoning to train a smaller model, enhancing its capabilities.
Smartphones lack compelling upgrade reasons; on-device generative AI could change this by offering tangible improvements.
Spec wars may shift to parameters, with larger models offering better performance and user experience.
On-device AI creates hardware requirements that older devices can't meet, driving upgrades.
Apple and Google are leveraging distillation to scale AI models for different hardware tiers.
On-device AI could become the primary differentiator between phone tiers and generations.
Consumer interest in AI as a purchase driver is growing, with immediate feedback on model performance.
The next major phone upgrade reason will likely be AI model capabilities rather than hardware specs.

Hasty Briefsbeta