Full threadshoker·Generative text AI (Large language models) require a ton of RAM for inference which just isn't a thing on mobile devices yet. It makes more sense to keep these features in the cloud vs making phones with unrealistic specs.View on HN