Apple Intelligence sets a baseline: a small on-device model integrated into the OS for summarization, writing tools, and Siri. Standalone apps push further but require you to install and trust them. Both are valid; they solve different problems.
What Apple Intelligence does well
Deep OS integration. Writing Tools work in any text field. Image cleanup is a Photos.app primitive. Siri can call into private cloud compute when the on-device model is too small. Apple's own paper confirms a ~3B on-device model — small enough to be fast, large enough for OS-level helper tasks.
Where Apple Intelligence stops short
Apple Intelligence is a helper — it summarizes notifications and rewrites your emails. It does not own your data. It does not build a knowledge graph. It does not coach you toward a goal. It is OS plumbing, not an application.
What standalone apps add
On-device LLMs let standalone apps do things Apple Intelligence cannot — long-running coaching, custom system prompts, full ownership over your knowledge graph, and bigger models when your phone can handle them. Sovereign currently runs Gemma 2B; on a recent iPhone Pro it can run 9B comfortably.
Trust trade-off
Apple has the strongest privacy brand in tech. A standalone app needs to earn that trust separately. The way it does that is open code (verifiable claims), zero-knowledge architecture (technical guarantee), and conservative defaults (opt-in cloud sync only). Sovereign does all three.
When to use each
Apple Intelligence for system-level convenience. A standalone app for the data you actually care about — your journal, your private notes, your skills, your second brain. The two coexist; they aren't in competition. See how Sovereign is laid out for an example of what a standalone private AI looks like.
About Sovereign — A privacy-first AI personal assistant that runs entirely on your iPhone. On-device LLM, zero-knowledge encryption, and a coach that learns from your own words. See how it works or visit the homepage.