If I were on Apple's board, I would force three priorities onto the agenda — none of them incremental, all of them existential.
First: actually ship something for Apple Intelligence. Not a roadmap. Not a keynote slide. Code that runs. The AI industry is operating in 2026 and Apple is still running on 2020 assumptions — that taste and ecosystem lock-in buy infinite runway. They do not. Siri, the supposed foundation, is so broken that the first thing a technically literate user does upon unboxing a new iPhone is navigate to Settings and execute it. I have never once used Siri intentionally. I have, however, had it detonate in the middle of conversations, triggered by ambient music or a friend's sentence that rhymed, phonetically, with 'Hey.' Kill Siri. Build something real. The absence of a concrete Apple AI product makes specific critique impossible — which is itself the indictment. There is nothing to critique because there is nothing.
Second: stop the battery theater. A watch that requires nightly charging is not a watch — it is an anxiety device on a strap. My threshold for purchasing my next Apple Watch is 7 days of battery life (or I am staying with my solar powered G-Shock). For a phone under average use, 72 hours is the floor, not an aspiration. This is not purely a hardware problem. It is a software discipline problem. Kill the background processes, the silent syncs, the ambient telemetry — all the metabolic overhead Apple has quietly normalized. And end, with full public acknowledgment, whatever architectural decisions appear designed to accelerate device fatigue and drive upgrade cycles. Users have noticed. The cynicism is priced in.
Third: iCloud sync is, in 2026, still a problem that should have been solved a decade ago. Not optimized — solved. The baseline contract is simple: create a Shortcut on an iPhone, and within seconds it exists on the iPad, the work MacBook, the home MacBook. That's it. That's the whole ask. Instead: minutes of latency on 5G. Apps that require a force-quit to even attempt a pull. Conflict dialogs that surface after the damage is done. 638 notes on the phone, 592 on the MacBook, and some unknown third number on every other device — no single source of truth, just a distributed argument between machines that will never resolve itself. The performance tradeoffs are understood. Push cadence, fetch scheduling, conflict resolution strategy — these are real engineering decisions with real costs. But those decisions are supposed to operate within the constraint that sync does not break. That's not a feature request; it's the premise. Everything else is optimization on top of a foundation that, right now, doesn't exist.
Fourth, as a bonus: revert the liquid glass design language and admit it was a mistake. No one asked for a real-time ray-traced simulation of frosted glass rendered over UI text. The capability is real; the judgment to deploy it is absent. It harms legibility. It taxes the GPU. It accomplishes nothing a flat, disciplined interface does not accomplish more cleanly. Microsoft went through this exact indulgence — their Fluent/Mica material phase — and I called it wasteful then too. Google, for all my reservations about Material Design's rounded corners and box shadows, at least understood where to stop: they applied the metaphor only where it clarified, then held the line. Apple and Microsoft both confused technical demonstration with design. They are not the same thing. iOS 7's flat vocabulary was not perfect, but its sins were sins of omission. Liquid glass is a sin of commission, and commissions are harder to forgive.