Time to Deep Six Siri
Tim Cook is out. John Ternus is in. The analysts are hyperventilating about Apple’s AI deficit. The stock dipped. The hot takes are flying.
Here’s the Apple AI executive summary and call to action: kill Siri. License Claude, Gemini, or GPT. Ship it. Go home.
That’s it. That’s the plan.
Siri is like the Dead Parrot from the Monty Python skit. Apple keeps insisting it’s resting, it’s pining, it’s about to do something remarkable. It is not. It has ceased to be.
Siri launched in 2011 and was fine for its moment. Then ChatGPT landed in late 2022 and changed what “AI assistant” means. Even Google, who had the underlying technology sitting in their own lab, got caught flat-footed. The game changed. Time to move on.
Here’s the question nobody is asking: what does Siri actually do that Claude, Gemini, or ChatGPT can’t do better?
Take your time. I’ll wait.
The answer is nothing. Siri can set alarms, call your contacts, and tell you the weather. Those aren’t AI problems. Those are API calls. Any model can make an API call — and agentic AI already handles HomeKit, CarPlay, and device control. Keep the wake word, keep the hardware button, keep the on-device APIs. Just put a new face on whatever agentic model you license. Leave Siri there for legacy if you must — but give the AI version a new identity. How about Iris?
Of course, none of this is probably happening in such a blunt manner. The reason Siri isn’t already dead is likely political, not technical. There’s an entire organization built around it: careers, budgets, VP titles, performance reviews. Nobody in that fiefdom is going to recommend their own elimination. Microsoft dealt with exactly this problem for years — entrenched teams protecting turf while the world moved around them. It took brutal top-down decisions to cut through it.
We don’t need a live chatbot talking through a dead parrot.
Beyond Siri, the rest of AI integration into iOS and macOS is a long game. There’s a lot to think through and a lot to dogfood before touching an OS that hundreds of millions of people depend on daily. Microsoft learned this the hard way — Copilot got shoved into Notepad, Paint, File Explorer, and a dedicated keyboard key nobody asked for. Recall, the feature that screenshots everything you do, became a privacy nightmare before it ever shipped. Parts of Windows 11 had to be walked back entirely. Users noticed. They weren’t happy.
Apple doesn’t need to repeat that. The platform works. The AI that needs to be integrated deeply — the kind that acts across your apps, understands your context, takes real action on your behalf — requires time, testing, and honest internal dogfooding before it goes near a production release. There is no urgency to make a mess. Ternus knows how to build things that don’t embarrass you. Apply that standard here.
Nobody is buying an iPhone because Siri is great. Nobody is switching to Android because Siri is bad. The AI that iPhone users actually want is already on their phones — it’s called Claude, or ChatGPT, or Gemini, and they downloaded it from the App Store.
Apple owns the front door. The wake word. The hardware button. The ecosystem. That’s the moat. The intelligence behind the door was never the point.
The analyst complaints about the Apple AI deficit are a tempest in a teapot.
Ternus is a hardware engineer. He spent 25 years building things that have to work. The first real test of his tenure isn’t an AI horserace. It’s whether he has the clarity to look at thirteen years of Siri and say: we lost, we’re done, here’s what we’re doing instead.
That’s not a retreat. That’s strategy.
Davey Jones is waiting. Send Siri down.
Cranky Old Guy publishes at mecrankyoldguy.com. He has roughly 50 years of experience in computing and software. He holds a long-term personal investment in Apple.