Opinion: Siri is not that behind

[fonzy91]

Siri still has its drawbacks and doesn’t sound the best, but I use it constantly. To run scenes in my home, play music, start workouts, start navigations, and make phone calls. Everyone wants an “LLM Siri,” and sure, that’ll feel like a huge leap: more natural, conversational, and back-and-forth. But honestly, I don’t think that’s the real challenge Apple is dealing with.

What Apple announced a year and a half ago still goes far beyond what any other OS or assistant is doing today. Google only just made similar announcements, and Alexa Plus is getting pretty mixed reviews. I think Apple’s focus is on solving the harder problem, making Siri genuinely useful across your apps and devices, not just giving you LLM-style answers. The real challenge is integrating those two layers, as Craig Federighi mentioned.

When you think about it, what Apple described back then was essentially an agentic AI, one that can understand app UIs, navigate them, and carry out complex tasks on your behalf. At the time, nothing like that existed. Even now, there’s still no equivalent on Android. Gemini can handle some basic actions, but not at the depth Apple originally outlined. So it makes sense that it’s taking them time to make it work. But in reality, Apple isn’t “behind” — no one else has delivered those features either.

And if you’ve tried Workout Buddy on the Apple Watch, I think that’s a glimpse of what’s coming. It sounds incredibly natural (ChatGPT-level voice quality) and generates context-aware feedback in real time based on your pace and workout history. That shows Apple is more than capable of building an LLM-level system.

Interested in hearing everyone's thoughts!

 

[Paddle1]

Until Apple actually releases something this doesn't really mean anything. They have yet to deliver on what was promised for the iPhone 16.

 

[Lounge vibes 05]

I agree.
I’ve had Alexa+ for about two months now, and it is, to put it lightly, a mess.
Because of how unpredictable large language models are, it pretty much wears its flaws on its sleeve.
Compared to the previous command system of the regular Alexa, it is dramatically slower. Like, noticeably slower. Responses that would come between 3 and five seconds on the old version now take between 7 and 12 seconds, specifically because of how LLMs work.
Frequently it doesn’t quite understand what you are asking, even simple home commands get lost somewhere, and it’s capability to accurately answer questions has significantly gone down.
And its conversational aspect is probably the worst part of it to me, it just rambles and rambles on. It is extremely chatty, to an annoying degree.

So there’s the problem. It sounds more natural and it’s responses are more human like, but that doesn’t exactly equal a better experience.


On the other hand, Siri I have actually found to be able to answer more Internet based questions recently, using a wide variety of sources. The problem is outside of that, it’s still very much stuck in 2012 with its capabilities, and there is absolutely no consistency between platforms.
That’s the biggest issue with Siri, it’s very inconsistent, even across Apple‘s own platforms. The Siri on the iPhone isn’t the same Siri that’s on the HomePod, and that isn’t the same Siri that’s on the Apple TV. And even on the iPhone, sometimes Siri requires the phone to be unlocked to do certain functions and sometimes it doesn’t and they’re really appears to be no consistency with these things.

And I honestly don’t think that using LLMs for everything is going to fix the problem, if anything, it could make it significantly worse.

 

[HouseLannister]

Gemini can handle some basic actions

Understatement of the year. You need a 30-day Siri detox and some time with a Pixel. You are experiencing the Stockholm syndrome.

 

[fonzy91]

Understatement of the year. You need a 30-day Siri detox and some time with a Pixel. You are experiencing the Stockholm syndrome.

My best friend has a Pixel 10 and we tested Siri against Gemini assistant for fun, and yes, Gemini did better in certain things, but it also failed in others, it honestly didn't feel like a crazy difference. A lot of features Google announced weren't quite working. But it was a quick, generic test.