The whole thing is buggered. My latest, in car, on 26.6.2 with phone connected via bluetooth. I hold the voice button on my steering wheel down to activate Siri:
"Call Fred."
"You'll need to unlock your phone for that."
"Call Fred."
"You'll need to unlock your phone for that."
"Call Fred."
"OK, calling Fred."
Phone is locked in my pocket the entire time. Like, make your damn mind up. Every time I try to call from the car now I have to ask it at least twice, usually three times.
And a completely baffling one recently, not even anonymizing names, also in car:
"Call George."
"OK adding L.S.D. by Hallucinogen to your Liked Songs playlist."
That's also what drives me absolutely insane - I want a guide from Apple that gives me a regular syntax I can use to get predictable results with Siri. The fuzzy, natural language stuff in Siri is so bad that it's completely unpredictable what will work.
Super frustrating. It's the best thing about LLMs, that you can be so imprecise and human with your discussion, with context maintained from input to input. Siri doesn't have the predictability of an old-school fixed syntax voice agent (like Wildfire from the '90s - remember that? It worked!) and doesn't have the flexibility of a proper natural language interaction with an LLM. It lacks any sort of working memory - "Create an appointment on Tuesday at 3pm." "OK." "Name it 'Mechanic visit'" "Now playing songs by Mike and the Mechanics."
I've switched to Home Assistant and the same devices on the same network work instantly, as they should.
Apple is so terrible at this. They have nobody to blame: it's their system, their protocol (I have pre-Matter devices), their home hub, their certified devices.
Home Assistant with a pile of community-maintained python hacks runs circles around Apple.
I mean, they moved it from a deterministic set of actions to an AI-pilled LLM nightmare with probably a paucity of all test cases covered given the near-infinite permutations and actions available.
A smarter release would have been extremely limited in scope, heavily tested, and aimed simply for feature parity. Or even scoping it to just one section, like editing/adding/updating calendar events and all associated functionality. Or even exploring routing to much-less hallucinatory, more performant models like CNN for certain tasks instead of just throwing everything at probabilistic model and hoping for the best.
But that's not the way of AI-pilled middle manager thinking now.
Huh interesting idea and quite easy to test by putting the phone in a blanket in the trunk. It's definitely using the car speakers, and when I do eventually get a call through I am heard clearly, but perhaps that initial interaction is using the phone mic.
Interesting to see people are still hating on Siri after this newest update. So far I have not been having problems and it works better for me than before. I've even found a few new use cases that have worked wonderfully, like creating Reminders lists or calendar events based on what's on the screen.
Here's my tiny anecdote about Siri. It's so garbage I had to name my garage door "The garage door" because without the extra THE it would often refuse or tell me that device can't be opened.
Siri needs to be spoonfed the most basic things, and it's not getting any better. Google doesn't get any passes since my Google Home appears to be getting alzheimer's after the Gemini "upgrade".
Do a factory reset, then set it up again before doing the Gemini upgrade. This helped me a lot and it works way better now, even compared to the original non Gemini version.
Apple Intelligence broke the ability for Siri to run a shortcut that I use daily. I have a shortcut named, "What's for dinner?" I used to be able to activate Siri and says, "What's for dinner?" and it would set off to my personal mealie server, use an API key to fetch the information, then speak the result. Now, Siri looks through my messages and finds a dinner reservation from OpenTable from months ago and declares that to be fact.
Siri is the only thing in the whole universe which can make me rage. Not computers, not $dayjob, not my car, not getting lost while riding a rally in the desert, not people. Siri. Only Siri can do that.
The only voice assistant that has impressed me recently is the new ChatGPT Voice Mode with background reasoning. It's miles ahead of the other assistants.
The only downside is that it lacks the OS integration of Gemini or Siri. I've been trying to figure out a way to trigger it with the button on my pixel earbuds, but no success - seems it only works with Gemini.
Grok's voice mode is surprisingly good. I find it better than ChatGPT voice. Sol is normally better than Grok 4.7 of course, but in voice mode Sol is much dumber, and Grok edges ahead. It is also integrated with Grok Bot for a Muse-like agent that you can have a phone call with, which works really well.
I was optimistic that Apple would get this right but Siri AI is yet another embarrassment in a long continuing series of Apple embarrassments.
I went from being overly exuberant on AI to now thinking it probably should be banned outright. From a consumer perspective, the tech industry is a dead end and the internet now is at best a pipeline for content for your TV and sound system and the rest feels like it should be ignored and avoided.
Unless there are radical changes on the horizon, there is no way I’m investing any more in the Apple ecosystem and I’m saying this as a former Apple employee and fan.
Agree - I've had almost all of these problems with Siri.
Another thing I've noticed is the horrible lag with updating the UI. An example: On Apple Watch I have the temperature complication bottom center. It sometimes shows the current temperature, sometimes it hasn't updated in minutes or hours (bug #1). For example right now, it shows 66 which I know is incorrect. I tap it to go to the Weather app, a different temperature will be shown (56) (bug #2) and then about 10 seconds later (bug #3) it finally updates to the actual temperature (72).
If I go back to the home screen, finally the complication shows the correct temperature (72).
These kind of async refresh bugs are not unique to WatchOS, it's endemic throughout the Apple ecosystem.
Why? Is this poorly coded JavaScript web view apps? Is this React? Some sort of other async framework that works fine when you are testing on the Apple campus but doesn't work in the real world?
The question I'm asking is "why" - why would Apple ship code that is so horribly broken (from a UI/UX) perspective.
I thought for sure that this slow-temperature-update bug would be fixed in WatchOS 27, but it's just as bad as in 26 and before.
Sometimes the temperature UI will show "--" instead of a temperature, until it finally updates, so clearly they do have code for the "no temperature data" case.
It just boggles the mind that "6 hour stale temperature data" doesn't trigger this same behavior.
There are some very weird rules for this where you get a limited number of updates per hour/day and if you eat your budget then you don’t get any more. Or if you crash a bunch the system will just stop asking you for new data. It’s an odd, somewhat inscrutable system
If only Apple computer had employees that would fix this system (that they are 100% responsible for building) :-)
Even if Apple is too cheap to pay the $0.00001 to get me real time temperature updates, you would think there would be an OS API to query the temperature, and that the various apps and complications would query this to get the source of truth.
But from the actual UI/UX, it must be that every App, and complication, and UI element are caching their own value.
That's the only way I can explain seeing 3 different temperatures displayed within a few seconds.
(NB: these are all Apple apps, I'm not using any third party apps)
But aside from the accuracy, I’m finding requests take much longer to action. 15 seconds to add a reminder yesterday. Not sure if that’s just me or normal now?
Remember when you asked 2012 Maps for directions to the supermarket, and it’d tell you to drive off a bridge? Looks like they finally integrated it into 2026 Siri!
Gemini in android auto just refuses to place calls for you now for "safety" but will also have an extended argument with you about that at the same time.
The bug that I think was actually fixed (or maybe I just haven't hit the spot on the roulette wheel again since the beta):
- Call Mom
- What app should I use? (only offers the default Phone app
Immense rage - The only freaking app I've ever used to call her. You didn't even give me a reasonable alternative. We don't talk on FB Messenger/WhatsApp/etc (nor do I use them at all).
New Siri is _very slightly_ better for some tasks and infinitely worse in others.
I'm not sure if it's Gemini's fault or Apple's, I lean heavily towards it being Apple's fault and at the end of the day it doesn't matter if the underlying model is blame, it's the model Apple picked so it's Apple's fault.
I had hopes that "New Siri" would help wash away the stain of WWDC 2024 and the incredibly embarrassing new (old) Siri launch. Instead they've doubled down to prove their complete incompetence.
It's telling that I sit here with AirPods on my desk, an Apple Watch on my wrist, an iPad sitting next to me, an iPhone on a stand in front of me, all while typing on a MBP, and I have zero interest in touching their "AI" with a 10-ft pole while I use millions-billions of tokens other models every month. My phone would be the ideal AI device if it actually worked worth a damn. Instead I do use AI on my phone all the time, through the Claude and ChatGPT apps which would both blow away Siri if they had the access to my phone's data. I recognize the dangers in that, I've even agreed with Apple's line in the past about how they are the only ones that can be trusted with that access but they are terrible stewards who clearly don't have a clue what they are doing.
I have better results remote connecting to a Claude Code session on my computer to do anything of value instead of reaching for Siri.
It's the exact wrong place for an LLM. It SEEMS like it should be the right place, but it isn't.
The old way of doing commands and voice was far superior and made far fewer mistakes.
Funnily, I saw a skyrim mod I believe here on HN that got this right. They did fast commands through more standard voice and language recognition and shunted off more complex conversations only when that failed. It enabled them to do really impressive things with their ingame dog.
I was optimistic when I updated MacOS, iOS, and got a new series 12 watch to replace my series 8. But... it's terrible. It doesn't seem to be able to do much of what I ask, and Siri on my watch is somehow less responsive than it was on my Series 8 watch. When I talk to Siri on my watch, the Siri icon thing spins around as if it does something, then it does nothing? I don't understand.
Apple makes great products, and I am well vested in its ecosystem.
With that said - there are three areas of their products that are substandard.
- Siri, as discussed here
- autocorrect - which is insanely bad. I switched to Gboard a while back - and despite its occasional lagging (it depends too much on an internet connection) it’s such a relief. No more fighting autocorrect over words, especially after committing by hitting space.
- and… the whole way you register devices with Apple home - incredibly painful. Thinking they were trying to make it more secure, but they utterly failed at friction and usability. I suspect their ecosystem would be a lot bigger if they simplified how you interact with Home to register devices.
Try turning off autocorrect. It seems like a non-option, but I found it made my mobile typing skill greatly improve over the following months, to the point I’m now just as fast while also fully removing that frustration from my life, and also not reliant on a particular phone. Retrospectively, it makes sense: autocorrect numbs feedback, removing the mechanism by which we learn to subconsciously and quickly do what works and avoid what doesn’t.
I still think they’re in denial that their “Home” product as a whole isn’t going anywhere and still wait for them to just drop it…
That being said - not adding any more examples of “Siri bad because” for my blood pressure staying on normal levels - one thing that also is one more example how hard Apple fail at basic things:
Ok, so there’s this feature that charges your phone until it presumes you’ll need it?
Why on earth does it arbitrarily decide to be finished at 04:00 when I did set an alarm to 06:30?
I don't see the purpose in this feature even if your alarm was at 4:00. Why wouldn't you keep charging until the user picked his phone off the charger?
One thing you might do is only charge to 80% or 90%, to improve battery lifetime, unless you detected the user is going on a long flight tomorrow and needs it at 100%. But I don't see any reason for the opposite behaviour.
Siri IS bad at many things, but the Reddit post is not a particularly compelling example. I am continually astounded by the fact that Apple has had an on-device voice assistant for 15 years, on a vertical hardware/software stack that they own completely, and yet which cannot reliably perform even the most basic tasks.
Inconsistent capabilities: it can toggle do not disturb, but not "silence unknown callers". It can disable auto-brightness but not automatic screen lock.
Completely brain-dead responses even within things it is capable of affecting: "delete all alarms except for my 7:30 AM alarm" → "I deleted your 7:30 AM alarm"
And even with features that actually do work, subtle details of that feature inexplicably change over time, and they happen to be details I was relying on: asking Siri to set an alarm for a given time used to respect whatever ringtone I had last chosen (and which IS respected when I physically tap the + button to add a new alarm), but sometime over the past five or so years it switched to creating alarms with the default sound.
And so many more things that I can't recall the top of my head right now. Siri is an inexcusable failure by a company that supposedly prides itself on user experience and which has had every opportunity for a decade and a half to do better.
edit: One more I just remembered. For a decade, me saying "call [WIFE'S NAME]" would 100% of the time result in my wife being called. Back in the spring, Siri inexplicably decided to call the number of a woman I hadn't seen in the past decade, and whose name is pronounced the same but spelled differently. I had to delete that contact from my phone to be able to call my wife again using Siri.
As far as I can tell, the AI team assembled in Apple has a fundamental instinct problem and I was hoping that the exodus earlier this year will fix it, but looks like it did not.
Almost as if the non-deterministic matrix weight program doesn't take explicit instructions and produce the same answer each time. If only we had some sort of language we could use to "program" something to work that way.
While certainly drastically better than before, it still fails at so many of the basics. I use primarily in the car, when need to be hands free. My requests are almost always say, "Directions to Starbucks". Which Starbucks.. and proceeds to list multiple then have to remind it each and every time the one first one... the one closest to me.. the one always go to... etc. So then try to do Directions to Starbucks (Town Name) (that I'm currently in, and still asks did you mean this one or this one, or everyone in a while will say did you mean City (state I'm in), aka the one down the street or did you mean that city but in another state a hundred miles away that you have never gone to. Those types of things always happened with old Siri and persist with new Siri. But now almost always when asking for reminders etc, before it would handle fine, not it says did you mean "what I said), or swapping out word for another someone similar word, etc. The problem before was it had essentially no memory or ability to learn about me. With LLMS thought that would finally change, but it has not ing general. I can say call John. And it will say did you mean John (the one you call every day for the past decade) or do you mean the John, you have never called but is in your contacts from High School? ChatGPT is amazing, and knows me and never makes any of those mistakes but of course is less integrated than Siri can be. Here's for hoping with the new LLM foundation Siri will be able to grow more quickly, but right now, it is the dumbest of the bunch. I get they don't want it to be a full chat bot, and I'm ok with it being more limited, but if it's going to be limited to a certain domain, such as control the home, setting alarms, reminders, directions, it should be absolutely rock solid at that.
Why don't you just say which Starbucks you want to go to? Or even better if you go there all the time and it's the closest one to your home, just remember the way
It's unmitigated infantilism. Outside of the physically disabled, who would need such a feature? Is touching your phone to make a call less convenient than speaking to it? If touching your phone is so inconvenient, then why do most adults spend hours per day on their phones?
What most of these sorts of features do is not to add convenience; instead they remove the need to exercise impulse control. They reduce "friction," which means you do the action more often. But they don't improve convenience. Convenience is not needing to walk outside to the well to fetch a bucket of water. It's not someone making an app so automatic that you no longer need to use your executive function.
Just no good reason. No reason why someone would actually need one. It's like an automatic lift-gate motor in an SUV. It's not more convenient. Pressing a button to close your trunk is only more convenient than manually closing your trunk if you have a physical disability. An in the long term, it's less convenient: you actually have the option of closing it manually removed since this is bad for the motor. If the motor fails in the wrong position, then your trunk may be stuck open. Replacing an old-fashioned piston on a trunk is cheap an easy. Replacing an internal motor is much more expensive and is beyond the skillset of most people.
A voice assistant is similar in an analogous way. A button works every time you press it, but a voice assistant works very sporadically. It's hardly convenient to repeat yourself frequently and randomly. A voice assistant requires internet access, which wastes your data (again, when you could have just pressed a button) and the quality of the voice assistant will change over time. It's less consistent, more expensive, and more complex. And for what? So you can use your phone more often? You already use your phone too much.
if people find a feature for the physically disabled useful they'll use it. should manufacturers ship disabled-only versions with those features enabled and disable them in the other devices?
"Hey Siri, continue playing my audiobook" used to work, now it starts a random book from my library, even ones I've already listened to. "Hey Siri, play" used to continue playing whatever was last playing, now it starts random songs on Apple Music. Those are the only two things I use Siri for apart from controlling lights.
The product managers and developers at Apple have a unique capability to gaslight an atrocious user experience as excellent.
Folks who parrot the same are just a symptom of the underlying epidemic that permeates the ecosystem.
Some of their engineering deficiencies just happen to be masked by integration with better hardware.
iOS mmap limit is ridiculous.
But it paved the way for us to work around the restriction by rebuilding Dip's architecture resulting in a better, more simplified mental model along with increased performance and parallelism.
These benefits are applicable cross-platform and not just for Apple's OS. This subsequently also acted as an indirect but causal trigger to later build our own KV engine: Singularity.
So my experience with Apple has been bittersweet. The frustration with their ecosystem was beneficial in hindsight.
The brigades are relentless and much bigger than any of the real comments we can make.
You're getting downvoted here (possibly flagged within the next few hours), but I see this happen on half the posts made here daily. Someone notices out loud how fake the engagement has become and they get railroaded.
The whole thing is buggered. My latest, in car, on 26.6.2 with phone connected via bluetooth. I hold the voice button on my steering wheel down to activate Siri:
"Call Fred."
"You'll need to unlock your phone for that."
"Call Fred."
"You'll need to unlock your phone for that."
"Call Fred."
"OK, calling Fred."
Phone is locked in my pocket the entire time. Like, make your damn mind up. Every time I try to call from the car now I have to ask it at least twice, usually three times.
And a completely baffling one recently, not even anonymizing names, also in car:
"Call George."
"OK adding L.S.D. by Hallucinogen to your Liked Songs playlist."
WTAF.
I mostly like it, but it randomly changes what commands it accepts.
Last week:
Me: Hey Siri, turn off the bedroom lights. Siri: I’m sorry, I can’t change a scene. Me: Hey Siri, turn the bedroom lights off. Siri: OK!
This week:
Me: Hey Siri, turn the bedroom lights off. Siri: I’m sorry, I can only do one thing at a time. Me: Hey Siri, turn off the bedroom lights. Siri: OK!
Next week:
Me: Hey Siri —— Siri: The red zone has always been for loading and unloading of passengers. There's never stopping in a white zone.
Like, please just pick one and stick that in the integration tests. I promise I’ll learn the syntax if you choose an official one. Any official one.
That's also what drives me absolutely insane - I want a guide from Apple that gives me a regular syntax I can use to get predictable results with Siri. The fuzzy, natural language stuff in Siri is so bad that it's completely unpredictable what will work.
Super frustrating. It's the best thing about LLMs, that you can be so imprecise and human with your discussion, with context maintained from input to input. Siri doesn't have the predictability of an old-school fixed syntax voice agent (like Wildfire from the '90s - remember that? It worked!) and doesn't have the flexibility of a proper natural language interaction with an LLM. It lacks any sort of working memory - "Create an appointment on Tuesday at 3pm." "OK." "Name it 'Mechanic visit'" "Now playing songs by Mike and the Mechanics."
“Good night.”
Literal 10 seconds later.
“Setting the scene good night.”
Another 2-3 seconds later, the lights go off.
I don’t understand what it’s trying to do, simple tasks should be instant, what am I waiting for? The entire Siri became unusable and simply trash.
Edit: For reference, it used to turn off the lights as I complete my sentence and put my wrist down.
HomeKit is so painfully slow and flaky.
I've switched to Home Assistant and the same devices on the same network work instantly, as they should.
Apple is so terrible at this. They have nobody to blame: it's their system, their protocol (I have pre-Matter devices), their home hub, their certified devices.
Home Assistant with a pile of community-maintained python hacks runs circles around Apple.
Yes! it's so darn slow, then there's a 50/50 chance it either does the thing, or does nothing at all. It's driving me a bit bonkers.
I mean, they moved it from a deterministic set of actions to an AI-pilled LLM nightmare with probably a paucity of all test cases covered given the near-infinite permutations and actions available.
A smarter release would have been extremely limited in scope, heavily tested, and aimed simply for feature parity. Or even scoping it to just one section, like editing/adding/updating calendar events and all associated functionality. Or even exploring routing to much-less hallucinatory, more performant models like CNN for certain tasks instead of just throwing everything at probabilistic model and hoping for the best.
But that's not the way of AI-pilled middle manager thinking now.
One more for you.
While I’m away from my phone: “hey dingus play podcasts”
“I’m sorry you don’t have Apple Music subscription”
Apple should stop pushing for subscriptions. Especially when nothing is working anymore.
I've seen bugs like this which have led me to wonder if apple has a problem with generating UUIDs which aren't really unique.
It's the only way I can explain some of the totally bizarre Siri (and other iOS and macOS) bugs.
Yeah I get the same thing when trying to open the garage door. Have never been able to figure out what the rule is.
Good thing Siri and Alexa are going through a phase, otherwise Siri would’ve added that LSD to your Amazon shopping list!
At least then it would be trying to make me happy.
I wonder if it’s using the microphone in your pocket instead of the cars mic.
Huh interesting idea and quite easy to test by putting the phone in a blanket in the trunk. It's definitely using the car speakers, and when I do eventually get a call through I am heard clearly, but perhaps that initial interaction is using the phone mic.
Wait, so you're telling me security can be overcome by simply repeating the request 3 times?
Uh, can anyone do that? Can pickpockets and government stooges do the same thing to get into my phone?
Interesting to see people are still hating on Siri after this newest update. So far I have not been having problems and it works better for me than before. I've even found a few new use cases that have worked wonderfully, like creating Reminders lists or calendar events based on what's on the screen.
Here's my tiny anecdote about Siri. It's so garbage I had to name my garage door "The garage door" because without the extra THE it would often refuse or tell me that device can't be opened.
Siri needs to be spoonfed the most basic things, and it's not getting any better. Google doesn't get any passes since my Google Home appears to be getting alzheimer's after the Gemini "upgrade".
Do a factory reset, then set it up again before doing the Gemini upgrade. This helped me a lot and it works way better now, even compared to the original non Gemini version.
Apple Intelligence broke the ability for Siri to run a shortcut that I use daily. I have a shortcut named, "What's for dinner?" I used to be able to activate Siri and says, "What's for dinner?" and it would set off to my personal mealie server, use an API key to fetch the information, then speak the result. Now, Siri looks through my messages and finds a dinner reservation from OpenTable from months ago and declares that to be fact.
Siri is the only thing in the whole universe which can make me rage. Not computers, not $dayjob, not my car, not getting lost while riding a rally in the desert, not people. Siri. Only Siri can do that.
The only voice assistant that has impressed me recently is the new ChatGPT Voice Mode with background reasoning. It's miles ahead of the other assistants.
The only downside is that it lacks the OS integration of Gemini or Siri. I've been trying to figure out a way to trigger it with the button on my pixel earbuds, but no success - seems it only works with Gemini.
Grok's voice mode is surprisingly good. I find it better than ChatGPT voice. Sol is normally better than Grok 4.7 of course, but in voice mode Sol is much dumber, and Grok edges ahead. It is also integrated with Grok Bot for a Muse-like agent that you can have a phone call with, which works really well.
Yeah but then you're using Grok.
I was optimistic that Apple would get this right but Siri AI is yet another embarrassment in a long continuing series of Apple embarrassments.
I went from being overly exuberant on AI to now thinking it probably should be banned outright. From a consumer perspective, the tech industry is a dead end and the internet now is at best a pipeline for content for your TV and sound system and the rest feels like it should be ignored and avoided.
Unless there are radical changes on the horizon, there is no way I’m investing any more in the Apple ecosystem and I’m saying this as a former Apple employee and fan.
Agree - I've had almost all of these problems with Siri.
Another thing I've noticed is the horrible lag with updating the UI. An example: On Apple Watch I have the temperature complication bottom center. It sometimes shows the current temperature, sometimes it hasn't updated in minutes or hours (bug #1). For example right now, it shows 66 which I know is incorrect. I tap it to go to the Weather app, a different temperature will be shown (56) (bug #2) and then about 10 seconds later (bug #3) it finally updates to the actual temperature (72).
If I go back to the home screen, finally the complication shows the correct temperature (72).
These kind of async refresh bugs are not unique to WatchOS, it's endemic throughout the Apple ecosystem.
Why? Is this poorly coded JavaScript web view apps? Is this React? Some sort of other async framework that works fine when you are testing on the Apple campus but doesn't work in the real world?
This is just background updates being slow
Right, that seems to be what is happening.
The question I'm asking is "why" - why would Apple ship code that is so horribly broken (from a UI/UX) perspective.
I thought for sure that this slow-temperature-update bug would be fixed in WatchOS 27, but it's just as bad as in 26 and before.
Sometimes the temperature UI will show "--" instead of a temperature, until it finally updates, so clearly they do have code for the "no temperature data" case.
It just boggles the mind that "6 hour stale temperature data" doesn't trigger this same behavior.
There are some very weird rules for this where you get a limited number of updates per hour/day and if you eat your budget then you don’t get any more. Or if you crash a bunch the system will just stop asking you for new data. It’s an odd, somewhat inscrutable system
If only Apple computer had employees that would fix this system (that they are 100% responsible for building) :-)
Even if Apple is too cheap to pay the $0.00001 to get me real time temperature updates, you would think there would be an OS API to query the temperature, and that the various apps and complications would query this to get the source of truth.
But from the actual UI/UX, it must be that every App, and complication, and UI element are caching their own value.
That's the only way I can explain seeing 3 different temperatures displayed within a few seconds.
(NB: these are all Apple apps, I'm not using any third party apps)
First request after enabling Siri AI Beta:
“Play Powerwolf on Apple Music.”
“Okay, here is Power by Katy Perry.”
“Play PowerWOLF on Apple Music.”
“Okay, here is Power by Katy Perry.”
“Fine, just play Sabaton”
It got that right I guess :/
But aside from the accuracy, I’m finding requests take much longer to action. 15 seconds to add a reminder yesterday. Not sure if that’s just me or normal now?
Remember when you asked 2012 Maps for directions to the supermarket, and it’d tell you to drive off a bridge? Looks like they finally integrated it into 2026 Siri!
I'm sure there will be someone to defend Apple here. "Maybe it's a privacy thing"
What a pile of garbage Siri is.
It's so private, even Siri can't do anything when you ask :D.
Gemini in android auto just refuses to place calls for you now for "safety" but will also have an extended argument with you about that at the same time.
The bug that I think was actually fixed (or maybe I just haven't hit the spot on the roulette wheel again since the beta):
- Call Mom
- What app should I use? (only offers the default Phone app
Immense rage - The only freaking app I've ever used to call her. You didn't even give me a reasonable alternative. We don't talk on FB Messenger/WhatsApp/etc (nor do I use them at all).
New Siri is _very slightly_ better for some tasks and infinitely worse in others.
I'm not sure if it's Gemini's fault or Apple's, I lean heavily towards it being Apple's fault and at the end of the day it doesn't matter if the underlying model is blame, it's the model Apple picked so it's Apple's fault.
I had hopes that "New Siri" would help wash away the stain of WWDC 2024 and the incredibly embarrassing new (old) Siri launch. Instead they've doubled down to prove their complete incompetence.
It's telling that I sit here with AirPods on my desk, an Apple Watch on my wrist, an iPad sitting next to me, an iPhone on a stand in front of me, all while typing on a MBP, and I have zero interest in touching their "AI" with a 10-ft pole while I use millions-billions of tokens other models every month. My phone would be the ideal AI device if it actually worked worth a damn. Instead I do use AI on my phone all the time, through the Claude and ChatGPT apps which would both blow away Siri if they had the access to my phone's data. I recognize the dangers in that, I've even agreed with Apple's line in the past about how they are the only ones that can be trusted with that access but they are terrible stewards who clearly don't have a clue what they are doing.
I have better results remote connecting to a Claude Code session on my computer to do anything of value instead of reaching for Siri.
It's the exact wrong place for an LLM. It SEEMS like it should be the right place, but it isn't.
The old way of doing commands and voice was far superior and made far fewer mistakes.
Funnily, I saw a skyrim mod I believe here on HN that got this right. They did fast commands through more standard voice and language recognition and shunted off more complex conversations only when that failed. It enabled them to do really impressive things with their ingame dog.
I was optimistic when I updated MacOS, iOS, and got a new series 12 watch to replace my series 8. But... it's terrible. It doesn't seem to be able to do much of what I ask, and Siri on my watch is somehow less responsive than it was on my Series 8 watch. When I talk to Siri on my watch, the Siri icon thing spins around as if it does something, then it does nothing? I don't understand.
Apple makes great products, and I am well vested in its ecosystem.
With that said - there are three areas of their products that are substandard.
- Siri, as discussed here
- autocorrect - which is insanely bad. I switched to Gboard a while back - and despite its occasional lagging (it depends too much on an internet connection) it’s such a relief. No more fighting autocorrect over words, especially after committing by hitting space.
- and… the whole way you register devices with Apple home - incredibly painful. Thinking they were trying to make it more secure, but they utterly failed at friction and usability. I suspect their ecosystem would be a lot bigger if they simplified how you interact with Home to register devices.
Try turning off autocorrect. It seems like a non-option, but I found it made my mobile typing skill greatly improve over the following months, to the point I’m now just as fast while also fully removing that frustration from my life, and also not reliant on a particular phone. Retrospectively, it makes sense: autocorrect numbs feedback, removing the mechanism by which we learn to subconsciously and quickly do what works and avoid what doesn’t.
I still think they’re in denial that their “Home” product as a whole isn’t going anywhere and still wait for them to just drop it…
That being said - not adding any more examples of “Siri bad because” for my blood pressure staying on normal levels - one thing that also is one more example how hard Apple fail at basic things:
Ok, so there’s this feature that charges your phone until it presumes you’ll need it?
Why on earth does it arbitrarily decide to be finished at 04:00 when I did set an alarm to 06:30?
Why oh why…
This may be due to the Clean Energy Charging setting, which must make some guesses about your energy provider's usage tier and supplier mix.
I don't see the purpose in this feature even if your alarm was at 4:00. Why wouldn't you keep charging until the user picked his phone off the charger?
One thing you might do is only charge to 80% or 90%, to improve battery lifetime, unless you detected the user is going on a long flight tomorrow and needs it at 100%. But I don't see any reason for the opposite behaviour.
Is there any way to tap into Facetime from a non-Apple device?
From what I can see the only thing that durably keeps me on iOS at this point is Facetime.
(Siri AI is bad. Lots of "Did you mean". I'm back on Siri Classic now.)
Siri IS bad at many things, but the Reddit post is not a particularly compelling example. I am continually astounded by the fact that Apple has had an on-device voice assistant for 15 years, on a vertical hardware/software stack that they own completely, and yet which cannot reliably perform even the most basic tasks.
Inconsistent capabilities: it can toggle do not disturb, but not "silence unknown callers". It can disable auto-brightness but not automatic screen lock.
Completely brain-dead responses even within things it is capable of affecting: "delete all alarms except for my 7:30 AM alarm" → "I deleted your 7:30 AM alarm"
And even with features that actually do work, subtle details of that feature inexplicably change over time, and they happen to be details I was relying on: asking Siri to set an alarm for a given time used to respect whatever ringtone I had last chosen (and which IS respected when I physically tap the + button to add a new alarm), but sometime over the past five or so years it switched to creating alarms with the default sound.
And so many more things that I can't recall the top of my head right now. Siri is an inexcusable failure by a company that supposedly prides itself on user experience and which has had every opportunity for a decade and a half to do better.
edit: One more I just remembered. For a decade, me saying "call [WIFE'S NAME]" would 100% of the time result in my wife being called. Back in the spring, Siri inexplicably decided to call the number of a woman I hadn't seen in the past decade, and whose name is pronounced the same but spelled differently. I had to delete that contact from my phone to be able to call my wife again using Siri.
As far as I can tell, the AI team assembled in Apple has a fundamental instinct problem and I was hoping that the exodus earlier this year will fix it, but looks like it did not.
Almost as if the non-deterministic matrix weight program doesn't take explicit instructions and produce the same answer each time. If only we had some sort of language we could use to "program" something to work that way.
So it’s pretty obvious that Siri sucks.
Is Apple acknowledging it? Do they have an explanation? What’s their narrative?
Apple would never ever acknowledge that anything of theirs is less than ecstatically fantastic.
I updated my barely 1-year-old Apple Watch Ultra 3 to watchOS 27. Never installed any betas on it.
Tried to set a timer via Siri
after a spinner for several seconds:
"Cannot set a timer because the Clock app is not installed on this device."
F to the ninth heck whoever is in charge of that crap at Apple.
Who would have thought that a watch needed a Clock app?
Setting timers is literally the only thing I ever use Siri for, so it's not like this is some edge case.
After seeing how bad the Apple Watch became, I skipped updating it this year.
There is no use for it if I cannot simply raise my wrist and set up a timer or create a reminder without any problems.
I’m not going to have a full conversation with Siri when I’m moving about and need to remember something or need a quick timer for a meal.
Do these engineers ever use these products or do they simply throw it away after the fancy presentations they do?
Hopefully AI will eventually do to operating systems what it has started to do to Photoshop and Office etc.
These corpulent Hutts are long overdue for a choking
At this point if people keep buying Apple products expecting Siri to get better, that's on them.
Couldn't they use something like Jev to make Siri really good and fast at processing the most frequent commands?
Someone who knows how to do it should make a better Siri and then fuck Apple by selling it to them. For a really expensive price.
While certainly drastically better than before, it still fails at so many of the basics. I use primarily in the car, when need to be hands free. My requests are almost always say, "Directions to Starbucks". Which Starbucks.. and proceeds to list multiple then have to remind it each and every time the one first one... the one closest to me.. the one always go to... etc. So then try to do Directions to Starbucks (Town Name) (that I'm currently in, and still asks did you mean this one or this one, or everyone in a while will say did you mean City (state I'm in), aka the one down the street or did you mean that city but in another state a hundred miles away that you have never gone to. Those types of things always happened with old Siri and persist with new Siri. But now almost always when asking for reminders etc, before it would handle fine, not it says did you mean "what I said), or swapping out word for another someone similar word, etc. The problem before was it had essentially no memory or ability to learn about me. With LLMS thought that would finally change, but it has not ing general. I can say call John. And it will say did you mean John (the one you call every day for the past decade) or do you mean the John, you have never called but is in your contacts from High School? ChatGPT is amazing, and knows me and never makes any of those mistakes but of course is less integrated than Siri can be. Here's for hoping with the new LLM foundation Siri will be able to grow more quickly, but right now, it is the dumbest of the bunch. I get they don't want it to be a full chat bot, and I'm ok with it being more limited, but if it's going to be limited to a certain domain, such as control the home, setting alarms, reminders, directions, it should be absolutely rock solid at that.
Why don't you just say which Starbucks you want to go to? Or even better if you go there all the time and it's the closest one to your home, just remember the way
It's unmitigated infantilism. Outside of the physically disabled, who would need such a feature? Is touching your phone to make a call less convenient than speaking to it? If touching your phone is so inconvenient, then why do most adults spend hours per day on their phones?
What most of these sorts of features do is not to add convenience; instead they remove the need to exercise impulse control. They reduce "friction," which means you do the action more often. But they don't improve convenience. Convenience is not needing to walk outside to the well to fetch a bucket of water. It's not someone making an app so automatic that you no longer need to use your executive function.
You literally cannot think of any reason why someone would use a voice assistant?
Just no good reason. No reason why someone would actually need one. It's like an automatic lift-gate motor in an SUV. It's not more convenient. Pressing a button to close your trunk is only more convenient than manually closing your trunk if you have a physical disability. An in the long term, it's less convenient: you actually have the option of closing it manually removed since this is bad for the motor. If the motor fails in the wrong position, then your trunk may be stuck open. Replacing an old-fashioned piston on a trunk is cheap an easy. Replacing an internal motor is much more expensive and is beyond the skillset of most people.
A voice assistant is similar in an analogous way. A button works every time you press it, but a voice assistant works very sporadically. It's hardly convenient to repeat yourself frequently and randomly. A voice assistant requires internet access, which wastes your data (again, when you could have just pressed a button) and the quality of the voice assistant will change over time. It's less consistent, more expensive, and more complex. And for what? So you can use your phone more often? You already use your phone too much.
if people find a feature for the physically disabled useful they'll use it. should manufacturers ship disabled-only versions with those features enabled and disable them in the other devices?
The #1 reason is while driving. It's also the only reason I have Siri enabled at all since you can't use CarPlay without enabling it.
Edit: There are other cases where hands are not free as well - like cooking - where it's useful to be able to start music or add a todo.
>The #1 reason is while driving.
Just wait until you're not driving.
Ok “ever” “drive”
I listen to the radio and don't use my mobile phone in my car. It's awesome.
Shortcuts automation is significantly better than Siri, they should just train a small model to work with it instead of whatever Siri is
I wouldn’t mind it being useless if only I could remove it from my device. :(
"Hey Siri, continue playing my audiobook" used to work, now it starts a random book from my library, even ones I've already listened to. "Hey Siri, play" used to continue playing whatever was last playing, now it starts random songs on Apple Music. Those are the only two things I use Siri for apart from controlling lights.
Well, everything but privacy and efficiency because it runs on your device and not on a remote server. This also explains why Siri still isn’t good.
The product managers and developers at Apple have a unique capability to gaslight an atrocious user experience as excellent.
Folks who parrot the same are just a symptom of the underlying epidemic that permeates the ecosystem.
Some of their engineering deficiencies just happen to be masked by integration with better hardware.
iOS mmap limit is ridiculous. But it paved the way for us to work around the restriction by rebuilding Dip's architecture resulting in a better, more simplified mental model along with increased performance and parallelism.
These benefits are applicable cross-platform and not just for Apple's OS. This subsequently also acted as an indirect but causal trigger to later build our own KV engine: Singularity.
So my experience with Apple has been bittersweet. The frustration with their ecosystem was beneficial in hindsight.
How is a link to a reddit thread with 2 sentences on the front page? This place has gone WAY downhill.
Because everyone agrees with the sentiment, Apple lovers and haters alike.
Yeah, what news is this? Why are we having a discussion about a gripe from another discussion forum?
The brigades are relentless and much bigger than any of the real comments we can make.
You're getting downvoted here (possibly flagged within the next few hours), but I see this happen on half the posts made here daily. Someone notices out loud how fake the engagement has become and they get railroaded.