121–130 of 533 posts
I think that all the deep learning models for handling speech on the Apple Watch are run locally on the watch.
Earlier quoted context omitted.
Hey Siri, “add 2 minutes to the timer”. It can’t even do something so basic as that.
The inability to set multiple simultaneous timers is annoying. I compensate by setting multiple alarms. "Delete all my alarms" works when I'm done, but it's certainly a hack.
"hey siri, set a 30 minute timer for the rice."
it works just fine for me, use it all the time.
I generally like Siri and rarely experience bugs. I use it to set timers, get the weather, play music via Spotify, occasionally dictate text messages, and look up trivia mid-conversation. It's been my replacement for Alexa ever since I got creeped out by Amazon's audio data retention. In this context I'm suprised to hear Apple employees are unhappy with it.
> I generally like Siri and rarely experience bugs This is mind-blowing to me because I'd guess that ~50% of my interactions with Siri are buggy. I mostly use it through CarPlay. It even fails to call the correct contact in my phone, which should be the easiest thing it does.
I honestly don't know why Apple can't do their own ChatGPT. They have the money / resources to then implement something like the ChatGPT 'plugin' architecture so if someone wants to turn on their lights or get specific 'private' information that apple cannot give as 'training' data. i.e. callout to 'private' API's if appropriate to a sensible LLM generated response
- private, on-device language model execution (see llama.cpp for feasibility)
- a single, consistent AI available with you wherever you go
- total access to your personal information / documents (knows your birthday, can see your meeting notes)
Because they have the hardware to run it locally, they have three very hard-to-beat advantages:
1) privacy, because the LMs can see all your stuff but none of it goes back to apple. Microsoft can't do this; they get flak every time they try and phone home with telemetry, and they don't control their platform enough to run a massive LM in the background.
2) omnipresence: if you're in the apple ecosystem, you'll always have your iphone with you. That means the LM will have access to location data, maps, chat - everything. And since it never leaves the phone, privacy-oriented people may be ok with it. And that means the LM can be exponentially more useful than just summarizing documents.
3) evaluation costs - they are the only competitor who will not have to pay for a massive datacenter, which means that the LMs can be as powerful as the M2+ hardware they sell. Everyone else will have no alternative to running the LMs on their centralized, expensive hardware.
What do people use it for? mostly just to set alarms or turn of lights etc.
With all this breakthrough in AI and things like Whisper which brings incredible voice recognition, these tools are ripe for an upgrade even though they have been stagnant for years.
ChatGPT like abilities + voice assistant and these 'voice assistants' are relevant again and actually deliver on there original promise.
Earlier quoted context omitted.
The inability to set multiple simultaneous timers is annoying. I compensate by setting multiple alarms. "Delete all my alarms" works when I'm done, but it's certainly a hack.
"hey siri, start a 10 minute timer for the oven" "hey siri, set a 30 minute timer for the rice." it works just fine for me, use it all the time.
>There's already a [number] minute timer. Replace? Confirm/Cancel
Running iOS 16.1, so unless they added this feature in the last few months I'm not sure why we'd see different results.
edit: Perhaps you're using HomePod? Apparently Siri on HomePod supports multiple timers, while Siri on iPhone does not. Goofy.
Earlier quoted context omitted.
Is this any different than Alexa or Google, though? I guess Google is probably better at looking up facts and Alexa at ordering stuff (if buying things without even seeing a picture is your thing), but neither is that useful. Is this not just a case of voice interfaces being much harder to do right than people originally hoped? (I say this as someone who has a Google Home that I use regularly, but mostly as a kitchen…
I mentioned an example in another comment - basic manipulation of timers don’t work well - like, adding time to an existing timer. When I’m working in the kitchen this is really useful. There are also major performance issues - “working on it” and “sorry can you say that again” are probably two of it’s most common utterances. Random bus like “there’s nothing playing” and the wrong siri taking over the request from a…
Personally, I'm become convinced a good talking/listening clock is actually a useful thing, but Google isn't providing it, and it sounds like Apple isn't either.
I don't use Siri all day, every day. It's certainly not perfect. But did you know you can say things like... "Turn off the Apple TV." "Turn off low power mode." "Stop navigating." "Open [name of app I don't want to spend time finding in my iOS folders]" "What song is this?" (Listens and usually is able to discover the name/artist of the song in the background.) So when people say things like, "Siri is mostly useless,…
My favorite is “[Let’s] go home”, which is what I use to navigate back to my house via Apple Maps. There’s also sending messages to people - you can send voice messages (“send a voice message to X”), which works decently well and avoids transcription failures.
It doesn't help that Apple doesn't let third party developers use it. The only SDK available for Homepod, is locked behind a bunch of approvals and the only thing it does is allow developers to implement music players. It's very specific to playing music. Not even Podcasts. Why can't an eCommerce app implement, "Hey Siri, tell Online Store to add Tidepods to my weekly order". ? What about Audible implementing, "Hey S…
https://developer.apple.com/design/human-interface-guideline...