Not only is their speech-to-text of no use, but Siri doesn't work when it is asked. The AI in Apple phones is not getting better with the release of new software updates.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple's ChatGPT Rival Moves Forward, But Siri's Future Still Uncertain
- Thread starter MacRumors
- Start date
- Sort by reaction score
the good thing is: it can only get better!! i am waiting for improvements too, even still think about switching to Android first time of my life, because Apple Software sucks so much these days!
anyone else have the bug when hitting the dictation icon, it doesn't actually work, I have to press it maybe 2 times for it to kick in within any app.
Mine sometimes does this ... or gets stuck with the button activated but not actually dictating anything I say.
I use dictation all the time and it's amazing how poor the accuracy can be at times.
I really thought it would have been so much better after all these years.
Right, always the negativity, the snark, the denigration, always too late, always behind, always substandard. And when Siri finally does make it to the level of the big boys in AI it will still be dismissed as irrelevant. This is life in the tech blog universe where Apple is concerned.
Apple is learning the hard way that it takes unlimited and, seemingly as of now unaccountable by Wall Street, VC money to even come up with the 300 billion + models that hallucinate now much less their paltry 150 billion parameter model.
And people kavetching about “Apple should have scrapped Siri and started over sooner”….with what and by doing what that was fundamentally different from all the other A.I. companies ? Apple is not and has never been an A.I. company. What makes any of you think that Apple can suddenly turn into a single purpose company like OpenAI and do any better than OpenAI that’s been developing A.I. from the jump ?
Some commenters say Apple should just buy <insert AI company here>. Well…why not ? That’s how we got Siri. Just look at the Wiki on Apple buyouts of companies over the last 30 years. For crying out Apple BOUGHT STEVE JOBS himself along with his operating system NeXT which became the very operating system we all use now as owners of Apple tech.
Apple is never going to develop an AI that meets their hallucinating goals. OpenAI can’t. No one can at this point. Apple is either A: going to have to scrap the whole internal LLM project and buy <insert AI company here > or B: Live with a certain amount of AI slop and hallucinations and just try to make the on-device small models ( 7 billion or less parameters ) more useful and integrative than Siri.
And people kavetching about “Apple should have scrapped Siri and started over sooner”….with what and by doing what that was fundamentally different from all the other A.I. companies ? Apple is not and has never been an A.I. company. What makes any of you think that Apple can suddenly turn into a single purpose company like OpenAI and do any better than OpenAI that’s been developing A.I. from the jump ?
Some commenters say Apple should just buy <insert AI company here>. Well…why not ? That’s how we got Siri. Just look at the Wiki on Apple buyouts of companies over the last 30 years. For crying out Apple BOUGHT STEVE JOBS himself along with his operating system NeXT which became the very operating system we all use now as owners of Apple tech.
Apple is never going to develop an AI that meets their hallucinating goals. OpenAI can’t. No one can at this point. Apple is either A: going to have to scrap the whole internal LLM project and buy <insert AI company here > or B: Live with a certain amount of AI slop and hallucinations and just try to make the on-device small models ( 7 billion or less parameters ) more useful and integrative than Siri.
Last edited:
After using chatgpt a lot its a bit tough because it says things with 100% confidence and about 50% of the time its completely wrong and often it contradicts itself
And I would argue that Apple should not go all in on integrating a feature with that track record.
That's not what I want out of an Apple device.
What about Grok? It is the best!
yeah that's exactly the issue I have
This is the fundamental problem really. Nobody has solved this, Apple isn't going to solve this. Other companies are content just to stick a warning on it, but that's not really the Apple way is it?
Absolutely, if Apple can do what no other LLM company has been able to do, they'll have a winner, but what in Apple's track record of software and services suggests they have even a tiny chance of doing that?
This isn't quite accurate.
Apple has been an AI company for a long, long time.
AI is much, much more than Large Language Models. Much more.
Apple has a dedicated Machine Learning (ML) developer kit. They've had built in processors (neural engine) in their chips since at least the A11 bionic (first for iPhones/iPads, later for all Macs). The Photos app has had several ML enhancements for a long time. There are other examples, and among them: They - almost certainly - had a lot of AI engineers working on the self driving functions for project Titan.
What they've NOT been, so far, is an LLM company.
and its not learning at all. didnt they tell something like machine learning on device many years ago? I am correction some totally wrong understood words on a daily basis. its so useless dumb.
When did Apple decide to push privacy and 'on device' so hard?
It wasn't _always_ an Apple thing. And they're not even 100% all about privacy but do push it hard as a marketing thing and now it's become a constraint in the face of AI requiring cloud processing and learning.
Have they backed themselves into a corner?
It wasn't _always_ an Apple thing. And they're not even 100% all about privacy but do push it hard as a marketing thing and now it's become a constraint in the face of AI requiring cloud processing and learning.
Have they backed themselves into a corner?
can be, but they ****ed up now. the software is buggy and speech to text isnt even working how it should. Its a shame
Yep, it used to be that no one could refuse Apple because they were the undisputed tech leader, nowadays Apple is just another player with lots of money and no common sense. Nobody wants to be part of that. Nobody cares about Apple now that Jobs and his magic are gone.
The thing is that they did it to themselves. Siri used to already be much more conversational and useful than it is now. I remember articles about autistic kids conversing with Siri back in ~2010. Now it’s just a waste of memory on Apple’s various platforms where they won’t let you remove it.
Glad to see the common sentiment to drop the Siri brand entirely. This is perhaps more likely now that they are (supposedly) FINALLY scrapping all the old code for Siri and starting something completely new. They should have done this literally 10 years ago.
I remember writing here and elsewhere long before ChatGPT 3.0 splashed on the scene that Apple was dangerously losing ground with Siri's failures, as compared to Google and Alexa. I wasn't predicting AI or LLMs (I had no idea what LLMs were at that time). I was merely pointing out that the reason for Apple's existence more than anything was "it just works" and that was no longer true, especially in comparison to how well "Hey Google" worked in comparison to "Hey Siri." It was really a joke (still is). Apple seemed oblivious to this, or its importance. It is like they viewed Siri as a sideshow, and it was not. Watching interviews with Google's people, and watching parts of I/O this year, Google seems a decade ahead of where Apple is today (probably good marketing on Google's part), but maybe 5 years is more conservative, and they are pulling away fast.
Google knows search is toast, and soon. They are getting ready. The new paradigm is going to be that your devices will be able to actively predict the kinds of things you want from them--true "interactivity" meaning your devices will talk back to you, not just respond to you. They will be watching what you do, and what the world is doing, and create a bridge between you and the world. Right now, all devices are merely responsive, and that's it. You tell it what you want, how to find it, and then it tries to find it for you. In 5 years (or sooner), devices will be predictive, they will know the world, your hard drive, your surroundings, your daily routines, your relationships, and what's important to you, and the people in your life, and it will also already "know" vast external sources of data, ready to hand the most salient parts of that data to you in real time, interactively. Apple seems to have missed the importance of this until last year, and maybe still doesn't really see it. I hope I am wrong about Apple, or it will go the way of landline phones, Blackberry's phones, Microsoft's phones and c: prompts.
I remember writing here and elsewhere long before ChatGPT 3.0 splashed on the scene that Apple was dangerously losing ground with Siri's failures, as compared to Google and Alexa. I wasn't predicting AI or LLMs (I had no idea what LLMs were at that time). I was merely pointing out that the reason for Apple's existence more than anything was "it just works" and that was no longer true, especially in comparison to how well "Hey Google" worked in comparison to "Hey Siri." It was really a joke (still is). Apple seemed oblivious to this, or its importance. It is like they viewed Siri as a sideshow, and it was not. Watching interviews with Google's people, and watching parts of I/O this year, Google seems a decade ahead of where Apple is today (probably good marketing on Google's part), but maybe 5 years is more conservative, and they are pulling away fast.
Google knows search is toast, and soon. They are getting ready. The new paradigm is going to be that your devices will be able to actively predict the kinds of things you want from them--true "interactivity" meaning your devices will talk back to you, not just respond to you. They will be watching what you do, and what the world is doing, and create a bridge between you and the world. Right now, all devices are merely responsive, and that's it. You tell it what you want, how to find it, and then it tries to find it for you. In 5 years (or sooner), devices will be predictive, they will know the world, your hard drive, your surroundings, your daily routines, your relationships, and what's important to you, and the people in your life, and it will also already "know" vast external sources of data, ready to hand the most salient parts of that data to you in real time, interactively. Apple seems to have missed the importance of this until last year, and maybe still doesn't really see it. I hope I am wrong about Apple, or it will go the way of landline phones, Blackberry's phones, Microsoft's phones and c: prompts.
Heck, Apple's spelling and grammar checkers along with on-device search are pretty much below everyone else (i.e. terrible). If Apple cannot get the basics correct, why expect them to get any of the more complicated stuff correct.
me too. another annoying iOS and macOS Bug not getting fixed.
Or when you talk (on macOS) it writes what you talked 100 times in the text field. HOW do they not even fix these ongoing bugs. Apple is lazy, slow and has no QA anymore.
I’m talking about OS-integrated, future functions, like what Apple was hinting at with on-screen awareness and personal context. I’m specifically thinking of reasoning models having access to what is happening on your screen, and to your personal data. Apple won’t be opening that up to external providers. But you can be sure that Google and others are working on functionality along those lines, while Apple is playing catch-up just getting a basic ChatGPT equivalent to work.
“Apple has run benchmarks on the model that suggest it ‘approaches the quality of recent ChatGPT rollouts.’ However, there are still said to be concerns over its tendency to hallucinate”… that it approaches the quality of recent ChatGPT rollouts.
Last edited:
Just drop the AI nonsense altogether, IMO. Every LLM is supposedly getting better on metrics, but their real world performance, in my experience, is getting worse. It’s almost like how CPU and GPU makers optimize for benchmarking software instead of actual innovation that might not result in “number bigger now — bigger number good”.
I don’t think I have had a single LLM interaction in the past 2 months where I haven’t had to ask it if I was unclear (“No, you were very clear.”), ask it repeatedly to double-check and verify it’s clearly long answer, been gaslit on answers I know are wrong, have been told by the LLM to do a search for the answer, etc, etc.