whelmedjedi
macrumors regular
Lol. You told us it was coming with iOS 26.4 few weeks ago. Then 26.5. Then 27. And now its 27.2-27.3 already. Are you a joke? Hahahahahahahahhaha
Apple is and has been in over their head for some time now. Keeping Google services far away is but a dream. It’s too late, the Cupertino companies desperation is showing. Google is unstoppable.Please have an option to keep google services far far way from anything on my iPhone in settings. I am NOT intrigued not matter how revolutionary the tech is.
10-1 it’s gonna be Siri”beta” when they release it.
I’m also so excited to see what devices this new Siri won’t work on like my 8 HomePods, 5 Apple TV’s, 2 laptops, Mac mini, 3 iPads it will be so soothing to have a intermittent Siri experience depending on what device hears my request.
Guess it will be the same as usual if that’s the case.
Not interested even if it can make cappuccino? 😝🤣Please have an option to keep google services far far way from anything on my iPhone in settings. I am NOT intrigued not matter how revolutionary the tech is.
Yes but as I stated although the memory is unified the compute does not happen in memory with the current Apple Silicon framework. And even though there is zero copy of data and the CPU, the GPU and the NPU all have simultaneous access to the unified memory pool the model has to first be loaded into the memory pool from the SSD and THEN the CPU, GPU or NPU can compute the model residing temporarily in the memory pool. Eventually some of this compute will have to move to inside the memory pool itself and perhaps to a portion of the memory pool that is now made up of persistent memory that doesn’t flush the LLM upon booting down like RAM but is faster than the SSD.Apple's focus has been on software based optimizations. They have a paper called LLM in a flash. They have compressed existing models into very tight spaces and reduced latency. They can load 7b+ models into phone memory footprints.
And you're forgetting the GPU/CPU and TPU can already directly access the same memory objects, there's no copy need. And they put tensor cores in the GPU cores too in the M5. And their software abstracts the need to differentiate.
10-1 it’s gonna be Siri”beta” when they release it.
I’m also so excited to see what devices this new Siri won’t work on like my 8 HomePods, 5 Apple TV’s, 2 laptops, Mac mini, 3 iPads it will be so soothing to have a intermittent Siri experience depending on what device hears my request.
Guess it will be the same as usual if that’s the case.
It is much more than turning the lights on and off. Take a look at Perplexity's "Personal Computer" for Mac.Of all the top AI Gemini is at least better than ChatGPT but still garbage compared to Claude. At this point I'd rather just use my own LLM on a machine at home. Who needs AI anyway to turn your lights on and off. Will be disabled on my phone for sure
Running the models locally. Take a look at Google's Gemma 4 models.Apple partnering with a data collecting company to power a personal feature that is currently secure doesn't sound good to me. How is Apple going to protect the user?
Very likely does not know this. Just goes, “IT GAVE ME WORDS!” then copy/paste. Not even a cursory glance to see if it was even remotely accurate (like 26.4). 🙂You’re aware that it’s just drawing from publicly available information and adding some flourish?
Google and private don't belong in the same sentence.This is Google's version of Private Cloud Compute. It should give similar privacy protections.
![]()
Private AI Compute: our next step in building private and helpful AI
Introducing Private AI Compute, our new way to bring you helpful AI with the power of the cloud, while keeping your data private to you.blog.google
It's better than Apple doing it because they've been lying about it for 2 years. If Google is saying it's coming that means it's finally almost ready to deploy.How weird that google releases info on apples upcoming plans.
I wish people would stop saying what “Jobs would have done” as if they knew him on a personal level, and treating him as this infallible human being. He allowed the MobileMe launch, didn’t he? That was his decision, and while he chewed out the team who created it, he allowed it to launch in the disastrous state it was in.Honestly, this one will go on the history books.
The massive failure of Siri and the delays, ultimately resulting in Apple caving and doing business with direct competitor #1.
Jobs would've never allowed this. This was a clear mismanagement and lack of vision of Tim Cook and everyone involved with Siri. As an engineer and if I were to be an engineering manager at Apple, I would clean house entirely.
I have a few observations about that:Honestly, this one will go on the history books.
The massive failure of Siri and the delays, ultimately resulting in Apple caving and doing business with direct competitor #1.
Jobs would've never allowed this. This was a clear mismanagement and lack of vision of Tim Cook and everyone involved with Siri. As an engineer and if I were to be an engineering manager at Apple, I would clean house entirely.
Of all the top AI Gemini is at least better than ChatGPT but still garbage compared to Claude. At this point I'd rather just use my own LLM on a machine at home. Who needs AI anyway to turn your lights on and off. Will be disabled on my phone for sure
Apple has $66b cash on hand.4 trillion dollar company and they still cant it get right without another company bailing them out.
Gemini works good for me and my needs too.I use gemini, I'm fine with it. It works for my needs.
I have checked both of them too.I'm not a coder and don't )and am not allowed) use these tools professionally, but I've been trying out Gemini vs ChatGPT lately, and Gemini just keeps losing the plot every few exchanges. I was comparing Garmin watches for example, and Gemini kept inventing watches I hadn't tried, wasn't looking at, and didn't even mention. I'd correct it, it would apologize and it would be fine for a bit, but then they would come back or some other model.
I don't know a lot about AI bots, but at least ChatGPT has some short-term memory.
Well, Apple was then close to being dead, needed “partner investments”, had their back against the wall, and the Mac needed a well-working MS Office….The first thing Steve Jobs did when he returned to Apple in 1997 was create a strategic partnership with Microsoft.
like apple's wont either, its running off geminiThose services (Claude Dispatch, Personal Computer) run on servers and are very expensive.
Gemini has the worst Goldfish memory of any AI and it doesn't get better when you pay money. I've had it not only forgot long exchanges, but I've had it arbitrarily lose 95% of my chat with no recollection of it never existing. And by lose I mean it poofed out of existence. I scroll up and it looked like a brand new chat. Maybe with Gemini about to be on hundreds of millions of iPhones they'll finally address this issue, but I doubt it. ChatGPT lately is insanely negative and disagreeable (and more wrong on topics than ever).I'm not a coder and don't (and am not allowed) use these tools professionally, but I've been trying out Gemini vs ChatGPT lately, and Gemini just keeps losing the plot every few exchanges. I was comparing Garmin watches for example, and Gemini kept inventing watches I hadn't tried, wasn't looking at, and didn't even mention. I'd correct it, it would apologize and it would be fine for a bit, but then they would come back or some other model.
I don't know a lot about AI bots, but at least ChatGPT has some short-term memory.