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.
iOS 18 Apple Intelligence Feature Will Have Waitlist During Initial 'Limited Preview'
- Thread starter MacRumors
- Start date
- Sort by reaction score
I love WWDC... always brings the haters out for a day or two and fuels them up.
Then someone actually uses the new software and discovers things not announced and the "ah cool" comments start appearing...
Bring on Wednesday...
Then someone actually uses the new software and discovers things not announced and the "ah cool" comments start appearing...
Bring on Wednesday...
Probably if they gate it aggressively. Limiting in this manner in addition to by device gives them a better ramp-up and -I’m sure they hope- eliminates any nasty surprises. Of all the things shown today I’m more interested in their infrastructure for this. Hardware stats etc. I figure all this will be working good a couple years out once they build to scale.
Do you remember all that CSAM fuss. Around the time all these AI companies were scraping the **** out of your privacy....
For all it's hoopla about privacy, some of the companies it sleeps with are proof they don't give a ****.
It's safe to judge a person by the company they keep.
For all it's hoopla about privacy, some of the companies it sleeps with are proof they don't give a ****.
It's safe to judge a person by the company they keep.
It’s a priority when it aligns with their business interests simple as. Sometimes that is aligned with the user and other times not. I tend not to take too cynical an approach since they are a business answering to investors. Now the investment model of growth and profit or else? Yeah, terrible system
I thought the Apple Intelligence stuff was on device and everything else is being sent to new Siri or ChatGPT? So confused by what is powered by what and where it goes now!
Yeah the “Private Cloud Compute” part of this is (forgive the pun) a bit nebulous. I’m sure that is the part that is taking some of the most effort and not fully flushed out. There are some big and significant promises being made about privacy in a field that’s been operating under the modus of “screw privacy and go bigger/faster” for awhile now.
I am glad Apple is being Apple about this and scrutinizing every detail instead of just nudging the wagon over the top of the mountain and waiting to see what makes it to the bottom. On the other hand their competitors do not seem to have the same restraint and are willing to throw entire product lines and platforms at the wall to see what sticks along with how much privacy they are willing to give up for convenience.
It’s obvious Apple is completely unprepared. The keynote was vague and missed the usual depth and details. They simply announced what they could scramble together.
This stuff will definitely not be ready this fall. Therefore the waitlist, the beta label, the mentioning it will be rolled out gradually and the very disappointing English Only announcement
The ‘97 relationship with Microsoft continues! (Who own 49pc of OpenAI, apparently)
As far as I understood, GPT is opt-in and each time you want to use it Siri will ask smth like “may I use Chat GPT to process your request?”.
I then wonder what did they actally upgrade in Siri except being able to answer calls with voice🤣
Cook's new innovation, "You have to pay for it before we develop it."
Apple is just plain not good at developing initial cloud services. So I suspect it will take, based on previous Apple history, probably 3 or 4 API rewrites to get it right.
The reason they are not good is not because of the talent, but the push for something or anything, no matter how brain dead, ready for the keynote.
No one’s hating. They are so far behind the curve. They couldn’t even put the ai features with the beta. Making us wait proves it.
I spent a few years wanting to use just Apple Maps.
Google maps is still much better in every regard.
OpenAI had an outage just six days ago. How is that not accurate? Once you ask for more complex queries that need a larger LLM (in this case ChatGPT) it ll need to connect to OpenAI servers. They didn't license the Model to Apple and doesn't run on Apple servers....
Depending on the complexity of the query will have to reach ChatGPT. Usually on device-related enquiries will either be handled by the on-device 3B variable defusion model or Apple's server-side foundation LLM. The state of the Union video gives a very in-depth overview of Apple Intellegence