Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
I’ll provide you with an example. I input messy data from a branch manager’s email into ChatGPT, instructing it to sanitize and format the data in a manner that I can utilize. In a matter of seconds, ChatGPT successfully performs these tasks. It would have taken me significantly longer to accomplish these tasks manually.

Say it takes me about 5 minutes to manually clean up that data. If I have 10 branch manager emails, that’s already almost an hour spent. AI can automate this process, allowing me to allocate that 50 minutes to more productive tasks.
You organization allows you send your internal email to chatgpt? That’s breaking information security policy I can imagine.
 
image, text and handwriting recognition are some of the most useful AI features. I take photos of receipts (for business) and it is so freakin nice to be able to type a specific item and have it find the receipt with that item on it.

I’m willing to bet you use AI all the time, but just aren’t aware of it.
This is OCR and we’ve had it around for at least 25 years now. The fact it stores the OCR data with the photo doesn’t make it AI.
 
image, text and handwriting recognition are some of the most useful AI features. I take photos of receipts (for business) and it is so freakin nice to be able to type a specific item and have it find the receipt with that item on it.

I’m willing to bet you use AI all the time, but just aren’t aware of it.

I've been able to do that with receipts since the 90s and Windows 3.1.

OCR is not a new concept.
 
How good is the integration with Safari?

The use case I really care about is having to re-enter stuff like my name, address, phone number, known traveller ID, etc etc EVERY DAMN TIME I interact with an airline or hotel or whatever. The semantic web was supposed to do this for us but that crashed and burned (I still haven't seen a good post-mortem as to quite why). So for twenty years we've had this pathetic half-solution that can usually fill in your name or email address, can do your phone number about half the time, and fails at everything else, like your address (presumably, at least in part, because every damn web page formats and labels their forms differently?)

Apple Intelligence certainly, IN PRINCIPLE, can do this - has both the language knowledge to know what's asked for in each field, and the user knowledge to fill it in. But I don't know if they've yet hooked up this (obvious, but perhaps more complicated) use case.
You can already do that by creating text shortcuts. For example when I type ph my phone # appears, same for ad & my address. You can create any text shortcut you want say td for your traveler ID… 😉
 
It looks nice. I'm not that sure about privacy implications. Let's hope Apple keeps e2ee for the most part. Sending most of my pictures to OpenAI would make me concern.
None of this involves OpenAI.

The ChatGPT integration is entirely separate and will only come up when asking Siri for web knowledge or other requests that need a massive server to interpret. then you'll get an annoying popup asking if you want to use ChatGPT for this.

Even on the ChatGPT requests, Apple's contract with OpenAI prevents them from using your data for training or otherwise retaining any of the request sent to them. They have to throw it all away once the request is done. Apple lawyers likely put some teeth into that contract to ensure compliance. Apple doesn't want data leaks.
 
  • Like
Reactions: peneaux
Is there anybody out there who is completely disinterest in anything relates to AI?

Just cannot seem to get my head around why? its seems great on the surface, but to me its adding layers of complexity to things that are pretty simple to achieve with your own intelligence.
The only part of this that appeals to me is the ability to surface plans regardless of where I wrote them down (notes, calendar, email, etc).

As a neurodivergent person that feature will be extremely useful…but I’m not upgrading my 13 mini anytime soon.
 
I disagree that these features will make us dumber. "To" many people are already "to" dumb. Like, how does someone get through school without learning that it's "too", not "to"? 😇
I agree that many people are already no too bright, but, it will also erode creativity, writing abilities and the ability to research something. IQs won't drop, but abilities/skills will.
 
They must have an option to prevent someone phoning you and recording the conversation, right ? Or do they just expect you to hang up if you disagree ?
That's how it works. You are notified that the call is being recorded and you can either ask them to turn that off or hang up. It's also how it works on Zoom and other video conferencing systems.
 
It will erode creativity, writing abilities and the ability to research something. IQs won't drop, but abilities will.
I’m still shocked to this day how many office workers rely on Grammarly…and the nonprofessional tone people still email in anyway.
 
Thanks for the reply. You may be surprised to hear that there are laws against that here in the ummm EU...
Ask Zoom and Teams who are already doing this. You are notified of the recording and can either ask them to stop or hang up. Anything else becomes too cumbersome to use. This is more like real world. If you walk up to someone and start talking and they say, just a minute while I record this conversation, you can say "stop" or you can walk away.
 
Running the beta, it’s really difficult to justify use cases for any of the AI. Everything is an exercise in finding a solution to a problem that’s not really a problem.

That seems to be the reflexive reaction here to many Apple product introductions.

iPod -> Who asked for a thousand songs in your pocket? (commenting on Jobs' iPod introduction)

iPhone -> Nobody will use a mobile phone with a non-mechanical keyboard.

iPad -> It's just a large iPhone!

Watch -> Who needs it? If I need to know the time I'll just look at my phone!

With respect to the AI features (that just scratched the surface) covered in Dan's video up above... Anyone who does research, processes information, reads (professionally/pleasure or as a student), writes papers/documents/summaries/books/etc. can benefit from the initial set of features covered in the story's video.

What was covered in the video will certainly be useful for me.
 
When writing an email by the time I load an AI, get the correct prompts, get the result and review it, I've already written the email without help.

Doesn't matter for my job as AI is 100% banned for employees. I work in healthcare, and for security and HIPAA reasons, we can't use it. Couldn't try if I wanted to as chatGPT, copilot, gemini and the like are domain-blocked on our network. I suspect with will be MDM-blocked on or company phones, too.
Normally, if I am writing an email of importance, I will write it, run it through AI, and say something like, “Make this more professional” or “Make this more friendly,” etc.
 
I’m still shocked to this day how many office workers rely on Grammarly…and the nonprofessional tone people still email in anyway.
Guessing you’ve ever used grammar check or spellcheck in Word, Grammarly, and other writing tools are just an upgrade to those.
 
  • Like
Reactions: Tagbert
I’ll provide you with an example. I input messy data from a branch manager’s email into ChatGPT, instructing it to sanitize and format the data in a manner that I can utilize. In a matter of seconds, ChatGPT successfully performs these tasks. It would have taken me significantly longer to accomplish these tasks manually.

Say it takes me about 5 minutes to manually clean up that data. If I have 10 branch manager emails, that’s already almost an hour spent. AI can automate this process, allowing me to allocate that 50 minutes to more productive tasks.
Not a use case for me, because it would be a breach of data protection to expose company communication and data to OpenAI. It would be nice to have a local LLM for stuff like that, though.
 
Not a use case for me, because it would be a breach of data protection to expose company communication and data to OpenAI. It would be nice to have a local LLM for stuff like that, though.

Just like this shoe doesn’t fit everyone, it’s important to remember that there’s no one-size-fits-all solution.

Interestingly, Apple Intelligence's writing tools work offline. I used it to re-write the above to be "friendlier" while in airplane mode.
 
Apple Intelligence is limited to developers at the current time, with Apple working to refine the Apple Intelligence features and eliminate bugs. A beta version of Apple Intelligence will be available for all users later this fall, after iOS 18, iPadOS 18, and macOS Sequoia launch. You will need an iPhone 15 Pro/Pro Max or later or an Apple silicon Mac or iPad to use Apple Intelligence due to the processor requirements.
The only problem with doing a Youtube video on this is that the demo is only on iOS, looking at MacOS I see writing tools greyed out. I wonder how equivalent the implementation is across iOS/iPadOS 18.1 and MacOS 15.1? So yes prefer editors discussing Apple Intelligence across all the OS's.
 
Last edited:
You organization allows you send your internal email to chatgpt? That’s breaking information security policy I can imagine.
Sure the current problem would be that for some. But organization-approved AI, either running locally or through a trusted provider will exist and bring benefits which I believe was the point of the post (AI tools being useful).
 
  • Like
Reactions: Tagbert
Writing articles, cleaning up my grammar and errors in my writing, spending less time looking for stock photos that are really royally free, and just creating images, etc.

Edit: I'm a member of a few non-profits. I’ve been able to improve our marking items with AI, this includes websites, flyers, posters, ads, etc.
If you couldn't do those thing prior to AI, should you really have been in that position to begin with?
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.