You greatly discount the lazy and the illiterate. I work in IT and it’s amazing that people would use this as a replacement for having to learn how to use an email client or even something as idiot proof as webmail.
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.
New: 'Apple Intelligence' AI Service for iOS 18 Will Have These Features
- Thread starter MacRumors
- Start date
- Sort by reaction score
ML cores have been in Apple silicon since like the A8 or A9 series chip. Apples machine learning has also been there as well. Those smart photo albums that know when it’s your birthday. That reminder that pops up when you’re arriving at your house. The location aware system that remembers where you parked your car- that’s all machine learning algorithms.
Part of it might be new ML cores. But it’s also RAM intensive to run ML algorithms on device. Especially generative machine learning. That’s why most “smart” devices like Siri and Alexa, etc. have traditionally sent those requests to the cloud.
On the contrary. I think it’s quite alarming how amazingly you can’t find articles about how to turn off “AI” features, chat bots, and integrations like “Copilot”. Everyone seems mesmerized by this and so excited to see if they can let it do everything.
I work in the IT space and I’m no Luddite. I don’t want a great deal of machine learning algorithm integration much beyond what Apple already uses. And it’s a lot! As in people really don’t realize how much machine learning has to do with features and services in Apple products. Especially iOS and iPad OS.
But LLM generative ML- I don’t want that. I don’t care how on device it is or how temporary and anonymous cloud computing requests are. An LLM generative ML bot like Microsoft’s Copilot is just one malicious attack away from being used as a precision tool for a threat actor to compromise your data across your device, run targeted algorithmic searches, shorten the payload to its minimum size, and squirt it off to an IP address to somewhere in the world.
To have it deeply integrated into an OS without an off switch is like connecting your car to the internet so the manufacturer can push updates down to it. Sounds good in practice. Brand new attack vector for criminals to steal and harm people.
And the average user is so ignorant about technology and even more so about information security. They get mad about complex passwords. The scoff if you enforce multifactor authentication.
But let’s unleash a complex set of algorithms deeply integrated into the OS that runs our computers without knowing that “AI” isn’t a thing yet and that machine learning has been around for a long time.
It’s also not virtue signaling to not be lazy. Autocorrect and spell check have made it such that even I have a hard time remembering how to spell basic words.
I think having a little baked in ML as an enhancement is fine. It makes better pictures in the camera app. It finds contact information in emails and text for my contact cards. It helps my Apple Watch track my fitness more accurately and it helps me to find where I parked my car, even if I didn’t pin it.
If you want the ML features, fine. If you want the generative ML features, great. But I’m not 5. I have a workflow. I make myself more efficient by having processes. And what’s more is that I haven’t forgotten how to craft a professional letter or a professional email.
Let’s not call it virtue signaling when the other side of the coin is to rush ignorantly forward at full blast for all the machine learning bots and their generative parlor tricks- consequences be damned.
While I agree with you and plan to turn off AI if I have the option, it’s clear we probably won’t have much choice after a couple of years. That’s because more and more apps and general functions will be written to require some level of AI in order to work properly. I’m sure the fanboys will love it.