3-5 Evaggelikis Scholis, 14231 Nea Ionia, Greece

Apple Intelligence Just Made All Other AI Look Pointless

Originally posted on lifewire.

But it hasn’t solved the environmental nightmare

Apple Intelligence finally bring us the sci-fi promise of computer assistants from TV shows and movies.

So far, the AI hype has been all about generative junk, Chat GPT and Google spewing made-up nonsense cobbled together from unauthorized data scraped from across the internet. And while Apple did debase itself by allowing some of that into its WWDC 2024 keynote, it also showed the tech industry what the future of AI really looks like, just as it showed the world what laptops, smartphones, and tablets would look like. It’s far from perfect, but it makes all attempts thus far look a little pointless.

“Yes, Apple is going all-in on AI features. But that’s the thing: they are FEATURES,” Ryne Hager, of privacy-forward cellphone company OSOM, said on social media. “AI isn’t a product in itself. This is a watershed moment in the same sort of way that data analytics were, and we are gonna see AI happen to every single thing from here on out. But you aren’t gonna buy an AI box or pin or any other magically wishful novel product category.”

Apple Intelligence

What Apple has done is put the large language models (LLMs) that power generative AI inside of apps. The most GPT-like features are those where it offers to proofread or rewrite emails for you, but those are a distraction from the main event, which is that Apple’s AI has deep access to your personal data, and uses that to become a really, really useful virtual assistant.

To begin with, it does helpful things like summarize emails, and shows those summaries in your inbox, instead of just the usually-useless first line of the mail. Or it can parse notifications, showing you the ones you need to know. Or it can learn your handwriting style, and then add pasted text to your notes, in your own handwriting.

But what really impresses is Siri. Your phone already knows who you are, where you live, who your friends and family are, what they look like, and more. It has your calendar, you address book, your emails, photos, and messages. So, instead of offering pizza recipes with glue, Siri can now a) actually understand your questions, and b) answer them, drawing on all that personal data on your iPhone.

So what can you use this for? One great example from the Apple keynote was asking Siri what time your mother’s flight lands. Siri grabs that info from your email or messages, and then checks in with flight-tracking services and gives you the answer.

“[I]f you can integrate personal information, then this takes generative AI to the next level and it becomes much more valuable,” Muddu Sudhakar, co-founder and CEO of generative AI company Aisera, told Lifewire via email. “You do not have to write many prompts and tweak them. The LLM will generally understand what you want […] and carry out the actions. It makes for a much better experience, such as when searching photos or planning a trip.”


This seems basic, but it’s exactly what sci-fi computers do in the movies, and it finally frees us from the busywork of searching for and correlating information. And because it’s all based on your own data, it should hopefully be accurate and useful.

“The concept of system-wide access to personal information is set to revolutionize our interaction with devices. By breaking down the silos that currently limit their functionality, devices will not just serve as a starting point for connecting other tools but will actively create efficiency on our behalf, sometimes even in spite of us,” Kathleen Perley, generative AI professor at Jones Graduate School of Business, told Lifewire via email.

Another example we like is from Apple’s press release, where you ask Siri to “Play that podcast that [your friend] recommended,” and it works it all out for you. Or what you husband asked you to get from the store, and so on.

But AI is still AI. While Apple has solved the privacy angle by doing all of the above locally on your device, it still has to train those models, which takes a lot of electricity, and it still kicks some queries out to the cloud. AI clouds are both less private and use reckless amounts of power and water, not to mention the ethical concerns around Apple’s new AI partner Open AI.

“I really wish they would have addressed the environmental story. That would help me personally accept these features more. But I know they can’t… because the story isn’t good,” veteran Mac and iOS developer Cabel Sasser said in a post on Mastodon.

Given that Apple is so hot on pushing its green credentials and commitment to renewable energy, it’s no wonder it didn’t mention this side of things in its keynote.

“Genuinely, if you are rolling out generative AI features, and not talking about the ethical or environmental implications of them, you have entirely failed to change the industry in a way that only Apple could,” ex-Apple software engineer and longtime Mac and iOS developer James Thompson said in a post on Mastodon.

Apple might be showing the world how to use AI, but it missed out on the most important part.

Source: lifewire

Related Posts