|
Post by CdnPhoto on Jun 7, 2024 3:12:01 GMT -8
|
|
|
Post by CdnPhoto on Jun 7, 2024 4:03:09 GMT -8
|
|
4aapl
Moderator
Posts: 3,845
|
Post by 4aapl on Jun 7, 2024 6:01:52 GMT -8
We're still so early of stages that Apple can define what AI is. Many here ask "what is AI", or "how do you define AI"? My current thought would be on something where the processor seems to do thinking, instead of a set task. So something like "clean up this image" might brighten or darken it, reframe it, tweak the colors, and remove some seagulls or people in the background. All of those are doable, but it's a complex enough task that someone might consider it AI, especially if there is leeway on it. Whereas just throwing one filter at it, like a fisheye lens, is just computing power doing a set set of things.
I don't know the process behind it, but I'd probably put the "photo set" things as AI, where it combs through your photos and finds a group of them. Some of these aren't necessarily magical, using location data for an overseas trip, though it then has to decide which photos to use. And maybe the computing behind "winter trip" isn't that exciting, knowing what the light characteristics are of snow (our Nikon 995 from 20 years ago had a snow setting), or what a ski or snowboard looks like, or even just ski goggles. I haven't looked closely at the picture sets, but there is something magical about putting some of those together, while also managing to create value with our own stuff that for many people doesn't get looked.
There is so much current thought about generative AI just being to make pictures or an essay/article. And that might be useful to some people, but it is also a way to create something that doesn't necessarily have a defined correct answer, and so there is some room for the product to be a little different than what a person might make, because different people would likely make different products. This isn't "the answer is 42", but rather "here is a vase". How do you define what is good or not? "I'll know it when I see it".
I think that is the tough and yet malleable spot that Apple is in. They need to make some cool stuff that "AI" can do that ideally is actually useful instead of just cool, label it to the masses as innovation, and ideally have it not easily duplicated while possibly also showing off some of Apple's advantages. It's a combination of HW, SW, and marketing. And that is a place where Apple has an advantage, though they still need to focus on some great things.
We'll see what Apple has next week. It might just be baby steps, and if so I don't think the market will like that. Maybe it will be some big things. Or most probably, it will be like the Vision Pro of showing off some cool things (baby steps), while opening up some options to developers (the point of WWDC), and possibly even a showcase some 3rd party things that are already in the works after "only working a few weeks" on it.
But we're still at that "defining moment" in AI, while also still defining what we are after.
|
|
mark
fire starter
Posts: 1,624
|
Post by mark on Jun 7, 2024 6:07:13 GMT -8
"Apple Intelligence" is an ABSOLUTELY terrible idea. Really, really, really bad.
That's because every misstep, and there will inevitably be missteps, will be directly associated with the very valuable Apple brand name. Imagine if Siri were named Apple instead!
Really bad choice.
|
|
|
Post by zzmac on Jun 7, 2024 6:49:15 GMT -8
Apple hopefully has a lot of great things planned to incorporate AI into there products but I think it's key that they make SIRI a hundred times smarter if they want to sell an absolute boatload of new phones.
|
|