mjtomlin

Got my first Mac in 1987, a Mac SE, and the damned thing still runs. Bought my latest iMac in 2009, and it still runs, although the display is slowly dying. My current main system is the base model of the M1 mini and it’s still going strong - but I will probably upgrade to an M4 Pro mini, when it comes out.

Apple’s gonna want to release some new Macs alongside the next version of macOS so Apple can put Apple Intelligence in front of some Mac users. I think we’ll see the desktops updated much sooner than people think; new mini and studio models will arrive this Fall. laptops in the Spring. iMac, new iMac Pro and Mac Pro at WWDC


Update for clarification:
What I meant by that, is that Apple is going to want to push the whole “AI” thing further with the release of M4 Macs, which, I can only assume, will have vastly superior “AI” capabilities, especially with the Max and Ultra variants. As it is now, the base M4 has a more performant NPU (38TFLOPS) compared to the M2 Ultra (32TFLOPS)

appleinsideruser said:

So, hang on. Isn’t this like the ChatGPT Mac app storing conversations in plain text, which was a privacy breach recently?

Umm, nope…

After verifying with Face ID

Not sure how you missed that in the article?

The data is always kept encrypted on device. You have to authorize the system to generate the report, which is just plain text.

CrossPlatformFrogger said:

While Apple kept pushing for unnecessary unused horsepower in their chips Google chose to focus on AI with the Tensor chip, and Apple was caught flat footed. I remember when Google first announced Tensor and everybody talked about how it wasn't as fast as Apple's chip and Google came out and said they weren't going for speed. This article makes me want to check out the new Pixel, I know the "AI" on it will be the real deal.

This article has nothing to do with the Tensor SoC’s in the Pixel. Those SoC’s aren’t even close to Apple’s M-Series in any respect - maybe the NPU is close.

This is about Google’s server Tensor Processing Unit, which is an entirely different beast….

Is there any doubt Apple didn’t use 3rd party hardware early in their LLM R&D? Of course they did. Apple didn’t have the specialized chips or servers to do this work. I’m sure they used this time to better understand how to build and optimize their own chips/servers that would benefit their own LLM plans in the future - Apple has already stated their “Apple Intelligence” server farm will be running on Apple silicon.