Marco Arment

Like the next process shrink with the chiplet architecture frees them up to leave behind the ultra, the, what has been the ultra architecture, which is taking two maxes and sticking them together.

Because I think that is a cost savings and it's been cool, but it has essentially run its course and,

it was never as good as it could have been, but it was a reasonable compromise to get something out better.

And by the way, one of the other big advantages of the ultra has, which you see in some of these reviews is, and we don't talk about that much because it's not in the world we're in, but it has twice as many media units and lots of people who do like these, uh, you know, benchmarks of like, Oh, I'm escorting from final cut.

And like, huh,

No matter which M chip I use, it exports from Final Cut at like the same speed or like the Pro versus the Max.

Why is it the same speed?

Because that's all happening on the stupid media engine.

And the only way to get it to go faster is get the Ultra because it's got double the amount of that stuff because it's two Maxes stuck together.

And lo and behold, it goes twice as fast.

Again, if you built a dedicated high-end chip, maybe you'd put four media engines.

You know what I mean?

Like, you'd make different decisions.

But to your point, it's like, are they ever going to... Is it ever going to be economically feasible to make anything custom for the Mac Studio?

Or is even the Mac Studio stuck with...

You just get the leftovers from what the laptops use.

And maybe we can stick two Maxes together and you should be happy with what you get, happy with your non-linear scaling, whatever.

It's better than a Max.

It is.

And people will pay for it if you want something better than a Max.