The benefit of running the same chips as everyone else was about leveraging the investment of other products into the Intel architecture. Now they're doing that, just with their own chips.

As a long-time Mac user who's still basically happy with the product, this is what worries me the most long-term. YMMV, different strokes, go in peace with Ubuntu, but for me personally, macOS has been my favorite desktop Unix, hands-down, for about fifteen years. There's nothing that I want to do on even a semi-regular basis that I can't do, and a lot that would require, well, heavy adjusting if I moved to Linux. (Yes, I've used Linux within the last couple of years.)

I also love iOS and the iPad, but it's not a general computing platform; it's a computing appliance. It's very good at what it does, but by design it's difficult-to-impossible to do things on it that Apple doesn't want you to in ways that aren't true for macOS. (I'm sorry to those of you who had to disable SIP to recompile your own version of Apache, or who are infuriated you can't replace Finder with ratpoison, but you know that's not what I'm talking about here, right?) If "Project Marzipan" is about creating a new UIKit/AppKit hybrid that allows developers to create codebases that run on both iOS and macOS, that's already a little worrisome; if it's about "letting iOS apps run on macOS," as some of Gurman's reporting has it, that's a lot worrisome. I have a lot of apps that exist for both macOS and iOS, and in every single case, the macOS version is more capable. And iOS's "sandbox everything" model--and, I suspect, attitudes it engenders--make every app feel like an island not just in terms of data but in terms of functionality: there's much less of the "learn the basics of one app, learn them all" feeling that makes macOS, well, macOS.

If the Mac line moves to Apple's A-series chips, that's...not necessarily bad, but if it's being done in conjunction with sweeping software changes, it makes me extremely uneasy about the line's future in a way that even the Touch Bar doesn't (and trust me, I do not take the Touch Bar as a good sign). I'm not planning to switch platforms any time soon, but I'm starting to wonder if maybe I should buy an inexpensive Linux-compatible laptop so I can, you know, practice. Just in case.

In the long term this is a play to push Cook's iPad Pro post-pc idea, yet again...

They switched to Intel first and foremost because PowerPC was not going to get them the performance/power ratio that they wanted. It was holding them back in laptops. For example, Apple never shipped a laptop with a G5 chip because it took so much damn power to run a G5. Intel was basically their only option.

A secondary benefit, which they marketed heavily, was the opportunity to boot your machine natively into Windows. They thought this would help overcome the reluctance of "switchers" who were worried that they would miss something about Windows.

These days, the role of the OS as a gatekeeper is basically gone. There's greater diversity in client OS (Windows, macOS, Linux, ChromeOS, iOS, Android, etc.) and most anything important is available through a browser or a native app.

And Intel is no longer their only option. They've proven they can run big businesses making their own chips.

I do wonder how this will affect VM performance on Macs though. I don't see how it can be maintained, unless Apple chips are so much faster than Intel that they can absorb the translation overhead.

I guess that depends on what people think is "important". If by that you mean Facebook and Youtube, sure. But if we mean things like compilers, imho it would be a net loss if the only way to compile anything for any Apple OS was through Xcode/Swift.

That said there is no reason in 2020 they couldn't dual boot to MS's Windows ARM which is already shipping supports 32 bit x86 apps and should be polished enough to do legacy to the limited degree most users need.

I have to say this isn't how I expected Intel to go down. It is basically microcomputers killing minicomputers all over again (which was before my time, TBH).

I believe the main driver though, as pointed out in this CNET article https://www.cnet.com/news/four-years-later-why-did-apple-dro... was that IBM could not deliver on a powerful enough powerpc chip that would also meet the other constraints, namely, at this time, one of the biggest markets Apple had was the notebook market, and its sales there were exploding. IBM was unable to deliver a lower thermal envelope for its portable chipsets on top of performance issues.

I think this more than anything else pushed that reality.