Hreb

Metal 3 may very well be great, but it does seem like a major (and perilous) assumption that any game developers will care to bring games to macos in the absence of native Vulkan support -- aside from the 3 games already announced of course.  Metal may be well established for mobile games but it is a remote island when it comes to AAA.

Another glaring omission is the complete lack of support for adaptive sync on external displays.  Apple is years behind in this department.

One thing that strikes me, perhaps cynically, is that if your workload is compute-heavy and at all amenable to distributed computing, it makes more sense to buy two M1 Max Mac Studios than to buy one with an M1 Ultra.  There is zero price advantage to scaling up from M1 Max to M1 Ultra given these configurations.  Cleary Apple doesn't think the "Studio" market is ready or interested in distributed workloads.

Free from the need to support a modular GPU there's not much need for getting very many watts of power in and out of your workstation, so Apple can repeat all the mistakes of the 2013 Mac Pro.

I have been using the 2019 MBP 16" for the past 6 months and it is by far the ugliest laptop I have ever owned.  Not the MBP itself necessarily, which is already way too heavy on its own, but the chunky Thunderbolt dock from Kensington, which besides being ugly was some $180.  And without the dock the laptop is worthless.  The M1 MBP is a step in the right direction, at least.

At these price points, consumers have a lot of choices.  If I needed a netbook with no extendability, I could have saved a lot of money and got a Chromebook.  That's just not the segment the macbook pro competes in.

So the M1 Max, announced yesterday, deployed in a laptop, has floating-point compute performance (but not any other metric) comparable to a 3 year old nvidia chipset or a 4 year old AMD chipset.  Somehow I don't think this comparison is going to be useful to anybody.