> And oh yes, the modern Mac shares those 8GB with the video RAM...
The Unified Memory Architecture is why these Macs are so fast—no wasted cycles moving data between RAM and GPU. And the data is compressed in real-time so less data has to be transferred and there's less ware and tear on the SSD, which is directly to SoC [1].
UMAs aren't made for speed, but for power savings. You are ignoring the fact that a discrete GPU accesses VRAM and caches at much higher bandwidths (and power) than an iGPU does RAM. Shared mem also comes at the cost of keeping it coherent between CPU/GPU. So you can't just look at one part of the system and then claim that UMAs must be faster because there are no data transfers.
And by the way, even on UMAs, the iGPU can still have a dedicated segment of memory not readable by the CPU. Therefore UMA does not imply there won't be data transfers.
That's not really a thing with Apple Silicon. The A series chips and the M series have the same CPU and GPU core designs.
Because you don't need to support Thunderbolt 4/5 controllers, PCIe lanes for NVMe storage, ProRes encode/decode engines (on Pro/Max/Ultra tiers) and multiple external displays in a device like a phone, Apple TV, or a HomePod these features are absent from A series chips.
The A17 Pro corresponds to the M3, the A18/A18 Pro corresponds to the M4 and the A19/A19 Pro corresponds to the M5. Same core design, different implementations.
It's not like Intel where there are many server processors, desktop processors and mobile processors. Apple uses the same core design they scale up or down as needed, for example the S series chips in the Apple Watch. The S9 is a scaled down A13 or A15.