My 13 outperforms my 16 in every case I have minus an eGPU playing a game (and only get a few frames more -sometimes- on the eGPU.)Well, that'll always depend on the configuration.
I compiled Node in a very short time compared to the 16 and could not be happier. Joining the efforts to port software over and get virtualization working well (not too bad right now just early.)
No question the M1 is far superior and it will only get better. This "low-end" entry is miles ahead and will only be low-end compared to it's siblings going forward.
And no, the ports do not run lower speed, my drives are running faster on these ports than I had before. These are two thunderbolt ports not usb-c. Just like the 16, the ports are segregated in two buses. So the fact there are two vs four is only a matter of physical difference not bandwidth.
I've also observed memory handled far better and using far less in workloads I've been throwing at it. Non of those workloads are light either. Also I tested out my question on how unified memory would work with high-texture high-usage gaming. The tiled rendering does a far better job handling this and does not require the stupid amounts of memory to stream framebuffers to as with dGPUs.
Again, this is the first entry in the lineup, and it will support even more later.