They are impressive-looking parts. Quite an errata on the N6, which might explain why it spent years stuck in limbo. And that doesn't include the various accelerators, which may also have or have had issues, but we wouldn't know because they are secret peripherals.
Personally, I would be more interested in an M4 -> M55 upgrade for the flagship low-power series (U5), as that is more of a lateral move in processor. Basically, give me Helium with all the other low-power goodness. The neural accelerator is interesting, but only to the extent that it can be abused. If it can only be operated with vendor tools in ways that the vendor intends, then as far as I'm concerned it's useless silicon that I'm forced to pay for. But maybe that's just me.
If I'm looking for high performance (w/o regard for power consumption) then I want the M7 -> M85 upgrade. This is rumored to be on its way, but sounds highly ambitious (18 nm process, phase-change memory), so I don't expect it any time soon.
I do hope the N6 is a runaway hit for ST so it can help fund development of the parts I actually want to use

Or, perhaps we will have compelling RISC-V MCUs (in segments other than "value") by then.