I’m not sure if you read dcrainmaker’s post thoroughly, but this excerpt should give you pause:
I’m cautiously optimistic about accuracy potential on the new 2nd gen Shimano power meter. At no point during our conversation did Shimano shy away from the challenges and limitations they had trying to make the 1st gen unit ‘work’ on a crank arm that wasn’t designed for it.
In short, there are problems not so much with the power meter, but rather in how the drive side of R8000/91000 cranks flex when force is applied is not consistent, which makes crank based PM’s (of any brand) problematic on Shimano cranks.
I’m not sure that this weakness was known when Ray published his original 9100-P review in 2018, but thanks to the work of @gplama, this issue is known now.