While Charlie has the reputation he has, in this case, lets not forget, 10nm was originally a 2016 product. Charlie started telling us 10nm was having problems in 2015.
Nobody believed him.
2016 came, and no Cannon Lake, it was pushed back till 2017.
Charlie started saying, still problems with 10nm, don't expect Cannon Lake in volume in 2017, which indeed didn't happen.
at CES this year Intel told us "ramp up of 10nm late in the year" Charlie said bull. Come April, Intel pushes back 10nm mass availability to early 2019, Charlie again calls bull, July Intel says "they meant 2h 2019"
Frankly despite his bad reputation Charlie has been correct on the 10nm issue every step of the way, and Intel has actually lied multiple times (SEC, you should probably get on that, they probably lied to investors too)
We are at the point that yes I will believe Charlie over anything Intel has said till they actually have more then neutered 2 core sans IGP 10nm parts out the door.
Regarding the reality of 10nm being dead, I think the original 10nm is deader then a door nail, the "new" 10nm is going to basically be a lot closer to how the foundries are doing their "7nm" using pretty much similar techniques and machines.