I think the answer lies in the small plot in the article: Difficulty vs Complexity.
As the intrinsic complexity of the problem at hand increases, the (intrinsic) difficulty of using Excel just rockets after a certain point.
The more experienced R-user have a lower threshold of preferring R over Excel, since the (accidental) difficulty of using R is low enough for them to do even trivial stuff. It might be much smarter and faster to do the same thing in Excel for just about everyone else.
On the other hand, if people with no previous programming experience whatsoever keep building excel-workflows for larger and larger problems, it WILL turn into a catastrophically incomprehensible mess eventually, since they know no other alternatives than hitting the complexity wall.
(For myself - a fairly decent professional programmer, it requires a bit of swearing and stackoverflow to use R since although a solid core, a lot of syntax and conventions are decidedly non-cs... But so does Excel.)
As the intrinsic complexity of the problem at hand increases, the (intrinsic) difficulty of using Excel just rockets after a certain point.
The more experienced R-user have a lower threshold of preferring R over Excel, since the (accidental) difficulty of using R is low enough for them to do even trivial stuff. It might be much smarter and faster to do the same thing in Excel for just about everyone else.
On the other hand, if people with no previous programming experience whatsoever keep building excel-workflows for larger and larger problems, it WILL turn into a catastrophically incomprehensible mess eventually, since they know no other alternatives than hitting the complexity wall.
(For myself - a fairly decent professional programmer, it requires a bit of swearing and stackoverflow to use R since although a solid core, a lot of syntax and conventions are decidedly non-cs... But so does Excel.)