I once had a discussion with some other OGS players in the English chat.
They argued that in order to achieve any certain strength in Go, a player would have to learn the skills of the ‘previous’ level(s) first to get a solid foundation. On this you would build the next level of skill and learn new things in addition, like a house or other structure that you construct one part at a time.
I stood with the philosophy that a local dan player sensei told me:
The strength of a Go player is heavily determined by the extent to which they have mastered the basics. When two professionals with equal reading ability play against each other, the one who wins often does so because of their more refined understanding of shape, direction, proper timing, all the usual stuff.
To expand on this, I have drawn a parallel to a diamond, or a painting, piece of music or other art. These are not things which you ‘construct’ (maybe initially, just a very little bit), but which are refined. You start out with a simple, crude core, e.g. some chord progressions, or a crude understanding of Go based only on the rules. Then you refine and polish that understanding until it becomes better and better.
I was pretty much alone with that standpoint. Maybe it didn’t help that the discussion took place in the context of the OGS rank feature. I was lamenting the removal of the custom rank setting and the introduction of the forced rank that you get to pick just once at account creation.
I said that there is always more learning value in games against stronger opponents than weaker ones.
This is really my core message. Trying to learn ‘the fundamentals’ from weaker players (or bots) is like trying to build a rocket engine from a faulty design.
The rank system part of the discussion continued in this very interesting forum topic.
So that is how I see the issue - in Go, you teach pro soldiering from the start and hope that most of it sticks the first time. 
Admittedly, this is not the majority opinion.