And now, back to specificity. One of the main findings of research on motor control and learning is that there is very little evidence for one skill actively interfering with another. This is so much the case that getting twice as strong in the squat doesn't translate into being able to jump twice at high, and those are two skills that share a lot in common (at least at the start).
OK one more thing to add since your post counters mine. I am talking about totally changing the coordination of which muscles and which structures manage load across the whole body-- so 2 independent strategies of movement. That's what I mean by the synaptic weights across the whole movement control system needing to be changed. As opposed to the interaction of 2 kinds of exercise (squats vs jumping) that utilize the same control strategy.
So within one movement type-regime, there may be little learning type interactions.. but across the types of regimes, there is a conflicting set of ideal synaptic weights, and we want to switch from one to the other.
Nothing ridiculous I think. Though it is still probably without any data to support it.