[xondono]

When presented with this kind of “simple advice”, I always like to think the counter examples.

Quantity is great, unless you are trying to change a paradigm. Iterative learning gets you trapped in local minima, so be sure to aim for quality once in a while.

I would also not take that “experiment” that seriously, because the requirements where obviously different, each class optimized as needed. “Quality” is an abstract concept, and most of the time you are better of saving the time than spending it on “Quality”.