I voted "it depends".
If I see a skyline which I like, I can visualize a result which is likely much wider than it is tall, so I expect to lose the top and/or bottom in the darkroom. Other times, I may be shooting 6x6 and see a rectangular composition, again I will lose the unwanted parts in the darkroom. Sometimes, I will choose a composition on 35mm film which will be more square than the 3:2 ratio, so I will lose the edges in the darkroom. Additionally, one can frequently consider the proportions of the size of paper one anticipates printing on, as a factor in choosing variations of some particular composition. When the subject is available, it makes sense to shoot alternatives, for different end products.
I think the vehement anti-crop idea came about because of a combination of the lower film qualities available historically and reaction to a "no composition - we'll sort it out later" style of shooting. The latter point would have been heavily stressed on most educational courses and has become a dogma of it's own. Even M.Cartier-Bresson didn't make every picture in 3:2 format - but I'm fairly sure he knew what he wanted before clicking the shutter, as should the rest of us.
There was a prototype digital camera in Japan which was made for placing on party tables. When it saw some programmed arrangement/quantity of faces it would focus and make a picture. Even that cost-less (per image) party camera wasn't simply random.