The biggest difference between Giro and Tour has been the consistent ability of the dominant team(s) in the tour to castrate the race. the best teams in the tour are miles better than the best teams in other races and it helps controlling the race even more. GC action is almost impossible to be had when Poels, Henao, Nieve, Thomas and Landa are there to pull the leaders' group. Take the one sort of GC action that had an impact on the GC in the Bettex stage where Bardet attacked. Where did he attack ? in a descent where teammates' usefulness is a lot more limited than on flat or climb.
This thread is interesting, I disagree with some of the analysis, I think to some extent we have been focussing way too much on parcours and route for the last 10/15 years because it was the only lever of action that could really be freely experimented with. But actually other levers would be way more potent. In this I agree with OP : less riders per team (5 on one day races, 6 on stage races, 7 on GTs), no radio beyond radio tour (channel 1 local language, channel 2 english : safety info + info about time gaps and such, but all generic and not team specific), no SRM or rather blind SRM, data is recorded for later analysis but not accessible in race (riders keep telling you "I paced myself based on my SRM, I had practised this effort and knew I could keep x for y long"... this is not cycling anymore, this is formula one engineering), more limitations on what type of equipment is allowed (I am personnally against TT bikes, I think the frame should be the same model for a rider for a whole stage race, adjust wheels, saddle, handlebars, gearbox and such as you will, but keep the same frame).
All these would be very good to be experimented with, if only the UCI had just a tiny bit of courage.