As someone on here mentioned, I agree also that ASO shouldn't be going out of their way to pave roads for new climbs. There are plenty of good roads to use as it is.
The problem with the recent routes, is that in total, there is probably more climbing now than there was in the 'good old days', but a lack of a genuine queen stage/s. If you put a route together that has 12 stages where GC action could happen, then the GC riders will choose not to race on half of those stages anyway.
A big improvement in recent routes seems to be the greater inclusion of the medium mountain ranges of France; the Jura, Vosges and Massif Central. I don't hardly ever remember these stages in the Armstrong era, or at least they were not as predominant as now. To me, having a number of these stages (and making them raceable) makes the Pyrenees less relevant, or at least potentially less important, when it comes to designing a good route. My geography is not great, but if the Pyrenean area is so much smaller than the Alps, then why race over it for multiple days?
I think that 5 high mountain stages is enough. 3 in the Alps and 2 in the Pyrenees, then 4 in the Alps and 1 in the Pyrenees (in which case you can make the latter a 220 km, six climb stage). Then you also go to two of the three medium mountain ranges for two days each, and that's 9 climbing stages, which is plenty. What I would want to see is at least 3 of these stages being properly hard and decisive. Then of course in a proper route, you also have a couple of time trials (or maybe a prologue, what happened to these?).
That doesn't mean that you have 10 flat stages. Most of these should be designed so that the winner is largely unpredictable. A 4km climb 30 kms out, or a 1km hill 8kms from the finish, etc. Possible crosswinds. Sprinters should have to work for their victories, but should be able to win multiple times if they are good enough.