I have liked PRO for some time. I must admit that I don't have the need to make maps, but I use the analysis tools in Arctoolbox extensively. I also do a lot of programming and Python is a great language, particularly on the teaching front.
I must comment esri, on their efforts to add a channel to other open source programs like those in the SciPy stack (scipy, numpy, matplotlib etc etc), and little things the R-bridge. The only place that I think they blew it was with the ArcGIS module and bringing Pandas into the fold while skipping direct import of numpy data as well (it might make a lot of the Spatial Analyst redundant I suppose)
I have always found Pandas bloated for doing simple things with arrays, whether they be raster images, raster data or tabular data. Numpy is lurking beneath the skin and they have some good functions in the numpy to featureclass/table conversion suite, but they make it a pain to bring array data back into Pro. It isn't a game stopper, just a pain, and everyone loves pandas.
I would like esri to spend less time on the cool display stuff and put more effort into analytical and statistical tools. Shuffling that to R is catering to a particular market and missing out on opportunities as the newer generation begins to take over the helm. I have this conversation with a wall everyday, but I suppose it won't change.
I think they also should draw the line on trying to deploy all analysis in all platforms and devices. If a complex network analysis needs to be done, then do it on a standalone machine. Trying to push some of these analyses to the web might have a market, but it causes more problems than not.
I will retire, thankfully, before ArcHydro is deployed on the iPhone , but until then, if people would remember that ArcMap isn't Pro and vice-versa, then they might enjoy the good bits that both have to offer.
For now, Pro is the future, but I also thought every piece of software since the mid-70s was the future