The obvious problem here is that Facebook, Twitter, Path, an airline website: these are all public or essentially-public sites and programs that most designers can and do use and therefore know where they succeed and where they fall short of being optimally designed.

A POS system? Dental office software? These aren't systems that your average Behance and Dribbble user uses or even has access to. Besides some basic cosmetic enhancement based on some screenshots, or a very abstract outline of suggestions, I don't see how a designer could possibly work up a sensible redesign of something like "Dentimax". It's pure hubris to assume that any random person with no domain knowledge, no understanding of the needs of the market for a product, and likely not even a working copy of the product can provide more than platitudinous advice about how it could be improved.

The ATM example is poor, in my opinion. ATMs that still look like that almost certainly do so because they're too old to look any other way. If it's really a problem to use an outdated ATM without a beep-boop touchscreen technicolor interface, tell your bank to replace it. "Imagine the difference you could make even in just one hour." Sure, if that's all the time it takes to yank the old one out and put in a new one.