Any my point was, once you push you 'science' to the point where it is indistinguishable from 'magic', you have walked away from the 'hard sf' corner and into the 'fantasy' corner.
This was why I mentioned Von Neuman's War. The base assumption is that self replicating machines have invaded the solar system. I'm cool with that.
They range in size from nano to macro. I not so cool with that. Macro machines are cool but building them from nano machines raises questions about where the plans are stored and nano machines stop being nano when you ask them to cart around the data to build and program large machines. However, the biggest problem is the energy budget. Where do you get the energy to power/motivate these machines? Entropy asserts that whenever you do something, there is energy lost. The busier your machine is, the more energy is lost.
So, the story requires that I believe that a cloud of nano machines, floating over the landscape, can disassemble a missile, in flight, with no energy release.
Somehow, the writing community got the idea that nano machines will be able to manipulate matter one atom at a time to build any desired molecule or material. Obviously, if you can put it together, you can take it apart. To the best of my knowledge, molecules are generally hard to take apart. Diamond is, perhaps a little extreme. On the other hand, some molecules come apart all to easily (the various fulminates for example).
Anyway, hopefully you get the idea. If the reader has a background that tells him a particular premise isn't going to work, the story cannot, for him, be hard sf.
When I was in high school, I felt good about the range of science that I was learning about through hard sf. Now, many years later, that means I have a huge knowledge base and that means that lots of stories are built on premises that I don't believe will work. I would like to read more hard sf but, what I feel to be, flaws keep getting in the way of the story and so I read lots of outright fantasy instead.
Outright fantasy has whole different set of problems but I'll save them for a more appropriate discussion.