All the criticism of the Richards pick assumes that optimizing at each position results in optimization of the defense as a whole. That exposes the critics lack of understanding and experience in dealing with large complex systems. You need to look at the system constraints and where bottlenecks are preventing the system as a whole from being optimal. I am convinced that is the actual approach of BB and the staff.
Perhaps with the perfect player at each position, you could have BOTH local optima and global optimum satisfied. However, with the salary cap, injuries, and other factors that is a unicorn. Given that there will always be inefficiencies at a local level, what can you do to ensure that you get the best global result from the pieces of the puzzle available to you?
As we've seen so many times, mistakes cost games. Don't make mistakes (turnovers, being out of position, etc.). On the flip side, forcing other teams to make mistakes or beating them to the position wins games. My argument is that adding a player like Richards is all about getting a player who is good enough to play every down, but because of his aptitude to consume and process data quickly he can have a greater impact by putting his teammates where they need to be to do their job and maybe where they need to be to make a big play. He may be 20% worse than the guy they may have selected at #64, but if he makes the other 10 guys around him 10% better, that's a bonus for the defense as a whole.
If you want to read more about the concept of theory of constraints, there is a non-technical (kinda hokey) book that I can't recommend highly enough called
The Goal.