I would say that there is definitely a lot of overlap among the five components. Some are more related than others. E.g., it's possible to have great interpretation with only simple transitions or vice versa, so I in theory there shouldn't be as much correlation between those two marks, but both are much more directly related to choreography.
Speed is certainly a product of good skating skills and also enhances the impression of everything else, both the other components and the technical elements. It's not the only aspect of skating skills, though, so there would be some cases where a slower skater would deserve a higher SS mark than a faster one. (E.g., I'd expect John Curry to score higher than Surya Bonaly.)
Not all moves are easier or possible to do at all, or to do well, with no speed because the technique may need to harness the momentum of the preceding skating. But that will usually show up in poorer quality of the move done at slow speed, or in inability to do it at all (if the skater tries the move at speed, s/he can't control it; if s/he tries it from a standstill, s/he can't use the momentum correctly).
In the old system, the technical mark included the basic skating skills, and even when judges were encouraged to separate the two marks they usually tended to be within 0.2 or so of each other (before any short program deductions).
In the new system, it's a lot easier to separate the elements marks from the component marks because they're not even on the same scales. In general the best skaters tend to be good at both elements and skating skills, and also pretty good at presentation, and the weaker skaters tend to have more weaknesses throughout. But there are exceptions. Also, strong skating skills will affect the ice coverage, edge quality, and so on of elements such as step sequences and spiral sequences, and also jumps and lifts, so the skaters who deserve higher skating skills marks are also more likely to get higher GOEs for those elements, and probably better able to achieve the features for higher levels in steps and spirals.
I think that judges and judge trainers are still figuring out the best way to approach assigning component marks to best reflect the real differences in the skating and the guidelines have changed over the four years the system has been in use. Differences in how judges use the marks will vary depending on their own personal systems and how/when/where they were trained.
If we come back in 10 years, I expect there to be more judges experienced with the system to be using the PCS more confidently and consistently, with gaps of 2 or 3 whole points between individual skaters' components often evident where applicable. But I don't expect such gaps to be applicable for most skaters because the skills do tend to be related and the 0-10 scale needs to cover every possible skill level from barely able to stroke and turn three turns or mohawks to all-time great.
