Some random thoughts on VCV: I built it from source the other night on Linux and have been impressed so far. JACK support isn't compiled in by default but works great if you follow the instructions on the wiki. The plugin ecosystem is still in its earliest stages. There is this confusing sort of 'login' mechanism (both on the website and in the app) which seems to be laying the foundation for commercial plugin sales; you're not required to use any of it right now. In fact, if you build the bleeding edge sources in their default state, it is disabled.
I found that many of the available plugins are broken against the bleeding edge development version (annoyingly). Each plugin maintains its own version number which I think is supposed to track the overall release version number (Rack is currently at 0.5.1 and plugins seem to be at 0.5.x). The claim is made that all 0.x plugins are compatible which 0.x releases but it seems like backwards compatibility breaking changes are being made. Hopefully there will be a better strategy for dealing with plugin API level in the future.
The UI is the most impressive part of VCV to me. It is extremely responsive and smooth on my development machine, but might be too expensive for lower-end machines. It seems like considerable effort has gone into making the modules look pretty. I'm not totally sure how I feel about the "realistic" studio rack UI metaphor. There is nothing that feels like a real programming language here, no abstraction mechanism. You are purely using modules with their pre-fab UIs. It all looks super pretty though and feels very approachable, simple. Like you are just sitting down at a physical device that "just works", not having to deal with much software complexity. I think this last aspect is something we could improve on in Axoloti, making the experience for streamlined for new users, particularly less technically inclined ones. What's interesting about the UI implementation is that it is not using JUCE or QT or Java or any other very heavy UI library; it's using GLFW which is pretty close to raw OpenGL but helps out with normal UI stuff like input and events, etc. It's definitely inspiring me to think about directions for future Axoloti UIs. It might interesting to have a system that had the ability to work with both more abstract object representations (like Axoloti currently) and pre-fab skeuomorphic modules (like VCV) at the same time. I guess this is kind of like Reaktor maybe?
Unfortunately, the actual sound hasn't impressed me that much so far, but I haven't had a chance to go through all the modules.