Oliver's description is correct. We're two separate features development teams, which focus on different approaches to editor engagement. We collaborate on some projects, but there isn't a process where we experiment to find what works, then hand it off to E2 or someone else to build permanently. When we find something worth making permanent, we generally do that ourselves. 

I kind of like the way Terry puts it: E2 builds the levers. E3 figures out how to pull them. An example might be Echo, the new notifications system. E2 is building out that platform, and regardless of details we know that we want a modern notifications system to replace that damn yellow messages bar, including a way for developers on other teams and in the community to trigger a notification. The E3 approach to "pulling that lever" might be to test the effectiveness of particular notifications at driving editing activity. 
--