I watched the video. I took the code coverage "coverage" for what it was worth. While it's not the best testing tool that can be built, it IS useful. Any developer worth a paycheck understands its limitations and benefits. I don't recall hearing any unrealistic claims about code coverage's capabilities. So, rather than slam it, perhaps we should applaude the team for doing something positive rather than doing nothing? I fully expect the next Visual Studio to take the next step. Code coverage detection isn't easy. The next step in automated testing will certainly be more difficult, say: Validate that methods are called with every conceivable combination of parameters? Validate that all of your exception handling does what you expect? Working on an automated testing solution that would excersize parameter boundary conditions certainly would be interesting.
My biggest hurdle so far has been the "day one productivity." Just figuring out how to do simple tasks, like follow the MSF for Agile template process, is painful. I haven't been able to find much in the category of "Perform steps A, B, C, ..." to get started. When I do, it's unclear, like "Start the web form designer." What's that? After a bit, you discover that it's just Visual Studio. It seems like such a stupid question once you've seen the answer. Why not just say, "Start visual studio."