One of the points Casey is trying to make with this series is that, at least in games, dynamic languages aren't quite the silver bullet for productivity that people make them out to be, and you can write your gameplay code just fine in C or C++. It might even work out to be a net productivity win, because native code debuggers are far superior to "managed" language debuggers (most of which don't handle the boundary between managed and native code at all), and because even moderately well written C or C++ is much less likely to cause performance problems down the line than anything garbage collected.