The demonstration on youtube is really nice. The 500 x 500 matrix of objects showed a really fast speed up. I suspect though that internally theres still a lot

The demonstration on youtube is really nice. The 500 x 500 matrix of objects showed a really fast speed up. I suspect though that internally theres still a lot of plumbing that needs to be changed on top of what they've already changed. It was the same way for reflection, lambdas, threads. They really do a great job on iterating and keeping backwards compatability too.