Overall this presentation is really good, I liked how Mr. Hejlsberg explained the current problems of modern programming and the approach C# took to accomodate those problems.
However, I wish he talked more about the debugging and specific run time errors that can occur.
The problem with dynamic, as far as I can tell, is that since you could give a dynamic variable anything you want, you won't know if it will work until you run the application at run time; but even then, you won't even know where the error will be in the code!
If Mr.Hejlsberg dedicated sometime to show possible runtime errors, and if there is some kind of message handling that will explain viable reasons as to why something is not showing, I'd program using dynamics with higher confidence. But as far as now, I feel a little effy on making the leap and spending my time asking others why something isn't working.