He talks about how he got into Delphi (via Algol) and then some of the research work he's been doing lately with C# at MS.
Somewhere in the podcast he mentions that creating specific languages in C# would be quite difficult, because it would require tools that would be able to parse statement trees full of code.
I wonder what his exposure is to lisp... I've walked expression trees myself a little in lisp and I've found it quite useful. I imagine I could write myself a whole language if I wanted to (I'm sort of in the process of doing it at the moment).
He also says that using generics made the process of building LINQ into C# a lot easier. I imagine he's right. However, I wonder if lisp might be easier again for this kind of task.
I must sit down on the train sometime and try and write something like a LINQ for lisp. I imagine it would be possible.
At the moment still building my plsql generation language (mentioned above). Now it has a built in type inferencer (courtesy of Pascal Bourgignon sp?).