Eric Lippert, who has singlehandedly done more for purple Lucida Sans Unicode than anyone I know, links to my blog http://blogs.msdn.com/ericlippert/archive/2004/04/09/110556.aspx.
Although he's got plenty to blog about already, hopefully he will eventually get around to blogging about some of the cool work he's been doing on VSTO 2.0--if I don't beat him to the punch. For example, Eric has been working on the generated data/view programming models we create when you use the schema mapping features in Word and Excel. Picture mapping a schema like “Customer.xsd” in Excel, going into the code editor, and seeing nicely named controls that drop down from the left drop down list in the code editors--all named with names that correspond to the terminology you used in the schema.
Eric has also done some a lot of interesting work on the code generation we do for you in VSTO 2.0. To generate code on your behalf, we use an intermediate declarative XML representation that we then convert to C# or VB.NET. Of course, a similar approach is also used in Avalon's XAML files. Eric blogs about some of the implications of this approach here: