One of the features provided by RIA services is that your server-side model objects get automatically proxied in your Silverlight front end and are automatically serialized back to the server for operations such as saving to the database.
This seems to be a very similar goal to what "Self Tracking Entities" are providing, which is a technology developed by the ADO.NET Services team.
Are these two technologies solving the same problem? Or are they solving subtly different problems? Is one "better" than the other in certain scenarios? What are the pros and cons of each?
They appear to be solving the same problem, but for slightly different situations. The RIA services model is used for server side code that won’t work in SL, for example going directly to SQL with LINQ-to-SQL or EntityFramework. On the other hand ADO.NET Self-Tracking Entities is more generic and can be used to track things purely on the client side, such as WCF data contracts.
From what I see they are solving the same problem, but have a different target situation.