I worked for about a year and a half on the ASRG project, specifically doing reliability work for the Stirling engines. They had a bit of a ways to go to prove out the system (its a tricky problem trying to "prove" that your device will last 17 years), but they had made a whole lot of progress. The biggest problem for the project though was that the ASRG wasnt being designed for any one mission, so its performance requirements were a bit of a moving target.