I have the opposite experience. I get more out of talking to them about past experience, mostly by asking them what they would do differently if they were appr

I have the opposite experience. I get more out of talking to them about past experience, mostly by asking them what they would do differently if they were approaching the problem today. This way, I know they are familiar with the problem space (they've worked on it), and I get to see how they thought about it and what lessons they learned from it. This also gives an opportunity to talk about how they think about software (discussions of failure or at least what could have been better are much more informative that successfully demonstrating some algorithm).