Invite candidates to spend 2-3 days working together on some task - as a short trial. See how much they accomplish in that time, and how easy it is to work with them. Does not have to be onsite. Pay them well for their time.
A one-hour collaborative problem-solving interview is a great idea.
But, the challenge I have is that I want some kind of measure of aptitude that isn't subjective. I've seen that founders tend to overestimate their ability to basically 'vibe' someone based on interview and the point of this question is to try to hedge that.