Think about your/our design goals here. If that's 'make the best cube available for <$100' then ditching singleton to get that is probably a better plan than having the singleton restriction and compromising that goal. If that's not the goal, then I've missed the point a bit here; even still, concrete goals are useful to point at as a utility function in these kinds of questions.
My fear is simply that breaking singleton will, unfortunately, drive many people away from the final result. If the theoretical "best cube<$100" is full of duplicates and idiosyncratic card choices, the project is likely to be dismissed as Not-A-Real-Cube, but instead A-Limited-Draft-Format-From-Those-Crazy-Guys-At-Riptide-Lab. I do think we can work towards something that can both provide a widely useful structure for others to build on and a strong cube in its own right. There is going to be unavoidable tension between these, however.