Why it's difficult to compare SNARKs together, and why it would be nice to have a centralized table benchmarking and comparing them nonetheless, among other topics.