And in the DES s-box case, you'd have been right to trust them, because they fixed a vulnerability in DES 20 years before the theory behind it would become public. They couldn't disclose the reasoning behind the new s-boxes without disclosing differential cryptanalysis, because they generated the new s-boxes by first generating random candidates and then testing them for resilience against differential cryptanalysis.
True and that is the problem they are facing. Before when they came in and said "trust us it is better" NIST and everyone would say "yup we trust you". Now it has switched to "no way, you are just building in a backdoor". That is the sad part.
