wedrifid comments on Patternist friendly AI risk