At last, one would say: doing extreme AI research at the very forefront (aka reinforcement learning) while letting results available to every malicious party out there? Headless chicken hubris or naive daydreamers, tertium non datur.