People working on Reinforcement Learning or AI safety that don't have kids can not be taken serious. Anyone with small kids knows that "preventing agents hacking the reward function" or "designing agent incentives to avoid reward tampering" is categorically impossible and it is ok.
