"AI could kill off humanity" is... perhaps the most-remote, most-boring, most-distracting outcome to discuss in AI safety. GPTs have no chance of becoming an e

"AI could kill off humanity" is... perhaps the most-remote, most-boring, most-distracting outcome to discuss in AI safety. GPTs have no chance of becoming an extinction-threat level AI. Intelligence is more than just spitting out words when an answer is demanded (and the folks who are saying GPTs are as "intelligent " as some people make me really worry what those folk must think the role of other people is in their lives). Putting a fancy chatter bot in an open loop isn't going to suddenly create a goal-seeking paperclip optimizer. But there is real harm already being done to people by corporations and governments using AI tools. Yet somehow that doesn't ever come up in the "AI Safety" discussion. Facial recognition getting the wrong person sent to prison. Bad medical advice. Insurance policy denials. Where are the calls from "concerned AI researchers" for slowing down biometric data collection? And if they are so concerned, why are they so busy pouring gasoline on the fire?