Extinction-by-AI remains a fringe belief among actual experts, and even most who entertain it call it low probability and preventable. A single viral resignation post spread panic without ever explaining how software actually wipes out civilization. Banning a technology based on apocalyptic fantasy, rather than real risks like hacked infrastructure and sloppy software, is legislating by vibes.

Both camps are buying the industry's favorite premise that today's AI is on an inevitable march toward godlike intelligence. Doom rhetoric turns speculative superintelligence into a reason to panic, while dismissing that rhetoric lets companies keep deploying unreliable systems into infrastructure, warfare and surveillance. The real danger is more immediate: humans surrendering consequential decisions to technology that routinely fails, while its makers grow richer and more powerful either way.