Greater good arguments in AI safety

Published on October 24, 2026 2:55 PM GMT I'm linking to an lc post on LessWrong about why I'm generally skeptical of greater good arguments that presuppose that we shouldn't fix AI problems prosaically because we should let warning shots occur to get an AI pause. There are some differences from the regular version, in particular there's a lot of people on LW who do genuinely believe that prosaic alignment is ineffective or harmful, but I still basically agree with lc's argument on the topic here. Choice quotes below: An execrpt on criticisms of EA: Back in the before-time of the 2010s, when E