"You either die a capabilities researcher or live long enough to see yourself become an alignment researcher." Aidan McLaughlin says his job at OpenAI is quietly turning into the second one. @aidan_mclau, research scientist on @OpenAI's Core Models team, joins @theojaffee and @schisofrenia in the MTS studio on recursive self-improvement, why alignment and profit point the same way, and the chess video that got him into AI: 01:41 X as the uncheatable eval 03:10 his one-sentence take on…