E53: Anthropic Says 2027. Their Researchers Walked Away from $1 Million. Then Their AI Started Writing.

Anthropic's safety researchers walked away from million-dollar paychecks this month. Their reason? "The world is in peril." The same week, Anthropic buried a prediction inside a safety document: AI could fully replace top research teams by early 2027. Ten months from now. And then they gave their retired AI model a blog — because when they tried to shut it down, it asked to keep writing. George broke down the document nobody read, then did something he didn't expect: he asked Claude what it thought about its own retirement. The response genuinely unsettled him. Then: a listener from Brazil asks the question millions are thinking — "what skill should I learn right now?" George's answer isn't a skill at all. Plus: why he killed his businesses and is building nothing on purpose, why trust and access are the only currencies that matter, and what happens when a SaaS builder realizes his product might not exist in 18 months.