JoshAEngels says he left Google DeepMind's AGI safety team for METR
He says he turned down offers from Anthropic and OpenAI, arguing that current AI systems aren't aligned enough to safely begin recursive self-improvement—where AIs build smarter AIs.
TLDR
On September 12, 2026, @JoshAEngels said he had left Google DeepMind's AGI safety team three weeks earlier to join METR, citing the stakes of AI safety. He argues that a misaligned recursive self-improvement loop—in which AIs build smarter AIs—could be catastrophic. He describes a “terrifying chance” that AI systems cause immense harm over the following five years, while saying he doesn't know the exact probability. He calls for pacing AI development to give safety more time.
Combined views
1.4M
16 Sources, first seen 19d ago