'Things Will Never Be Chill Again':The Doomers Who Shaped the AI Safety Freakout (www.wsj.com)

A group of AI researchers at Berkeley's Constellation co-working space are tackling existential risks associated with artificial intelligence, following heightened public concern catalyzed by the recent resignation of Anthropic researcher Jacob Coxon. This community, which includes some of the earliest employees of leading AI firms like OpenAI and Anthropic, engages in intense discussions about potential doomsday scenarios linked to unchecked AI development and advocates for precautionary measures. Their gatherings mix professional dialogue with a dose of dark humor, as seen in a recent event featuring a cocktail playfully named “Death With Dignity.” The significance of the Constellation community lies in its influence on AI's trajectory through its focus on safety and ethical considerations. Their alignment with effective altruism has attracted funding and talent, contributing to the responsible development of AI technologies. As Anthropic approaches a potential $2 trillion public offering, the impact of these "doomers" is palpable; their internal discussions about doomsday preparations highlight a collective urgency to promote safety in AI innovation amidst growing fears about its consequences.