Tag
NYU finance professor Arpit Gupta shared observations after spending two days at Anthropic, covering progress toward recursive self-improvement, AI lab competition undermining safety, the dual nature of alignment problems, and Claude increasingly behaving like a person.
Google DeepMind launches the DeepMind Institute to address AGI's technical and societal implications through interdisciplinary research on safety, governance, and human values.
A Google DeepMind AGI safety researcher has resigned, warning that AI poses existential risks to humanity and time is running out.
The article analyzes recent AI policy developments including a White House memorandum on AI use, OpenAI's plan for AGI safety, and the release of Claude Fable 5, while contrasting the philosophical approaches of Anthropic and OpenAI.