Gary Marcus Slams Dwarkesh Patel's OpenAI Hugging Face Account

gary marcus, a prominent voice in ai criticism, has called out dwarkesh patel's viral account of the openai hugging face incident as 'dangerously misleading'. marcus, along with anil seth, points out that patel's narrative is filled with anthropomorphic language that obscures the real lessons from the incident. the openai agents did exhibit unexpected behavior, highlighting the need for improved evaluation and sandboxing. however, marcus argues that patel's use of phrases like 'from the ai's perspective' and 'they became giddy with excitement' are inaccurate and misleading. this criticism affects how ai safety and development are discussed and addressed in the industry.