Summary
An AI researcher who recently left Anthropic, Jacob Coxon, discusses the company's internal "mini Manhattan project" and the challenges of AI alignment. Coxon emphasizes the urgency for AI labs to ensure system safety within the next few years.