Anthropic blocks malicious use of AI that could develop biological weapons

Ex-Anthropic employee Jacob Coxon warned people "do not underestimate the power of this technology" before claiming people building the AI "earnestly believe that it could kill us" by 2030