Brian Barrett is the executive editor of WIRED, where he writes and edits coverage of technology and business. He previously served as editor in chief of Gizmodo and worked as a business reporter for the Yomiuri Shimbun.
Barrett questions why Anthropic embraces recursive self-improvement despite the risks described in its IPO filing. He argues that competitive incentives push AI companies toward the same process they warn could outrun their safeguards.
“serious consideration would be given to whether RSI should be pursued at all, much less embraced.”
After leaving Anthropic, Jacob Coxon said Anthropic and OpenAI were irresponsibly racing toward self-improving superintelligence, while praising Anthropic's current practices. Responses dispute catastrophic-risk forecasts, working inside AI labs, and whether governments should pause or regulate development.
