OpenAI claims Millennium Problem solved NYU mathematician claims they copied his work He'd run drafts through Codex all year Asks if the model was trained on their session No clear answer OpenAI wants Alpöge dropped from the paper He threatens to go public “Why would you ruin your career?” Sebastien Bubeck from OpenAI says this is false and inflammatory. Is OpenAI training internal models on the data that goes into Codex? @jason @lons @VCrebecca @benjlerer @yoheinakajima