Toward Training Superintelligent Software Agents Through Self-Play SWE-RL (Meta)arxiv.org1 point·xhevahir··0 commentsOpen articleSaveView on HN