Toward Training Superintelligent Software Agents Through Self-Play SWE-RLarxiv.org2 points·klipt··0 commentsOpen articleSaveView on HN