精选理由
GPT-5.6 Sol在Codex里花了5小时自己打穿了Slay the Spire 2每日挑战，这自主性有点强啊。
用户将GPT-5.6 Sol模型通过Codex平台赋予电脑控制权，要求它赢得游戏Slay the Spire 2的每日挑战。该挑战包含随机化因子，无法通过预知方式作弊。模型连续工作了5个小时，自主做出复杂游戏决策，最终成功通关。此演示展示了AI在长时间自主任务中的决策能力。
原文 · Ethan Mollick
This was one of those impressive AI thresholds for…
This was one of those impressive AI thresholds for me. I gave GPT-5.6 Sol in Codex control over my computer, and asked it to win the daily challenge for the game Slay the Spire 2 (randomized factors, so can't cheat).
It worked for 5 hours, making complex game choices... and won. https://t.co/IVsMiyNYEk