← back
x.com烟花老师Mon, May 25, 2026, 8:14 AM PDT
score 26.0
211likes43RT11reply

Long-running AI agents struggle with state, effort estimation, and self-evaluation

Original: 我最近两个月的探索和实践分享:长任务 Agent 的最小工程闭环

Source: x.com

Writing ELI5 summary…