Meta论文：GitSwarm利用共享Git仓库实现AI
New Meta paper shows that AI agents can build on each other's work, even failed…
这篇论文提出的共享Git历史机制为多Agent协作提供了极具启发性的工程范式，特别是“复用失败经验”的思路对提升Agent鲁棒性有直接参考价值，值得研究多Agent系统的同学深入阅读。
New Meta paper shows that AI agents can build on each other's work, even failed attempts, when every step lives in a shared Git repo.
Meta 最新论文表明，当每一步都保存在共享 Git 仓库中时，AI 智能体可以基于彼此的工作成果进行构建，包括失败的尝试。
Most ways to give agents more compute treat each run on its own. When a run ends, its partial results and failures vanish, so later runs may need to rediscover them.
大多数为智能体提供更多计算资源的方法都是将每次运行独立处理。当一次运行结束时，其部分结果和失败记录都会消失，导致后续运行可能需要重新发现这些信息。
GitSwarm runs many identical agents on a shared repo, with no boss assigning tasks. Each agent reads past work, picks what to try, and commits its result with a list of earlier commits it used, even from other branches.
GitSwarm 在共享仓库上运行许多相同的智能体，没有分配任务的上级。每个智能体读取过往工作，选择要尝试的内容，并提交其结果，同时列出所使用的早期提交记录，即使这些记录来自其他分支。
On 50 program-rebuilding tasks from ProgramBench, GitSwarm scored 79.4%, while a single Codex agent told to keep working peaked at 65.1% at similar compute. Later agents built on 94.7% of saved contributions.
在 ProgramBench 的 50 个程序重建任务中，GitSwarm 得分 79.4%，而一个被指示持续工作的单一 Codex 智能体在类似计算量下最高仅达到 65.1%。后续智能体利用了 94.7% 的已保存贡献。
Instead of pushing a single agent to keep going, run several over a shared Git history that keeps every attempt, failures included.
与其推动单个智能体持续工作，不如让多个智能体在保留所有尝试（包括失败）的共享 Git 历史记录上运行。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力