Google DeepMind：Gemini 4
Banger paper from Google DeepMind on long-horizon multi-agent coding.
Agent领域里程碑式成果，多模型协同解决复杂长程任务，值得Agent开发者深入研究其协作机制。
Banger paper from Google DeepMind on long-horizon multi-agent coding.
来自 Google DeepMind 关于长周期多智能体编码的出色论文。
(bookmark it)
（收藏它）
They report that, to the best of their knowledge, this is the first end-to-end autonomous completion of Minecraft by AI agents.
他们报告称，据他们所知，这是 AI 智能体首次端到端自主完成 Minecraft。
They report that four Gemini 4 Argon agents beat Minecraft end-to-end. They defeated the Ender Dragon in 27 hours and 14 minutes with no human intervention.
他们报告称，四个 Gemini 4 Argon 智能体实现了 Minecraft 的端到端通关。它们在无人干预的情况下，耗时 27 小时 14 分钟击败了末影龙。
Each agent controlled one player through the standard Mineflayer library. No one assigned roles, wrote game logic, or scouted a seed in advance. The agents planned, wrote, and debugged every program themselves, and coordinated through in-game chat and a shared memory.
每个智能体通过标准的 Mineflayer 库控制一名玩家。没有人预先分配角色、编写游戏逻辑或侦察种子。智能体自行规划、编写和调试每一个程序，并通过游戏内聊天和共享记忆进行协调。
Over the run, they wrote 2,783 programs and died 605 times. They split the work, shared what they learned, and recovered each other's items after deaths.
在整个运行过程中，它们编写了 2,783 个程序并死亡了 605 次。它们分工合作，分享所学经验，并在死亡后互相取回物品。
If you build agents for long software or enterprise tasks, this is a useful case study. The team had to keep one plan consistent across thousands of dependent steps for more than a day.
如果你为长期软件或企业任务构建智能体，这是一个有用的案例研究。团队必须在一天多的时间里，在数千个相互依赖的步骤中保持一个计划的一致性。
https://arxiv.org/abs/2610.22319
Chat with Paper: https://academy.dair.ai/papers/a-team-of-gemini-4-argon-agents-beats-minecraft-2610.22319
与论文对话：https://academy.dair.ai/papers/a-team-of-gemini-4-argon-agents-beats-minecraft-2610.22319
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