技巧精选
为 AI Agent 重建 COMPUTE 基础设施的 3 条经验
精选理由
给做 Agent 基础设施的人看的：沙箱安全、空闲成本、持久工作区三个坑都讲到了，实践味很浓。
作者在为 AI Agent 重建 COMPUTE 系统后总结 3 条经验：模型会犯错，沙箱里不能留任何可被攻击者利用的东西；Agent 大部分时间在等待，空闲状态应几乎零成本且唤醒速度快到人无感知；Agent 记忆越多越有用，工作区需要跨回合、跨天数、跨系统版本持久保存。
原文 · Shubham Saboo
3 leanings from re-building COMPUTE for AI Agents.
The model will make mistakes. when it does, the sandbox shouldn’t hold anything that an attacker can use against it.
Agents spend most of their time waiting. An idle agent should cost almost nothing and wake up faster than a person notices.
An agent becomes more useful the more it remembers. Its workspace should survive between turns, between days, and between versions of the operating system.
