一个月交付 2000 个 PR：Lauren Tan 谈如何建立对智能体的信任

“I shipped 2,000 pull requests to production.”

我交付了 2000 个 PR 到生产环境。

“we're still responsible for the final outcome”

我们仍然要对最终结果负责。

“the codebase is really, like, the best form of memory, because agents love to extend existing patterns that they see.”

代码库其实是最好的记忆形式，因为智能体特别喜欢沿用它看到的现有模式。

“So what if we designed a framework such that the shortcut, you know, the easy path is the right path for agents?”

那如果我们设计一个框架，让那条近路、那条最省力的路，恰好就是正确的路呢？

“if you have existing anti-patterns, you will actually find that these will spread kind of like a virus”

如果代码库里已有反模式，你会发现它会像病毒一样扩散开来。

“whenever you see tech debt or bad patterns, your instinct should be, "I need to write a lint rule against it."”

每当你看到技术债或坏模式，第一反应应该是："我得写条 lint 规则来防它。"

“you trust the environment so much that your agents can just be free, right?”

你无比信任这个环境，你的智能体可以完全放手去干，对吧？