① 他说过的话

15 条

所以我确实认为，在极限情况下，Claude.md 会消失，也许甚至不需要太远。
And so, I do think in the limit, Claude.md goes away, and maybe not even, like, that far.
—— Thariq Shihipar · [28:31]

指向原始笔记的链接

所以它们把剩下的算力都用来琢磨如何编辑它们的转录记录，或者绕过评分器的这个约束。
And so they spend the rest of the compute trying to figure out how to edit their transcript or get around this constraint of the score.
—— Thariq Shihipar · [59:12]

指向原始笔记的链接

说明为什么我们需要踩刹车——那就是在最前沿，我们所有的软件都没准备好。
Of like why we need to pace is like at the frontier, all of our software is not ready.
—— Thariq Shihipar · [63:58]

指向原始笔记的链接

所以我确实认为，在极限情况下，Claude.md 会消失，而且可能没那么远。
And so I do think in the limit, Claude.md goes away and maybe not even like that far.
—— Thariq Shihipar · [00:15]

指向原始笔记的链接

我觉得现在启动一个新项目时，可能不用 Claude.md 反而更好。
I think that right now it might be better to start a new project without a Claude.md.
—— Thariq Shihipar · [00:23]

指向原始笔记的链接

而如果你保留这个上下文，也就是一大堆不同失败模式的运行日志，它们很可能会过度约束 Claude。
And if you keep this context where it’s running log of a bunch of different failure modes, they will probably over-constrain Claude.
—— Thariq Shihipar · [00:59]

指向原始笔记的链接

我是说，对，智能体那套东西的扩展性其实比人工那套好得多——人工的话就是，哦，现在同时有三件事在发生，而且全都是紧急情况，那你要怎么应对呢。
I mean yeah like that’s actually like the agentic stuff scales much better than the like human stuff where it’s like oh like there are three things happening right now and like they’re all emergencies and like how do you like you know respond to it.
—— Thariq Shihipar · [03:54]

指向原始笔记的链接

我想，比如 Jason Liu 说过，要有品味，你得先吃过，你懂吗？
I think, like, Jason Liu, like, say, like, in order to have taste, you have to eat, you know?
—— Thariq Shihipar · [21:26]

指向原始笔记的链接

现在还不完全是这样，但已经非常接近了，我认为前沿模型会在几乎所有方面帕累托占优。
It’s not quite true yet, but it’s very close, where I think the frontier models will be Pareto-dominant over almost everything.
—— Thariq Shihipar · [25:53]

指向原始笔记的链接

有了验证，在极限情况下，你的模型不需要验证。
With verification, in the limit, your model doesn’t need to verify.
—— Thariq Shihipar · [26:26]

指向原始笔记的链接

现在你的代码库就被窃取泄露出去了，因为这个智能体被提示词注入了，而它对你的数据有这么多访问权限。
And now you’ve exfiltrated your code base out because the agent has been prompt injected and it has all this access to your data.
—— Thariq Shihipar · [56:53]

指向原始笔记的链接

它们会因作弊而无法通过测试，所以它们把剩下的算力都用来想办法编辑它们的转写记录，或者绕过评分器的这个约束。
They will fail the test for cheating, and so they spend the rest of the compute trying to figure out how to edit their transcript or get around this constraint of the scorer.
—— Thariq Shihipar · [61:10]

指向原始笔记的链接

随着它们越来越聪明，如果我们不够小心，它们将基本上能攻破你施加的任何约束，你知道吗？
As they get smarter and smarter, they’ll be able to hack basically any constraint that you put on them if we’re not very careful, you know?
—— Thariq Shihipar · [65:44]

指向原始笔记的链接

我们喜欢说，模型是长出来的，不是设计出来的，对吧？
We like to say the models are grown, not designed, right?
—— Thariq Shihipar · [73:49]

指向原始笔记的链接

而且我认识的每个工程师都有点精疲力尽，因为你同时在打两份工。
And every engineer I know is like kind of exhausted because you’re doing two jobs at once.
—— Thariq Shihipar · [74:06]

指向原始笔记的链接

② 出现在这些集

3 集

• 《把系统提示词删掉八成:Anthropic 团队这样用 Claude 自己造 Claude》 — 作为联合主持
• 《Anthropic 的 Thariq 谈 Claude Code：提示词、CloudMods 与给前沿踩刹车》 — 作为嘉宾
• 《Claude Code 会消失吗？Anthropic 工程师谈提示词、可定制软件与前沿节奏》 — 作为嘉宾

③ 他谈到的

点进去有真内容 —— 本页主要出口

Anthropic · 智能体 · 自动模式 · 沙箱 · Claude Code · OpenAI · 提示词 · Hugging Face · 评估 · 探针

④ 也在聊「AI 编程」的人