跨国深谈← 返回←Jerry Liu

① 他说过的话

7 条

归根结底,你可以拥有一个无限聪明的智能体,但你从这个能端到端做越来越多事情的无限聪明的智能体中真正获得价值的能力,其实是给它正确的任务去做。
In the end, you could have an infinitely smart agent, but your ability to actually get value out of this infinitely smart agent that’s able to do more and more stuff end to end is actually giving it the right things to do.
—— Jerry Liu · [05:38]

指向原始笔记的链接

有超过 10 万亿页的人类原生知识被锁在 PDF、PowerPoint、Word 文档和 Excel 表格中。
There’s over 10 trillion-plus pages of human-native knowledge locked up within PDFs, PowerPoints, Word documents, and Excel sheets.
—— Jerry Liu · [06:35]

指向原始笔记的链接

如今很清楚的是,越来越多的人只是用英语来构建东西,无论你是软件工程师,还是身处 go-to-market、市场营销这样的非技术职能。
It’s pretty clear these days more and more people are just building stuff using English, whether you’re a software engineer or you’re in a non-technical function like go-to-market, marketing.
—— Jerry Liu · [03:58]

指向原始笔记的链接

所以未来将走向一种状态:你甚至可能不必用英语定义任务,而是更多定义目标和评分标准,然后智能体会用它所有可用的上下文真正解决任务。
So the future will move towards a state where you actually might not even have to define the task in English, but actually more the goal and a scoring rubric, and then the agent will use all available contexts available to it to actually solve the task.
—— Jerry Liu · [04:57]

指向原始笔记的链接

但文档 OCR 的帕累托前沿,相比前沿模型在文档理解方面所处的帕累托前沿,总是会精确得多也便宜得多。
But the Pareto Frontier for document OCR will always be much more accurate and cheap compared to the Pareto Frontier for wherever the Frontier models are in terms of document understanding.
—— Jerry Liu · [12:19]

指向原始笔记的链接

有我所谓的高准确性区间,一些机构基本需要 99 到接近 100% 的准确性,因为错误提取的负面后果就是你完全搞砸你的财务模型。
There’s what I call like the high accuracy regime where like you know some institutions basically need like 99 to almost 100% accuracy because basically the downside of incorrect extraction is you completely mess up your financial model.
—— Jerry Liu · [15:13]

指向原始笔记的链接

我们把文档视为非结构化上下文的通用容器。
We see documents as universal containers for unstructured context.
—— Jerry Liu · [06:31]

指向原始笔记的链接

② 出现在这些集

1 集

• 《为 AI 智能体构建文档上下文层:LlamaIndex 谈 2026 年的 RAG》 — 作为嘉宾

③ 他谈到的

点进去有真内容 —— 本页主要出口

LlamaIndex · LlamaParse · LightParse · parsebench · RAG · 智能体 · 上下文 · 文档 OCR · VLM · MCP 服务器

④ 也在聊「智能体」的人