Reducto CEO谈PDF解析：99%准确率为何还不够
Reducto 老板做客 Latent Space，边做牛排三明治边聊 PDF 解析为啥难，模型会偷偷丢表格行，每月 10 亿页踩出来的坑。
Latent Space 播客邀请 Reducto 联合创始人兼 CEO Adit Abraham 对谈。Reducto 每月处理约 10 亿页文档，其文档解析产品 Reducto R1 已迭代到可自动路由难页面和表格的版本。节目中解释了前沿模型在处理 PDF 时会静默丢失表格行，以及为什么 99% 的基准准确率在生产环境中仍然不够用。内容还涉及何时该自己训练模型、持续评测、以及办公自动化何时会有类似 Claude Code 的爆发时刻。
In this episode, @reductoai co-founder and CEO @aditabrm joins @allenpark to make steak sandwiches with burrata, torched peaches, and chimichurri and talk about why PDFs still break AI: how frontier models silently drop rows, why 99% accuracy can still fall short in production, what Reducto learned processing a billion pages a month, and what it would take for office work to have its Claude Code moment. 00:00 Cold Open 00:13 Meet Adit Abraham and the Viral Steak Tweet 00:57 From Remembrall to Reducto: YC and the Pivot 01:57 Why PDFs Are Still Hard for AI 03:35 What AI Agents Need to Do With Documents 04:19 Model Routing for Tricky Pages and Tables 05:29 Build vs. Buy: When to Train Your Own Models 07:00 Deep Extract, Frontier Models, and Missing Rows 08:20 Why Everyone Wins Their Own Benchmark 09:29 Reducto R1: Accuracy, Cost, and a Billion Pages a Month 10:57 When 99% Accuracy Isn’t Enough 13:05 Continuous Evals and Production Failures 16:49 How Reducto Uses AI Agents Internally 18:28 Hiring With AI and the Eight-Hour Onsite 20:17 Early Hires and What Obsession Looks Like 23:17 Blowtorching Peaches 24:24 Scaling From a Small Team to Series B 27:34 Competition and the Data Flywheel 30:52 What Better Parsing Unlocks for Customers 32:11 Finding a Cofounder at MIT 34:34 The Anthropic Hackathon, Long-Term Memory, and RAG 38:49 From Reading PDFs to Redlining Contracts 43:19 A Claude Code Moment for Office Work 46:13 What’s Next for Reducto: Extraction, Citations, and Cost 47:51 Assembling the Steak Sandwich 49:55 The Steak Allegations and Taste Test Your browser does not support the video tag. 🔗 View on Twitter 💬 4 🔄 2 ❤️ 4 👀 463 📊 5 ⚡
