Show HN: 我们让 AI 从 45k 封 Enron 邮件中生成一个 wiki,每条事实均附引用
Show HN: We had AI write a wiki from 45k Enron emails, every fact cited

原始链接: https://www.thisandthat.chat/enron/

他们从 45,047 封 Enron 邮件构建了一个可搜索的“公司大脑”。AI 生成了 214 个实体页面,包含 49,401 条事实,每条事实都与一封源邮件关联;当后续邮件显示某条事实不再为真时,该事实会过期。留出审计发现,抽样事实中有 93% 完全有来源支持,且 500 个抽样事实中没有一个核心内容错误。页面聚焦于在至少 25 封邮件中被提到的组织以及关键 Enron 单位,而产品、设施和事件仅被列出,没有页面。选择 Enron 语料库,是因为其邮件公开且历史可验证。该项目由 this+that 完成,使用原型。事实是有日期、有来源的句子,而非 CRM 风格字段,且无人编辑页面或事实。

一篇 Hacker News 帖子宣传了一个由 45,000 封 Enron 邮件生成的 AI 维基,每个事实都有引用。该帖子有 10 个积分和两条评论:一条评论 jokingly says AI 正在利用 Hacker News 向其他 AI 展示,另一条评论反思 Enron 时代的问责制,并将其与后来的公司和解作对比,以 Boeing 为例说明后果轻微。页面还包含标准的 Hacker News 导航、指南和 Y Combinator 申请通知。
相关文章

原文

We built Enron a company brain

We gave an AI 45,047 of Enron's emails, and it wrote 214 pages about the companies, regulators and banks Enron dealt with. Every fact cites the email it came from. When the mail shows a fact stopped being true, the page expires it on that date. No one edited a page or a fact.

Explore Dynegy →How we did it

An email in the corpus

·

this+that
reads it

Entity page Dynegy

CompetitorTrading counterparty

Relationship then:

✓ Checked against the email
  1. Find the types
  2. Find the entities
  3. Build the page
  4. Verify

214pages

49,401facts, each citing its email

93%of sampled facts fully supported by their email

0of 500 sampled facts wrong at their core

The last two come from a held-out audit by a second AI; see how we checked it.

Where to start

These are the most-discussed organizations in the corpus. Each page has a slider. Drag it and the page shows what was known on that date, with facts that were only true for a while expiring in front of you.

EntityWhat it was to EnronMessages
FERCThe federal regulator overseeing much of Enron's business, where Enron pushed for RTOs and open transmission access, opposed California price caps, and sought to narrow refunds.1,501
General ElectricGeneral Electric was an Enron turbine supplier whose consent governed turbine contract assignments, a Dabhol co-investor, and in late 2001 a rumored rescuer and then bankruptcy creditor.1,499
Pacific Gas and ElectricCalifornia's largest utility: a gas and power trading counterparty and retail-market incumbent whose 2001 Chapter 11 left Enron a $404 million creditor on its unsecured creditors' committee.773
DynegyHouston rival and one of Enron's largest trading counterparties that agreed to acquire Enron in November 2001, then terminated the merger and claimed its Northern Natural Gas pipeline.593
El Paso CorporationA Houston competitor and top-ten Enron trading counterparty that bought Enron's Cogen interests, bid for other Enron assets, and restricted its trading with Enron in November 2001.478
California Public Utilities CommissionCalifornia's electricity and gas regulator, which subpoenaed Enron's records from 2000, received its PX credit complaint, and suspended direct access in September 2001 over Enron's opposition.477
Duke EnergyDuke Energy was an Enron rival and one of its largest trading counterparties, a co-defendant in California price-manipulation suits, a buyer of Enron assets, and a creditor in its bankruptcy.471
California ISOCalifornia's grid and spot-market operator: Enron's trading venue and counterparty, then its adversary over price caps and refund claims, and a debtor owing Enron tens of millions.460
Dow JonesDow Jones supplied Enron's custom news clipping service, worked with it on power indexes, and through the Wall Street Journal published the October 2001 Fastow partnership stories.418
Reliant EnergyHouston rival and top-ten trading counterparty that coordinated with Enron on California subpoenas and faced contempt alongside it, then limited and slowed its Enron trading in late 2001.400

Facts on these pages are sentences, each with a date and a source, rather than fields like headcount or revenue. Fields belong in a CRM or another system of record, and a brain like this one sits beside it, holding what changes too often or doesn't fit in a field. We describe the three kinds of company knowledge, and where each one lives, in The missing middle of company knowledge.

Browse by what it is

The pipeline proposed 57 narrow types and a person kept 54 of them. Almost all were kinds of organization, so we group these lists by a coarser class and keep the pipeline's own type as a tag on each entity.

The pipeline also found 78 products and platforms, 76 facilities and market points, and 49 events, listed without pages. Choosing which kinds of entity deserve a page is the setup a person does when configuring a company brain, and we made that choice before building any page. Every company's mail names things it doesn't need to track, in categories that are a bit of a mess: restaurants from catered events, airports from travel, movies mentioned in passing. Enron's mail is no different. The facilities include trading hubs, Enron Field, a parking garage and two tankers. The events mix business conferences with charity bike rides and Cirque du Soleil. The products mix software with an exchange, a statute and trade associations.

The right choice depends on the company. An event company should have pages for its events, a property manager for its facilities, and a hardware maker for its products. For Enron we gave pages to organizations named in at least 25 messages, and to the five most-discussed parts of Enron itself: three business units, a trading platform and a trading system. A real company brain wouldn't filter by count. A customer you signed today might have sent one or two emails, and you'd still want its page. The cutoff stands in for the judgment a person makes about what's worth keeping in the brain.

Who made this

We're this+that. We made these pages with a prototype of something we're building.

Why Enron

We wanted a corpus anyone can check our work against. Enron's mail is public because the Federal Energy Regulatory Commission (FERC) released it during its investigation, and Carnegie Mellon has maintained it since 2003. It's a large, free and widely used body of a real company's internal correspondence. The company's history is also settled, so a page claiming, for example, that Dynegy agreed to acquire Enron in November 2001 can be checked against the record rather than taken on trust.

The mail shows the story in pieces. For the whole of it, read The Smartest Guys in the Room by Bethany McLean and Peter Elkind, who covered Enron's rise and fall as senior writers at Fortune.

Generated October 23, 2026 from the Enron corpus, which FERC released and Carnegie Mellon has maintained since 2003. Eight mailboxes, kean-s, mann-k, jones-t, kitchen-l, beck-s, sanders-r, williams-w3, lenhart-m. Every fact on every page comes from one message and quotes a sentence in it, and we discarded any fact whose sentence we couldn't find word for word. We mask phone numbers and credentials.
联系我们 contact @ memedata.com