Show HN: We had AI write a wiki from 45k Enron emails, every fact cited (www.thisandthat.chat)

An innovative project has emerged in the AI/ML community, showcasing an AI's capability to synthesize a comprehensive knowledge base from 45,047 Enron emails, resulting in a 214-page document detailing interactions with companies, regulators, and banks that Enron engaged with. Each page is uniquely tied to specific emails, allowing users to track the context and relevance of facts over time, with information expiring if it ceased to be accurate. This project not only highlights the potential for AI in curating and organizing vast amounts of corporate knowledge but also employs a second AI as an auditor to ensure the integrity of the information. This endeavor is significant as it demonstrates how AI can effectively generate real-time, fact-checked documentation from historical records, paving the way for applications in corporate knowledge management. The use of a clear temporal context for facts, sourced directly from a real-world dataset like the Enron emails—a publicly available treasure trove due to regulatory investigations—sets a strong precedent for utilizing AI in compliance and governance contexts. By categorizing entities based on frequency in internal emails, the project also reveals how AI can discern important relationships within corporate interactions, potentially transforming how businesses manage their institutional knowledge.