Summary
Google reportedly bid $10 million for Spirit Airlines’ old work emails, and this video from Nate B Jones asks why a company with so much data wants records of ordinary office work. The answer lies in a leading theory among AI labs: that years of Slack messages, call transcripts, and documents can teach AI agents to perform knowledge work.
Jones explains the shift from verifiable work, such as code that either runs or doesn’t, to murkier domains where it is hard to prove what made a document good. Using an invoice-reconciliation example, he argues that work records capture coordination and messaging but often miss the judgment, tribal knowledge, and authority that actually get the work done.
The video also examines who benefits. Companies selling the data and brokers take a fee, buyers hope to resell agentic capabilities, and the employees who generated the records get no seat at the table. Jones notes the data often comes from failing firms, raising questions about quality, and describes how workers, including the Spirit flight attendants’ union, have challenged the deal over confidential information. He closes with what this means for how workers should value their own output.
📺 Source: AI News & Strategy Daily | Nate B Jones · Published October 08, 2026
🏷️ Format: Opinion Editorial
