The Post-AI Data Stack: Data Teams' New Job Is Guarding Company-Wide Reality

le_james94 · x · 2026-09-05

Ian Macomber, data engineer at Ramp, argues the Modern Data Stack is over and a "Post-AI Data Stack" has emerged. He traces the evolution: pre-modern stacks (2013) with small structured data, cloud warehouses (2016) expanding scope to all company data, and Modern Data Stack + Reverse ETL (2020) turning data scientists from reporters into operators.

The key shift: AI makes producing analysis cheap, but it doesn't make agreeing on reality cheap. As coding agents and Slackbot analysts let anyone answer their own questions, the scarce resource becomes company-wide consensus. Data teams now have two jobs: enable everyone to build with data and AI accurately and independently, and champion the singular reality the company operates on — encoding expert judgment into infrastructure so agents produce correct analysis without a data scientist in the room.

Original post →

More from AGI Musings

AGI Musings channel →