Built to bend: Postgres’s inventor and a steward on what AI really needs
Michael Stonebraker and Bruce Momjian discuss the intersection of AI and databases. Stonebraker notes that while benchmarks claim high accuracy for text-to-SQL, his own tests on MIT's data warehouse yielded zero percent accuracy initially, highlighting the difficulty of applying AI to idiosyncratic enterprise data.
Key points
- Postgres creator critiques text-to-SQL AI
- MIT data warehouse tests showed zero percent initial accuracy
- Argues for systems that are community-controlled and flexible
Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. The Register did the reporting; read the original for the full story.
Read the full story at The Register
