Leveraging AI as a PostgreSQL DBA
summary
Demonstrations use free-chat models (OpenAI derivatives, Google Gemini, Microsoft Copilot, Brave, u.com) and show concrete results: generated schema and runnable CREATE statements for an amateur‑radio example, bulk INSERT test data, and a test suite that checks FK and row counts. AI mislabels backups as MySQL when prompts are vague and invents a non‑existent deleteMeNow function; a bad query plan drops cost from 3929 to 4 after rewriting. Emphasis on perplexity as best performer and the rule trust but verify.
AI-written summary · about
chapters (14)
- 0:00Introduction and goals
- 1:44AI as a DBA tool
- 3:07Mastering AI prompts
- 7:07Demo: Bad backup script
- 10:20Demo: Hallucinated function
- 13:12Evaluating AI tools
- 15:00AI for learning and coding
- 17:51Demo: Database design
- 20:38Demo: Generating data
- 22:49Demo: Test scripts
- 25:17Demo: Writing functions
- 27:46Demo: Query tuning
- 30:58Final tips and tricks
- 33:29Summary and Q&A
related talks
27mPutting AI in Production
48mMastering the AI Shift: Adaptive Strategies for Software Evolution
20mOpen Source Database Operations in the Age of AI
25mPostgres and the Artificial Intelligence Landscape
01h 00mPostgreSQL and Artificial intelligence
45mForget SQL! Create your own AI Data Analyst
45mPostgres and the Artificial Intelligence
46mPower the next generation of AI apps with databases at scale anywhere