AI is reshaping the DB stack, and the latest chatter centers on Erdos—the open-source AI data science IDE from Lotas with built-in SQL integration. It's a data-science-friendly fork of VS Code designed for Python, R, SQL, and Julia, with an AI that can search, read, and write across those languages [1].
Erdos bundles a host of features: built-in consoles for Python, R, and Julia; a plot pane; a database pane for connecting and manipulating SQL sources; an environment pane; and a help pane that surfaces language docs. It supports remote development via SSH or containers and offers a zero-data-retention option or a local model for guidance [1].
On the DBA side, a separate post describes an AI agent built with Mistral that automates about 80% of PostgreSQL tasks via Postgres MCP—a secure bridge running in Docker. The agent analyzes pgstatstatements outputs and returns executable optimization steps, even simulating an index with hypopg without locking tables [2].
Meanwhile, a CMU team touts a vector-based approach that could turbocharge PostgreSQL [3], signaling a future where AI acceleration sits alongside traditional tuning. It’s early, but the trend is clear: AI tooling is inching into development, debugging, and performance tuning for PostgreSQL and beyond. Watch this space as AI tooling reshapes DB dev.
References
Show HN: Erdos – open-source, AI data science IDE
Show HN introduces Erdos, open-source AI data science IDE with SQL integration, multi-language support, and AI-assisted workflows.
View sourceAn open-source AI agent (Mistral) connects to PostgreSQL via MCP to diagnose slow queries and propose index optimizations in seconds.
View sourceCMU team claims vector-based system can turbocharge PostgreSQL
CMU team claims a vector-based system can turbocharge PostgreSQL, suggesting performance gains for database queries.
View source