AI

The useful question with an AI system is what it can actually do, where it fails and how its output is checked. These posts describe concrete experiments and tools rather than a list of model announcements.

Documenting source databases with LLM agents explores code investigation and human validation. Soufflé takes a different route: private, local speech transcription on macOS. Both involve trade-offs that matter beyond the demo.

Documenting a Source Database with Multi-Agent LLMs

6 February 2026·Updated: 27 October 2026·1488 words·7 mins

Data Engineering AI dbt LLM Data Engineering AI

Documenting columns in a source database is the kind of task nobody wants to do. You have an operational system with hundreds of tables, thousands of columns, and documentation ranging from “nonexistent” to “a 2017 comment that says TODO: document this.”

↑