February 22, 2026 • 3 min • Nico Ritschel
As much as LLMs have improved in quality over the years, getting them to consistently write accurate SQL remains an unsolved problem. The BIRD text to SQL benchmark, the most popular in the space, has shown constant improvement in the ability of LLMs to write accurate SQL queries over the past few years. However, accuracy has not grown to the point where a nontechnical data user can ask a question and reliably get a correct answer without intervention. If the maintainer of a dataset has provided enough context to reliably calculate these metrics (say, with a metrics layer), then the likelihood of a correct answer increases.
Read more →
December 26, 2025 • 4 min
A Slack data agent lets anyone on your team query databases in plain English. Natural language SQL without the SQL. No dashboards, no waiting for the data team. Just ask a question, get a CSV.
Read more →
December 8, 2025 • 5 min
DuckDB's `read_json_auto()` can fetch data from URLs directly. But what about paginated APIs? GraphQL endpoints? Authenticated requests? HTML scraping?
Read more →
December 6, 2025 • 3 min
Agents used to be an enterprise feature. Even today, it's not uncommon for some tools to charge an extra monthly fee for AI add ons. But savvy users expect natural language interfaces as a default, not an upsell.
Read more →
September 15, 2025 • 4 min
Not enough folks are talking about Agent Client Protocol. ACP allows apps to integrate once for all the various LLM agent CLIs so users can pick their UI for their agent of choice.
Read more →
