Code-first products present an unreasonably effective interface to AI models, because AIs excel at generating the very code these products run on. Today's release of Rill 0.41 introduces a related, powerful side effect of our BI-as-code philosophy: You can now create an operational dashboard from any data source with just one click, thanks to an integration with GPT-3.5 Turbo. rilldata.com/blog/one-click-… At the core of Rill dashboards are SQL aggregate expressions, which define metrics (like revenue, MAU, or ARPU) over a set of dimensions (geography, age cohort, or device type). Given a prompt that includes the columns of a data source, GPT-3.5 readily generates SQL & YAML code for a list of relevant metrics and dimensions, along with informative labels, that underpin Rill dashboards. Incorporating code generation into development environments like Rill Developer is not a gimmick, but a true productivity accelerator, and we're confident it will further speed up the data to dashboard journey for our users.

We built an interactive exploratory dashboard of the @duckdb code repository on Github... powered by @duckdb :). Amazing to see the project's acceleration, almost 300 committers have made 29k+ commits, with @mraasveldt @Thijsbruineman, (naturally) @hfmuehleisen, @lnkuiper, and @holanda_pe leading the pack. ui.rilldata.com/demo/rill-gi…

"SQL + YAML = dashboard" officially launched today by @rilldata. A BI-as-code stack controllable via git, powered by @duckdb. Ask and ye shall receive @josh_wills. docs.rilldata.com/notes/0.16