diff --git a/docs/src/content/docs/reference/_sql.mdx b/docs/src/content/docs/reference/_sql.mdx
deleted file mode 100644
index c763286d..00000000
--- a/docs/src/content/docs/reference/_sql.mdx
+++ /dev/null
@@ -1,74 +0,0 @@
----
-title: SQL
-description: A short intro to SQL.
----
-
-SQL is a functional general-purpose query language for relational databases. It
-is as old as relational databases themselves.
-It allows to define arbitrary data look-ups and computations w/o hard-coding a
-specific database implementation or storage structure (presence of indexes,
-storage layout, ...).
-Yet many foster a love-hate relationship with it, which has given rise to an
-entire cottage industry of ORMs: higher-level, often type-safe abstractions
-with bindings for your favorite programming language.
-They tend to work great until you need to break glass. In recent years, there
-has been a push to thinner and thinner abstractions.
-
-Instead of hiding or working around SQL, TrailBase embraces it as an evergreen,
-transferable skill.
-There's no denying that SQL can get tricky but it's
-also its greatest strength. SQL is a general purpose functional (sometimes
-imperative) programming language that lets you solve arbitrary problems in a
-high-level, portable fashion. In other words, learning SQL is a lot more useful
-than learning a specific ORM or similar abstractions.
-
-One thing that's pretty sweet about SQL is that, almost any relational database
-(Postgres, MySQL, MS SQL, sqlite, ...) supports some dialect of SQL, which
-makes this a pretty transferable skill. This is only possible because SQL is
-pretty abstract. Given a lookup or transformation, different databases might
-execute them quite differently depending on their capabilities, how the data
-structures are set up, and how the look is expressed. All of this is facilitate
-through the magic of the query-planner, which takes your query and turns it
-into a physical execution plan.
-
-## A gentle introduction
-
-At first glance, SQL may look just like some antiquated filter language but in
-practice it's a high-level, functional programming language that is optimized
-for table transformations. All that SQLs does is:
-
-1. inserting rows (INSERT),
-2. removing rows (DELETE),
-3. updating rows (UPDATE),
-4. and transforming NxM tables into PxQ tables (SELECT).
-
-5. to 3. are pretty simple, they operate on rows.
-
-SELECT is by far the most complex tool, think of it as writing a function that
-takes in one or more table, even combined tables, filters them, and them
-transforms them into a new table. They might look overwhelming at first:
-
-```sql
-SELECT
-  Output: outcol0, outcol1, ..
-FROM
-  Input: incol0, incol1,
-WHERE Filter: e.g. where incol0 > 10
-ORDER BY outcol0 ASC
-GROUP BY aggregation
-;
-```
-
-but most of it is optional. The simplest statement is `SELECT 1;`, it creates a
-new table with one row and one column containing the value of `1` (a scalar)
-from nothing.
-
-Now that we can create a new column from thin air, let's transform it:
-
-```sql
-SELECT col*2 FROM (SELECT 1 as col);
-```
-
-Now we created two nested SELECTs, inner one creates the table we already know
-but names the column `col` and the outer SELECT transforms that table into a
-new 1x1 table with the values doubled.