Going down the rabbit hole of Postgres 18 features (xata.io)

Interesting and comprehensive article.

Note to the author : on mobile, none of the examples are visible.

I think the author may have misunderstood the redundant GROUP BY columns with UNIQUE indexes. If you group by columns (A, B, C) and there is a UNIQUE index on (A, B), you can remove C (all parts not in the unique key), as there will only ever be at most one value of C due to the unique key constraint.