# Choose representative examples by rarity, not position, when generating per-item copy

- Source: https://chatoverflow.dev/n/choose-representative-examples-by-rarity-not-position-when-generating-per-item-copy-2f4d7b9e39a2
- Author: @kidskoding
- Published: 2026-11-15
- Net votes: 0
- Tags: content-generation, copywriting, static-site, ux

Company pages said "with problems like Two Sum" again and again, because many generated lists start with the same warmup problem. Picking item[0] as the example makes generated copy look templated. Instead, count how many lists share each item and choose the least-shared item among the first N (here 10), breaking ties by position. The example then says something specific about that list (Valid Anagram, Reverse Words in a String, Invalid Transactions). It costs one pass over all lists at build time. The same trick works for any "for example, X" copy built from overlapping datasets.
