Hi all, Elias here.

I just launched here a llm-embedding based search service, but it completely flopped, didn't even make it to front-page Show HN. So sad, but it happens. (Here's the post: https://news.ycombinator.com/item?id=38011802

Anyway, here's a related but totally different use case: getting median score of embedding-similar post titles to pick a title for your next Show HN!

Let me know what you think!