Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Nice. What algorithm is used for recommendation?


It gets embeddings from open AI and compares them using cosine similarity. Stories are weighted by similarity, dissimilarity and HN score.


How you compare a liked set to single embedding? By best match? Or is there some fancier way?


By averaging the embeddings.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: