I clustered and cataloged the archives of the NYRB and LRB with OpenAI embeddings. The substack is intended to highlight categories I think are good / interesting. There's a brief technical write-up as well.