Just brute force your embeddings

1 comments

• $Ellis_dev 9 hours ago

  The 1M-document numbers are a great reality check. I’d happily start with the NumPy version and only reach for a vector DB once it actually hurts.