Just brute force your embeddings

1 comments

• Ellis_dev an hour 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.