Yet another interesting comment from afar, this time from afar.
My thoughts (seem to be inline with the creators thoughts):
- This is a good learning moment.
- It costs about a half-a-buck per search; an embedding architecture would be about 1/60th of a penny.
- It is slow; search should be Doherty threshold or less.
- It replicates a lot of information in adjacent fields.
- It’s an impressive array of steps; a Make posterchild for complex solutions.
- Proof that if it can be done, it should not necessarily be done.
- Embeddings; 'nuf said.