. on Nostr: Oh that regardless of llm selection optimization and pay/per query it is still heavy ...Oh that regardless of llm selection optimization and pay/per query it is still heavy compute/token usage?