Hacker News new | ask | show | jobs
by treprinum 1002 days ago
This approach would remove one of the main benefits - the ability to run multi-task one-shot prompts where a single LLM call returns answers to multiple NLP tasks.