Yes, as long as context size increase and llm improve at least there's a way out through using AI but once the progress stops...