This is exactly what I’ve been thinking as I see LLMs being applied to all these complex problem domains. Humans did not conquer the world because our intelligence can solve every problem, we did it by using our intelligence to (1) break down complex problems into small, manageable pieces and (2) designing tools and machines that were exceptionally good at efficiently solving those subproblems.

The other recent example that comes to mind is the paper that explored the reasoning process used by LLMs to answer trivia questions like “Name a national capital whose letters can be rearranged to spell a common greeting in the language of a neighboring country.” (answer is Hanoi by the way)

The LLM responses show that they intuitively grasp the algorithm for answering such a question, but then they basically run the algorithm in their own thoughts (self-talk) which is horrendously inefficient.

Put differently, natural language reasoning is brilliant at turning the messiness of the real world into well-defined abstractions, but as soon as that is done it needs to hand off the task to a machine. For “solved” problems this might be a formally specified machine, but it could also be another class of model such as AlphaZero (along with a proper specification of the problem the “subcontractor” is to handle).