Early LLMs typically tried to do multiplication "in their head" by recall.

Now most LLMs do multiplication using a tool call to a programming language, akin to a person reaching for a calculator rather than relying on a learned table or working the problem out mentally.

The high level comparison between what LLMs do and what humans do" for this example is fairly parallel.

How long until LLMs are prompting LLMs to write a response to their user query?

Actually, this happens already in a modular way AFAIK…