To extend the software analogy, asking mathematicians to decipher LLM-generated math results is like asking programmers to understand computer programs using a decompiler instead of starting from human-comprehensible source code.
To extend the software analogy, asking mathematicians to decipher LLM-generated math results is like asking programmers to understand computer programs using a decompiler instead of starting from human-comprehensible source code.