Yet, we believe that an LLM is a good-enough measure of semantic ambiguity for our needs.

AI and believers, name a more iconic duo

On some days I wish back the ASM times. 3 chars that’s enough for a variable name. All those modern 4 char names … Pah … /S

Cool idea but Im pretty sure you can extremely optimize the ambiguity calculation by just vectorizing the terms and looking them up in a vector database.