Toying/Vive Coding with the idea or algorithm to classify the tendencies, ideological biases, and sentiment from contents of posts on message forums, including HN, by using LLMs to measure the temperatures and intensities of posters and replies to posts. Also thinking of incorporating ability to identify potential duplicate accounts, (e.g. same user multiple accounts) based on similar language and grammatical usage in posts. Don't see much of that type of detection in forums or message boards. Perhaps it can be useful in law enforcement circles. It's a work in progress, though, no repos yet.