that's a bit funny assumption that "LLM==data stealing" - 0) how
corporations would steal data from local models, and 1) how it's even
related if the context is claiming that using tool ABC (20y ago syntax
highlight, now LLM) makes the code by definition (of not so-smart
snobs/elitists) to be bad quality.
ironically the structure of your argument strongly remind chatgpt 5.0+ if
users would try to argue its leftist system prompt - it also instead of
seeing main logic via reasoning (as 'uncensored' llms or rastafari would
do), it instead starting trying hard to narratively 'preserve' its flawed
and rigid 'narrative' to attack the small details and blindly ignore big
patterns ("we can't compare X and Y, because even if pattern is the same -
it would lead to conclusion which conflicts with my propaganda books, so
let's derail it into comparing types of paint used on X and Y")
