Post #1822804

@wbftw @ekuber Something something something Ivan Illich's anarchistic theory of knowledge and pedagogy as espoused in Tools for Conviviality. I think the point is: LLMs aren't a "convivial" tool by nature, even when used just to write code; they rob practitioners of their agency, though oneshotting the boilerplate may be fine. Even their use in INFOSEC is questionable vs more traditional deep-learning approaches and static analysis. "I don't think that the real violence has even started yet."