ChatGPT seems to have difficulty accurately solving for the degree of an angle in a right-angle triangle when given equations in the form similar to a=cos^-1(4/7), a=sin^-1(4/7), or a=tan^-1(4/7). The answers it calculates are always slightly off.
Correct. ChatGPT is a generative AI which means it generates text based on a large language model (LLM).
This roughly means that, despite the hype, ChatGPT has no domain knowledge or “understanding”, it is simply a very powerful and very impressive “auto-completion” engine.
Yes, because ChatGPT is not “calculating” it is simply “generating text” because it is based on a language model, not a computing model or expert system.