Some even go as far as saying this makes them useless… But that’s not what they’re designed to do.
Large Language Models (LLMs) like don’t “think” in the way we do, even if they give the impression that they do.
They don’t calculate, they predict. Unless they’ve seen that exact maths question before, with a known correct answer, then they’re essentially guessing the outcome based on patterns in their training data.
It’s easy to assume these tools are some kind of magic bullet. But really, they’re just massive text prediction engines trained to respond to plain-language prompts by referencing training data that they’ve seen before.
They’re brilliant at simulating conversation, summarising content, or generating ideas… But not so great when .
So if you’re trying to check your expenses? You’re better off with a calculator, a spreadsheet… or even asking 😄
I’m Craig Pickles, a web designer, developer, technical SEO and social media specialist.