This is a limitation of the training data. If you were uncertain about something, you wouldn’t write a book about it. The kinds of people you’re talking about tend to generate far more text in their lives than others, because they can spend more time generating - writing books, blogposts, whatever - and less time thinking and working and actually doing things. The models never say they’re uncertain because we never say we’re uncertain, or at least we don’t write it down anywhere.
If you change it from asking a question to giving an instruction, how many humans do you know that have trouble saying no to things that aren't reasonable? I'd argue that pretty much every human will refuse to do most things you might instruct them to do, whereas an LLM will happily attempt most things you ask them to do for you, regardless of whether they're capable of succeeding, and it's up to you to figure out if they actually did it right or not. There are tasks where this is extremely useful, but they're ones that are extremely low risk and can easily be audited upon completion. This isn't anywhere near the level of what a human is capable of.