There is a very pernicious idea in epistemology, called induction. It's an imaginary, physically impossible process through which, supposedly, justified general theories are created from observations. It's still popular with some philosophers. Others realise it does not work (it was refuted by Hume hundreds of years ago), then wonder how we can know anything, and get stuck on the Problem of Induction (solved by Karl Popper, who should be super famous, but isn't). And, normal people hold many inductive ideas as common sense, too.
The primary claim of induction is that a finite set of observations can be generalised into a true predictive theory. However, any finite set of observations is compatible with an infinite number of predictive theories.
To see this, just imagine a paper with dots (observations) on it. We're going to draw a line from left to right (with the flow of time), and it has to connect the dots. The line is a predictive function, that gives values at all the points, not just the dots. So, how many ways could we draw this line? Infinitely many (go way up or down or zigzag between points). What inductivists do is pick one (whichever one feels intuitively right to them), and declare it is what will happen next. And people with similar intuitions often listen...
If you want a real-world example, think about the sun. We know it will rise tomorrow because it is a good explanation of reality (via our physics). Not because we saw it rise yesterday (and the day before).
I tried to write an entry that would be more helpful to people who don't understand, and it didn't go well. I have doubts about how helpful this will be to most people. I can answer stuff in the comments section.