Two flavors of 'predict the output'
Nearly every practical supervised learning problem falls into one of two buckets. Classification predicts a category from a fixed set of possibilities, like spam versus not-spam, or which of five product categories a listing belongs to. Regression predicts a continuous number that could, in principle, take any value in a range, like a house's sale price or tomorrow's expected temperature. The distinction sounds simple, but it's the first and most important decision in any supervised project, because it determines what kind of output the model produces and how you'll judge whether that output is any good.
It's worth noticing that some problems can look like either depending on how you frame them. Predicting 'will this customer churn' is classification (yes or no). Predicting 'how many days until this customer churns' is regression (a number). Same underlying business question, two entirely different modeling setups, because the shape of the answer is different.