Suppose you've never tried/done/seen/tasted X before. You want to predict how well you would rate/enjoy X. (X could be as small as a restaurant, or as big as a career choice.)
My knee-jerk way to approach this would be "(read about and) imagine doing X, see how much I like the thing I'm picturing, and form my prediction based on that." This approach seems to be subject to availability bias. Furthermore, this approach seems to take the inside view on prediction. For these reasons, I'm concerned that my knee-jerk approach will produce bad predictions.
On the other hand, how much you'll like something seems to depend more on 'inside' factors than whether you'll finish a project by a certain date. Does it make sense to throw out introspection entirely, and predict how much you'll like X based only on surveys/star ratings of X?