According to Paul Graham, The distinguishing feature of nasty little problems is that you don't learn anything from them. ... Working on nasty little problems makes you stupid. And this, in turn, makes you more qualified to work on more nasty little problems, resulting in an infinite regression.