In Defense of Doing Your Own Research / hacker news