The claims made aren't all that falsifiable (like anything else in psychology, it isn't ever possible to know to what extent a person is lying to themselves), a

The claims made aren't all that falsifiable (like anything else in psychology, it isn't ever possible to know to what extent a person is lying to themselves), and the specificity of the claims is something of a red herring; when 90% of a list of specific claims appear to describe a subject to at least some extent, if they want to believe that the system works, they're going to ignore the other 10%. The medium-sized fraction of the audience who fit at least that closely into the categorization system trumpet it around as a success, making those who fit somewhat less well more inclined to accept the errors and inaccuracies of their classification. These effects, plus the handful of results related in a very obvious way to the original categorization (emotional coping resources are often defined in such a way that the INTP/ENTP result you mention is entirely obvious, for example), make it appear that the MBTI means much more than it actually does. Much like an IQ test, used professionally, as part of a deeper analysis of a subject, the MBTI can help a psychologist or psychiatrist better understand a subject/patient. Used recreationally, such things lead to faulty generalizations unless treated as they are--glorified horoscopes geared toward looking at particular aspects of ourselves.