Every time the claim is made that MBTI is pseudoscience, comparable to astrology, meaningless, etc. Sometimes this claim is accompanied with mentions or links to other psychometric assays to use instead.
And every time the same two very rare MBTI types end up being the majority of HN readers. Which are already known as being the two most common types among software developers, something that has been a consistent result for decades.
(Side Note: the linked to humanmetrics web quiz correctly notes that it is not the same as the MBTI. No information is provided as to its validity or to how well it correlates with MBTI types that it uses the names of. Therefore, observations that the humanmetrics web quiz (or any other similar free web quiz/personality test/etc) does not provide consistent results over time for some people may not necessarily indicate that the same effect would be observed taking the MBTI assay.)