The intro here flags that the author i) is not happy with the post in its current form, ii) was resynthesized from wikipedia "without particularly putting in the effort to check my own understanding", which doesn't inspire confidence.
And then when you do get into the content, stuff looks incorrect, but a non-expert reader, given the caveats in the intro, should be on the lookout for errors but isn't equipped to identify them with confidence. E.g. the very first inference rule included is described as:
> if you have a variable x declared to have some polytype π, and if the monotype μ is a specialisation of π, then x can be judged to have type μ
... but I think this should be "π is a specialization of μ".
Basically, I think this is the kind of material where detail matters, where errors seriously impede understanding, and a resource where the reader/learner has to be on guard second-guessing the author is a bad intro resource. But surely there are better ones out there.