Controls engineer here. The author’s control theory knowledge is correct - a RHP zero indicates that corrective action will begin in the wrong direction (which the article compares well with countersteering on a bike). The technical term is a “non minimum phase system”. It’s possible this will lead to instability, but in general the long term stability of a system is determined by the location of its poles (which is more applicable to the author’s ice cream example). Poles in the RHP will cause a system to blow up.

My gripe with the article is that the author tries to wow you with some obscure technical points about a system which is unmodeled and he does not understand, to wave his hands at a vague conclusion. If he had made the same point using common English phrases that encapsulate the idea (“positive feedback loop”, “we have to let it get worse before it gets better”, etc), then it would be a lot clearer how wispy his argument is.