Socionics is based on partial truths which is what makes it rather dangerous. I enjoy some of the ideas which inspired it such as Jung's types and Kępiński's theory of information metabolism, but I don't think combining those two theories result in socionics in any meaningful way. I also find Jung somewhat questionable compared to the ideas which inspired Jung himself, and Kępiński seemed to have plenty of questionable ideas too, only that they seemed to have been significantly fewer and less horrible than Jung's.
However, some of the ideas of socionics do lead directly to what I see as being the more relevant, more interesting observations.
One, I think intuition has the characteristics of a "rational" function and not feeling. Socionics, however, has a dichotomy for this: abstract vs. involved, so I will use this word instead of "rational."
Ne = Sanguine temperament = air (hot/wet) = "Ej temperament" = dialectical-algorithmic
Ti = Choleric temperament = fire (hot/dry) = "Ij temperament" = causal-determinist
Fe = Phlegmatic temperament = water (cold/wet) = "Ep temperament" = holographic-panoramic
Si = Melancholic temperament = earth (cold/dry) = "Ip temperament" = vortical-synergistic
If you can't see here, the "wet" elements are the extravert ones vs. "dry" introvert, and the "hot" ones are abstract vs. "cold" involved. If you find this questionable, show me what's wrong. You won't be able to, because Fe is not choleric and Si is certainly not choleric. This is how a real hypothesis that makes counterintuitive-seeming predictions looks, but all these predictions hold up to scrutiny because I based them on evidence to begin with. Extraverts are the intuitive ones and not introverts, additionally, people stuck in their own little world are generally stuck in the senses.
What I would really like to question, however, is the idea no type is better than any other and that types don't change. Intuition, which overwhelmingly appears to be an extraverted element rather than an introverted one, seems to be the best, and sensing seems to be the worst. Additionally, intuition shows all the attributes of "rationality" much more than feeling does. The four humors theory related the four humors to health. You want to be sanguine and choleric is not too bad, while you don't want to be melancholic or even phlegmatic. The four humors clearly don't seem to cause the phenomenon attributed to them, but the connection between the mind and the body seems to hold, in addition to the idea of there being four of them which share certain attributes and differ in other ways. Something like the cognitive styles is especially modern, but Gulenko did not invent those, either, he simply decided to apply them to Reinin dichotomies. I think Jung's agenda really distorted things since Jung was trying to basically prove that things that were traditionally viewed as unfavorable weren't so bad, and other people's agendas after him distorted things as well.
This has been the reality of socionics. It's wrong, but it's not quite complete nonsense, and that's what makes it such a problem.