This seems to be the classic dilemma of wading into the pool or diving head first into the deep end.
FWIW I find the Steven's wrapper code to be annoying as well and I have struggled to get at least some of the examples working on every platform I tried them on. That said the wrappers hide mostly trivial error handling that would have been repetitive, unenlightening, and taken up more space in an already large book. (Though in fairness the editors can throw out the last 1/4th of the book covering XTI to make room.)
Network programming is deep enough of a topic that you can alternate riding the surface waves and doing cannonballs. Several good sources for a quick start have been mentioned that will get you past the "wow, I got it to work" stage. But when you need to ascend to the "I need it to be robust and reliable because my job/product depends on it" stage you will come to appreciate Stevens, warts and all, for his uncanny ability to anticipate your current problem.
There have been chapters of Steven's work I successfully ignored for years. But I was sure grateful they were there when I needed them.