Usenet was only predominant for about ten years. But programmers have been rejecting Lisp and the ideas behind it for nearly fifty. No matter how much we like it, it's not going to win. Think of it as your secret weapon à la pg.
Maybe you're young, but Usenet was the center of mass of programming discussions from shortly after its inception, in the early 1980s, until sometime around 2000–2005. So for about half of Lisp's lifetime, Usenet was central. And, as someone else pointed out, you can see a major difference in the tone of comp.lang.lisp before and after Erik's ascendancy — and its contributors. He may not have been the sole cause, but he was certainly one of them.
I think it's more accurate to say that programmers have been slowly accepting the ideas behind Lisp for fifty years: linked lists, recursion, dynamic typing, flexibly-sized data structures, conditional expressions, garbage collection, and a small core language with everything else in libraries were all present in (and central to!) 1959 Lisp, and very far from the mainstream in 1959. The main features of 1959 Lisp that have not been widely adopted are simple syntax and atoms. (Macros came later.)
For most of the time between then and now, new language features have tended to appear early on in Lisps and much later in more mainstream languages. Compare Python or Haskell list comprehensions with the Common Lisp LOOP macro, for example; or consider Loops/Flavors/CLOS.
I essentially don't use Lisp today because Lisp systems fell badly behind other languages in many ways: Python, Ruby, Lua, Objective-C, OCaml, and so forth are better options today. There's no mystery to why this is: people stopped working on Lisp systems. That was partly due to the AI winter, but it actually worsened in the 1990s. I don't know how much of that was because of Erik driving people away and how much was for other reasons.