LISP was once a language of choice for computer science research, specifically in the area of artificial intelligence. And so a few LISP startups were created in the 1980s with great optimism, the most significant of them being Symbolics, where I drank too much of the Kool-Aid. But in the 1990s, AI-oriented languages and paradigms -- such as LISP -- were replaced by more general object-oriented ones; beautiful and powerful though it was, LISP seemed slow and abstruse, destined for the scrapheap of history. But that's not quite how things turned out.
Dan Weinreb was one of the founders of Symbolics, as well as of object-oriented DBMS pioneer Object Design (now absorbed into Progress Software). He's working again in LISP development, at a very surprising place -- ITA Software, which makes airline reservations systems, one of the most transactional applications of all. (TP monitors were pretty much invented, 40 years ago, for airline reservation systems.) Beyond his day job, Dan writes an excellent blog, with a small number of long, insightful posts. One of those, back in April, reported on a LISP-specific conference, and detailed just how alive LISP still is.
If you have a really complex problem to solve, Common LISP might be an interesting language to try.