I can point you to an example where a book found a real publisher, because it did fairly well on Lulu and was written by someone with an internet following:

http://www.amazon.com/Seagalogy-Study-Ass-Kicking-Steven-Seagal/dp/1845769279/ref=pd_bbs_1?ie=UTF8&s=books&qid=1205436229&sr=8-1

All he had to do was pull the Lulu edition. I would think the overlap between your 500 page technical work and a popular book wouldn't be much greater than that between Freakonomics and Levitt's journal articles, and I don't think he lost too many sales because people were running to their libraries to access NBER.