I have had two papers published with IEEE when I was an undergrad. There is no provision for me, or any of my co-authors, to be paid by IEEE. In fact, I was required to pay steep conference registration and publication fees, and also a fee for an extra page on one of the papers, to have the papers published by IEEE. I recall that some conferences also actually require you to attend the conference in order to have the paper published, so, add mandatory hotel accommodations and an airline ticket to that as well (although I'm not complaining, since I got plenty of support from my universities, and conference trips are always a lot of fun).
The way the paper publishing world works is, after and if your work is accepted by a conference, you pay various registration fees and publication fees, transfer your copyright to the publishing house like IEEE (there are some tricks around this, like giving up the copyright instead, by placing your paper into the public domain), or, as is the case with some other publishers, give them a non-exclusive irrevocable royalty-free licence, and then they publish your work on their web-site (with a right of collecting fees from web-site users, too), as well as possibly into a printed proceedings of the conference that's given to all attendees (sometimes for an extra fee, too).
Many universities in the US, Canada, England and other contries, pay IEEE some kind of subscription fees, so, anyone on the university network is automatically given unlimited free access to all such papers that are published by IEEE. Otherwise, if you're using a home connection, or your university lacks any such agreement with IEEE, then IEEE collects individual fees for every paper directly from end users (and gives none of it to the authors of said papers).
As mentioned by other answers, many authors also place a copy of their papers on their own web-site. This is often done illegally, since they often no longer own the copyright to such papers, so, depending on the circumstances, IEEE and such can potentially resort to legal methods to enforce its copyright against the illegal copies. For practical reasons and bad-publicity considerations, this is not actively done in reality. It is also the case that after having the copyright assigned to itself, IEEE and other publishers generally give some kind of non-exclusive licence back to the authors with some limited rights on what could be done with an exact copy of the paper. IANAL, but I think the language of such licence is generally restrictive enough that you're not actually supposed to provide the very same copy of the paper elsewhere to the general public in an unrestricted access.
In contrast, these academic conferences are different to the technical conferences around open-source software. With technical conferences, there is a lot of members of the general public who want to see the presentation of the author, and such members of the public pay modest registration fees, which, when taken together, together with some contributions by big-name sponsors, are enough not only to support the web-site with free access to all the resources, and potentially a publication of the paper proceedings, but are also often sufficient enough to even cover the airfare and other travel expenses of the authors.