On top of those there are "recent research" for any problems I am currently working on, which is usually at least one window each with a search results tab and a few tabs I opened from there (after pruning the ones that were not relevant in the end.
Plus there are the tabs I am "actively" using (which is pretty much always at least five at the moment: two mail clients, two social networking sites, and a page containing the results of various service monitors which updates every minute or so).
I often see FF taking a bit more than 1Gb RAM particularly if I have complex pages that auto-update a lot (Zimbra, facebook, so forth) open, though often closing it and restarting with the "reload last session" option brings that down by at least 60% (I assume the "extra" is stuff cached in RAM and/or memory allocated and not yet released due to fragmentation in FF's internal memory management - both situations being "resolved" by the restart)
Maybe you don't feel bogged down but I feel bogged down for you :)
Similar usage with 3.x rarely exceeded 300MB and FF would remain open for weeks.
Tree-mode tabs help a lot (I'm still looking for a good Chrome plugin for this), "Tree Style Tabs" for FF5. Since trees group related tabs (I've got a subtree open for HN right now), and collapse, it helps to manage them.
As with grandparent poster, I'll have several windows, across several desktops, task-oriented. News, mail, site monitoring, research, reference(s) for languages / tools, etc.
200 tabs is probably a high count, but well over 100 is very doable.
It's just a way of keeping the 'flow' of a browsing session in the browser itself, so you don't have to keep all that in (mental) memory.
I do use Firefox, which works fine even in my 2GB laptop and with a bunch of extensions installed.
I've also switched back from Chrome to Firefox because Chrome doesn't scale (nor does it have all the extensions).
That's just one use case, but there are certainly many others.
I just dislike opening images inline or all loaded into one separate tab, because then it turns saving individual images into a mouse-driven task rather than a keyboard-driven task.
As an aside: what do you use to browse Hacker News on your phone?