Of course, you could argue that Perl has even richer libraries, so...
Ruby's domain is primarily the web and web systems, and when it comes to that area, Ruby has an vibrant ecosystem with plenty of libraries and tools that don't have python counterparts. Ruby developers have also readily borrowed good ideas from Python. For vast majority of web development, these kinds of things matter more than the existence of something like scipy. Each language has its strengths and weaknesses with regard to the quality and quantity of tools directly related to web development.
I think if needed you could change the framework code with django also but it just seems very large where as pylons or cherrypy is very minimilistic.
Also, Google App Engine kind of fits the bill (with a little finesse).
Fortunatelly it's ecosystem grows fast and into the right direction. On posts like which language to learn -- Django or Rails -- one can realize Ruby community is well equipped with tutorials, screencasts, components, news, collaboration, hosting and deployment/tuning tools
This really exposes how different this crowd is from the mainstream.
There are a lot of companies which simply never talk about what they use internally; as such, it's really hard to get an accurate picture of what's really in use, because the only glimpses we get are the (fairly rare) places which choose to be open about what they're doing.
Based on that, and on interesting tidbits picked up from talking to people in the trenches and paying attention to little clues (interesting addresses showing up on mailing lists or IRC, peculiar code quirks, etc.), I suspect that if everybody laid their cards on the table the results would be surprising to people who think they know what the "mainstream" is using (and not just in the "lots of people using Rails/Django sense -- there are many more exotic things out there).
What I'm objecting to is your apparent assumption that use of tools like Django or Rails necessarily sets one apart from "the mainstream", or even that there is such a thing and that it can be quantified.
Or, succinctly: there probably are a lot more people using not-PHP and not-Java than you think. Some are using hip, trendy new things like Django or Rails. Many are still using ancient, forgotten things like COBOL and IBM mainframes (some friends of mine deal with those types of folks on a daily basis). And there are a lot of places that aren't single-language shops, but have an amalgamation of code accreted over the years.
Which all adds up to the fact that you really can't make useful statements about "the mainstream"; all such statements tend, ultimately, to reflect hype machines rather than reality.
EDIT: Doh! It's too late at night, and I didn't catch what you meant in the title. My startup is bootstrapped, not YC-funded.
PHP is still strong because of it's stable ecosystem. Django must be at 1st place because of it's power and no-fluff media coverage.
Glad to see Rails in this position inspite the negative hype around it. Also the 37signals approach won supporters in a HN climate determined by PG who has his own philosophy.
Wonder who might be the 'others'? Haskell, Clojure, Erlang and maybe Arc?
It's all down to the tools for me.
That being said webforms suck man you can use MVC and still have the nice tools.