That is my hope exactly.

In this particular case, I agree with your assessment, and not arguing
the case.  Just got scared out of my mind, because we, and other
sites, depend on this so much.

However, let me make another case here for s's and g's.

For some of these SERP's that have "search result" pages listed, I
think those pages ARE relevant and helpful.  I mean, for my
theoretical shoes site, let's say I'm tracking the most popular terms
a particular product comes under, and create a link to those search
results.  Now, I have a page that is a search result, but still highly
relevant, because it's all about items that directly relate to the
term.

For instance, searching for "fender precision bass pickguard" gives me
this result

http://www.google.com/search?sourceid=navclient&ie=UTF-8&rlz=1T4GFRC_...

which is a SEARCH RESULT from the Fender website.

(upon further looking, this might be a page that's in danger, because
fender uses search queries as navigation)

And, it's exactly what I was looking for.  How is that not relevant,
or helpful?  I liked that result.  In fact, if I had gone to the
website and searched for that term, I would have gotten the same
results.  Why not allow it as a Google search result?

I completely agree that FLOODING the SERP's with thin result sets
might not be the best thing in the world, but I don't see the harm in
creating a link to a popular on site search term, that gives a result
set directly related to the term, that would be a great resource for
the searcher.

"Here's an example of said autogenerated search results, except that
each search result resulted in a new subdomain, which spawned another
subdomain, etc. "

I know and loathe that particular spammer, and have been in the debate
of how to work something like that out.  (By the way, that spammer is
back with a new spamming technique)
http://rumblepup.blogspot.com/2006/12/infamous-spammer-using-blogger-...

So it's a tossup.

I agree with both of your assesments, but it's always...scary, when an
algo change occurs.