Does MySql/Postgresql coupled with dynamically typed languages (Ruby/Python/Groovy) solve 80% of IT problems?
Of course this depends on how much of a load your average user generates.
The database is a bottleneck, sure. But usually you hit other roadblocks on the way up that are a lot harder to solve than how to make your database scale.
To put it in very pragmatic terms, on the web there are currently 1,000 companies with more than 1M uniques daily, if you're going to be one of the lucky 1,000 then you will haver a scaling problem some day. But that scaling problem will be a relatively minor problem, plenty of people have solved that and documented it.
Going noSql is one of the ways of getting around that but there are several ways of achieving that.
For example:
C++ -> Class-based object orientation
Java -> Virtual machines
Rails -> Web frameworks
NoSQL -> Scalability strategies
Erlang, Haskell -> Concurrency strategies
So really, the appropriate way to ask this question is, "are these concepts enough to solve 80% of IT problems?" Once the concepts are available to the mainstream, switching the specific technology is not a huge problem.
It's true that in the process of taking on fads, people tend to lose their way and become zealots, getting stuck on one concept as the end-all be-all, but that's true of most human topics.
I would venture to guess that at this point in time we're actually seeing a whole explosion of concepts thrust into the mainstream from many directions - some aged and well-known to academia, and others the result of front-line efforts.
However if the question you're really asking is "can those technologies help me reduce my effort or cost by 80% ?" then thanks to advances in hardware, services and open source in many cases the answer is yes.
1- http://research.microsoft.com/en-us/um/people/gray/talks/gra...