CSS was new when I started with it, so really when it started to appear in browsers I began to play around with it!
Perch came from a need we have as a company, we mostly did client work for design agencies rolling out custom CMS projects based on a larger framework we had developed. There kept coming up a requirement for something that had the same focus on structured content but was quick and simple to drop in. Solving that problem meant developing Perch v1. So it really was our requirement that we then realised might be useful to other people as a standalone project. The first customers were people who signed up to a mailing list we put on the holding page - some of those people still use Perch 7 years later.
It's very important. Firstly for your own career, if all of your skills are tied up in one framework and that stops being popular you are going to be back to square one having to learn something new. Learn the core technologies and you can have a lot of confidence that you'll be able to move between things.
Also, if you don't understand the technology, it is very hard to be creative and come up with truly new and innovative solutions working with an abstraction that gives you a limited subset of the power of the underlying technology.
Lastly, unless you understand the edges of what is possible with CSS, you are unable to offer feedback to make CSS better. The CSS Working Group are keen to hear what people want out of CSS, browser vendors want to know if you find bugs or which parts of the platform they should be focussing on. If your view of CSS is only what is provided by a framework then you aren't in a position to do that, and this all gets better the more of us push to get the tools we need to do our jobs.