Frontend development has come along leaps and bounds since the days of document.layers and MSIE6 alert() debugging, it seems to be slowly coming out of a renaissance period lately as many devs have reached a consensus that they are spoiled for choice and now all that's left to do is, well, build.

There is this trend of developers feeling just as you described: overwhelmed. But rather than feel that, I try to embrace it. Like anything on the web, if you're not building on strength, then you must be in it for other reasons, like trying to impress employers, or trying to learn code because apparently it pays the bills better than other gigs.

I would start small, and treat everything like an experiment. If an experiment works well, you can build on top of it, and import what you learned from experiments into full blown (hopefully paid for) development.

I sometimes have to remember to use <em> instead of <b> but only because I didn't think such things were above me. Indeed it's a miracle a visitor to your site can even read the content with the temptation that exists to include another slider widget, or inaccessible web component.