Regarding "tried many times over"... You mean like web search was before Google? Or MP3 players before the iPod? Or photosharing before Flickr?
I'm not saying I'm convinced this idea is a winner, but come on... Predicting failure for a startup is so laughably easy that I don't know why you'd bother, unless you just wanted to feel smug.
I found it buggy, too. But there's one thing I've learned from the software business: I have no idea what a given thing is worth, and whether it can be sold. I almost fell on the floor when that calendar company sold their (cough, cough) "technology" for some 200K. Xobni's widget got an offer for 20m or something like that. Wtf?
So, yeah. Don't be surprised if a semi-functional, 1000-times-already-done, text highlighter pulls in a couple million. I don't get it either.
Anyone know how they are calculating their limit and/or why it's there?