Scott Rosenberg has an article looking at Google as an example of the potential for innovation even when it seems like there is not room for any:
you have to recall the Web scene that Google was born into. In 1998, search was over. It was a "solved problem"! Altavista, Excite, Infoseek, Lycos, and the rest — all these sites provided an essential but fully understood service to Web users. All that was left was for the "portal" companies to build profitable businesses around them, and the Web would be complete.
Google burst onto this scene and said, "No, you don't understand, there's room to improve here." That was correct.
Mr. Rosenberg goes on to write: "it's a universal insight that never stops being applicable: there's an endless amount of room to improve, everywhere. There are no solved problems; as people's needs change and their expectations evolve, problems keep unsolving themselves." I'm not quite sure he's totally right that there are "no" solved problems — that might be an exaggeration. But he is right, I think, that there are probably fewer solved problems than most people assume, and that problems that look solved may become unsolved quicker than we think.