What's that you say? Make the search "ipad file management critique 'advice line'"? That works. It also demonstrates the importance of file management, because by adding "advice line" to the search, you identified the location of the article, along with the contents.
Don't get me wrong. I'd hate to do without search, even the irritatingly slow Windows search engine. There are times when it's the only tool for the job, even if it isn't very precise.
What search needs: Semantic indexing
Search won't be versatile enough to replace file management until someone perfects semantic search -- which actually is the wrong way to say what's needed. The right way to state the problem is that we need semantic indexing and won't have it for a very long time. Semantic indexing is the difference between finding the three files on my system that discuss "using search instead of file management" and only those files, versus getting either zero or a zillion hits.
Almost certainly, Apple won't be the company that develops it. Apple isn't a company that tackles hard computing problems. Even the very clever Siri is built on Nuance's speech-recognition technology. As is so often the case, Apple's brilliance was in its recognition that the technology was ready for adoption and polishing.
Microsoft also won't be the company that develops it. If it could, Bing would be more than a me-too search engine.
Nor will it be Google, because if Google could tackle the job, it would tackle the job. That's its core business, and Google has developers who are as good as they come.
Right now, the only company with the chops to solve semantic indexing is IBM. Watson shows that it's more or less possible. Scaling Watson up to the Web (not our subject right now) is a lot of scaling. Let me rephrase that: It's a lot of scaling.
Scaling Watson down to fit into an iPad, or a laptop, or even a server-based content management system? Easily as difficult as scaling it up.
Metadata: File management panacea
Quite a few commenters insisted that a well-organized, outline-based folder tree isn't just essential to finding what you're looking for -- it reveals the structure of the knowledge the system contains. That thread developed into quite an interesting conversation as to whether metadata can do a better job of sorting files into folders than folders can. The short answer is, yes, it can, and I want it. Right now.
The longer answer is (no, I'm not paid by the word): Yes, and taking advantage of this capability is both an opportunity and a burden. The opportunity comes from not having to store one file in just one branch of one tree. This is a big deal, because while folder trees are hierarchies -- trees, with branches and sub-branches -- information is cross-classified.