Oddly enough, though, I'm finding myself drawn to the same mechanical process -- aggregation and curation of material by hand -- as a way of fulfilling perceived demand in the niches I'm working in.
Back in the mid-1990s, I started working with (now-defunct) Free-Market.Net, hand-aggregating news stories and commentaries for their publication, Freedom News Daily (which, after going away and coming back under the auspices of the International Society for Individual Liberty, is now a re-brand of the publication I founded with some other FND veterans in 2002, Rational Review News Digest).
One of the things I'm doing now (for the last month or so) is called TE Juicebox. It's an online calendar (and daily email newsletter, etc.) of events called "surf promos" at marketing sites called "traffic exchanges."
What do these two things have in common? They require a human being to locate, summarize and link up -- in other words, aggregate and curate -- a bunch of material from different sites so that the end user gets "one-stop shopping" for a variety of products in a specific niche.
Could this be automated? Yes, but it would be difficult, because so many different sites use so many different formats to convey their information. And it wouldn't be as good, because that variety of formats (among other things) would likely cause almost any automation scheme to include material that doesn't belong and miss material that does belong.
Some of the bigger players in aggregation/curation do automate. Some of them do it quite well. But honestly, I haven't seen any that do it as well as a human or team of humans can do it, especially within highly targeted niches.
The problem, of course, is that aggregation/curation by hand is more labor-intensive on an ongoing basis and makes for lower profit margins. You get less content for more work, and monetization of web content is often a function of "throw as much crap at the wall as possible; hope that some of it sticks and moves ad impressions or results in affiliate program sales."
But, in my opinion, human aggregation/curation is still necessary, still results in a better product, and will be necessary and result in a better product for the foreseeable future. I think the market actually agrees with me on that, at least to an extent. People are more likely to get word of a development via Facebook or Twitter (where it has been hand-curated by their friends) than at an automated news site. The immediate adoption of Pinterest by so many people wasn't an accident, either.