I would say that adapting results towards expected customer preferences (eg giving pet owners and programmers different results for "getting started with python") is curation.
Perhaps more controversially, so too is the pruning of results at all – if most of what I'm searching for is spam/malware, then so be it.
Search engines are already very curated, IMO. There's room to differentiate by choosing NOT to curate. I would be curious to use a search engine that used nothing but PageRank circa ~2006.
Pedantically, any algorithm/ranking/weighting system is unavoidably curatory by design.
But to answer what I think you mean, yes, I think there's room for opinionated search engines for many values of "opinionated". Another way to think of these might be "contexts" that you apply overtop of a search engine, and perform many searches within that context.
For example, when I'm searching for front end things, I wish google were smart enough to time-scope my searches dynamically using knowledge I don't have. I search for Mongoose operations and get top results from 2011 saying X doesn't exist, but it was added in 2016. I would like to see only results from when the thing I wanted existed. But only ever searching the last two years all the time costs results that would be helpful in other times.
When searching for device repair instructions, I wish the engine were able to adopt my estimations about sites that aggregate forum replies or repair instructions in automated ways to gather clicks, and either banish them or loudly mark them.
I'm not sure how to do this with major traction, or without being eaten by google, but "ecosystem of many opinionated and accurate search contexts" feels like it would add value to the world.
However, I agree Google's results are getting worse, but I think that is more a result of the way they weight results, and personalize them.
Covering personalization - I've yet to see it work effectively for a large audience. It does drive certain markets, but by whales rather than everyone. So most people get things badly, and a few are driven towards a certain goalpoint, in Google's case adverts, where they tend to make clicks or purchases of the intended partners.
On weighting results - Google seems to have stopped weighting results for relevance a long time ago. They seem to prioritise things differently now. Exactly how I'm not sure, but it does seem to include Recent Events, using technology Google has a vested interest in, and Partners, and all those get addressed before relevancy.
If you can solve the relevancy problem (which is huge, from above, Recent does not always mean Relevant, but it can), then you can certainly attract the technological crowd.
Google seems to have aimed their search engine at popular events, news and social media.
Where would you aim your search engine?
- If there are too many relevant websites to display on one page, search vs. manual directory does not really make a difference.
- The curated model was long kept up by DMOZ, which had an army of volunteer editors but eventually shut down. Some editors took over and built curlie.org, but there is no innovation since DMOZ (which itself had no innovation)
- Curation is a big topic nowadays. Pinterest for example is based on that idea. However, things might be a bit more decentralised these days. I am just thinking about the awsome-* lists on GitHub.
- Manual directories may fail build a "searchable" taxonomy. I am searching maybe for a technical support company specialised in that product, located in this city, offering low rates. This is an easy example, but there are tougher ones. General purpose information retrieval (like Google does) might be a better fit for such problems than prebuilt taxonomies.
- Based on the last two points, curation might be more successful in small topics rather than in a Yahoo-like all-topics setting.
However, the value in curation is only on a particular subset of queries. Those where you want to discover information and have some introductory terms to go on ("Motorbike Safety", "Analytics SAAS", "Buying a Fridge" etc). What youre looking for is an expert in a niche area to point you in the right direction. Curation is very beneficial here.
The second type of query, where you are looking for a specific piece of information (specific business, specific book review, weird fact) etc Google already does fine on. And curation wont help (nor will it scale on such precise info).
Anecdotally, if you type a google search for something along the lines of spring break in Florida, then search Google Flights, it auto-populates the dates of Spring Break into the search.
Search for [flights to london], then start typing [things to do] and the first suggestion will be [things to do in london].
In so many ways google is failing many niches that it once served well. I have been doodling ideas on how to do some of those niches better, like adult.
Works pretty well in most cases.