Comment by cletus | Hacker News Reader

Comment by cletus | Hacker News Reader

This issue in a roundabout kind of way touches on Facebook.

The issue of social search has a lot of mindshare. Some think it is the future of search. I disagree.

One of the things that made search successful anduseful early on was scale. Instead of having to go to the librar or ask your friends you can effectively canvas the connected world.

I find the notion that friends' recommendations will replace that as nothing short of bizarre. It's like a huge step backwards. The argument is that you can filter out the garbage as your social graph will provide a level of curation.

Let me give you a concrete example. If I wanted t buy a camera I'd stil need t go to dpreview and other sites. It's highly likely that my friends don't really know a lot about this (but some will have an opinion anyway).

This same idea of human curation is behind such sites ad Mahalo and the garbage sites themselves to a degree. Of course at some point computers will be powerful enough to generate this garbage content.

Blekko's idea of slash tags s interesting (to a degree) but if it's successful its easily reproducible. Google is still in the box seat here but of course that's no barrier to a link-baiting TC title.

Personally I'm an optimist. I believe that, much like email spam, the garbage from AC, DM and others I'd a transitional problem (email spam is basically a solved problem now if you use a half-decent email provider). If they succeed we won't be able to find anything. I don't believe that'll happen so these services are therefore doomed.

So betting on Demand Media is (to quote Tyler) like betting on the Mayans (meaning betting they're right about the world ending in 2012: it doesnt really matter if you're right).

So my money is on Google being the better Google.

There are other options besides social or text-based algorithms. My company, Pikimal, is pursuing one - we're pulling together facts on items and are allowing people to weight those facts to dynamically create recommendations. We've added some "user rating" facets but are just barely past our minimal viable product.

I very much agree with you that social is limited. It's a filter which avoids spammers, but it also filters out experts and users. It also solves some of the problems introduced by the generic nature of search algorithms.

I'm betting my time on the idea that the solution which wins will combine an understanding of the product space, the value of new features, a current understanding of price, and which can be customized transparently to the needs of the user.

How does filtering based on a trust network imply that the network will only include your friends?

Indeed. For such filtering to work, a significant portion of your friends should be searching the entire network and not just filtering through friends. But this might just work out to an early-adopter/follower divide. Friend network filtering will work better for the followers. The early-adopters/leaders will continue to use other means of search and will feed the friend-network.

That assumes the same group of people will be the early adopters for everything, from recipes, to gardening tools, to what video card to get. If your peers don't happen to be knowledgeable about what you're seeking, then you're SOL.

Your best bet ends up being to simply include everybody, which improves your chances of finding an expert opinion, regardless of the obscurity of the query.

That assumes the same group of people will be the early adopters for everything

Logically speaking, no it doesn't. Why couldn't a search engine be aware of topic/product areas? Many already are.

Well, at that point, your 'trust network' may as well include the blog/review site that you would have found through a normal search to begin with.

You would only consider the people you trust on a specific topic when ranking results (not all backlinks).

Interesting question.

I'd divide the possible filtering processes into three approaches.

* Throw all your questions to all your friends and all the review sites you "trust". This means any third party site you trust gets a rather excessive ability to spam you - even a site with good user reviews on them are trying push pop-up windows on me. Just because I've gotten good info from X once doesn't mean I want any more from it.

* Throw friend recommendations and trusted sites to a third party "meta-filterer" who organizes things for you. You'd have to really trust that site and essentially there's no reason they'd be better than Google.

* Do the filtering yourself. Most people essentially do that now. I'm working on a project to create tools to automate and improve this this process. Create your own relevance and topic-weighting system that adaptively filters all the other filters. I believe that this kind of approach eventually going to be needed. Not so much because each person can or should do all their topic-relevance-weighting but because this approach would keep the other systems honest.

-----------------

Eventually, people are going to realize that both their social graph and their content-relevancy-weightings/algorithm are far too personal to farm out unquestioningly to a third party. The present social networking system is like AOL-email in 1992 except with the added provision that provide can look-at and alter your emails as part of the service. I'd envision this stabilizing to the present email system where webmail exists but has to work more or less the same of email to your personal computer.

The problem with 'social' search is that it reduces the universe of documents to what is known to the members on the platform. There is already data that is not crawled by most of the search engines, which itself is a problem. This only makes it worse.

> I find the notion that friends' recommendations will replace that as nothing short of bizarre.

Not for generic terms. But for _buying stuff_. If I am buying a camera, I'd be more interested in what my friends own, recommend.

Reply on news.ycombinator.com