A new start-up called Factery Labs has been getting quite a large amount of cyber-buzz lately and for good reason; with Factery Labs we see what could be the death of links and the birth of compact information.
Current search engines provide links to what might be relevant content. A user has to scan these different links, and decide which links to follow. Most links do not satisfy user needs and they proceed to other sites to find the information. Time is spent on this click-in click-out process and it wears on a user’s patience. Factery Labs now have a new search engine that does away with the whole link system and instead gives the user “facts” instead.
In recognition of PageRank, Factery Labs uses a technology called FactRank developed by Paul Pedersen. What FactRank does is crawl through web-pages or whatever source content looking for semantic information which is evaluated by an advanced language model. When it identifies a passage with a targeted message it cross references it against other similar sources. By doing this, the algorithm starts accumulating “facts”. These facts aren’t exactly a guarantee for truth, but more like a democratic fact. What the user gets in the form of results is a page with chunks of information that is organized based on its relevance. They get the gist of the subject they’re searching for right there on the search page in one glance.
In fact, most searches on a website are done just to get a small snippet of information instead of lengthy paragraphs. If these snippets come organized from different sources on one platform, it saves the user a lot of time.
One fall-out will be spammers trying to hack into the algorithm by manipulating content. This is no different from the challenges faced by a regular search engine. By maintaining a blacklist of potential offenders and constantly tweaking their system, results will get incrementally reliable.
The product isn’t fully functional as a consumer search engine yet though. The website offers demo searches as a proof of its concept. It searches inside real-time and social networks like Twitter and Facebook. They have made available an API for third party developers to use. Possible applications could be to deliver reading summaries rather than traditional links. This is particularly beneficial to mobile users, due limited screen space; condensed information contained in a tight package is highly valuable. The range of products that the technology can be morphed to fit into is pretty interesting.
The biggest issue though is accuracy of the results. The fact remains that judging the truth of any content is outside the realm of any existing technology. Even a human being is incapable of discerning a fact’s accuracy independently. Since nobody really knows how to tell fact from fiction, the best possible path to follow is the one of neutrality. By not placing a bias on any form of content, it relies on the democratic nature of the service. If enough people talk about some particular issue, then that snippet will be treated as a fact. The only real bias that the company can afford to have is toward malicious sites. Spammers will be included in a blacklist, and popular sites will be given importance, just as it is done in traditional search technology.
The current goal of the company isn’t to become a big search engine, at least for the new future. The real potential that is being targeted is the API service offered to developers. There are a number of obvious and non obvious applications that this technology can provide. The ones that do come out in the near future promise to bring with it a new paradigm in the web experience.
Just when we thought the search engine market was getting interesting, the stakes just went even higher!