Human-powered Search Engine
As the world’s leading search engine, Google.com, uses powerful computer hardware and software to rank web sites and deliver relevant results based on the search terms entered by users. For the most part, the system works great. Sometimes, however, it can take a little time to find exactly what you’re looking for.
Jason Calacanis, an internet entrepreneur, started Mahalo.com to provide spam-free, relevant links pertaining to the most popular search terms. Instead of using complex algorithms, he has a paid staff to hand-pick and organize the best links for each subject.
Mahalo is still in a beta version and there are a lot of things not covered. In fact, it seems impossible that Mahalo will ever have hand-built results for everything someone will want to search for. However, after experimenting with the site, it does look promising, particularly for when querying for vacation travel destinations or how-to articles.
It will be interesting to watch Mahalo.com over the coming months. Does human-powered search have a future? What do you think?
March 14th, 2008 at 10:52 am
Human powered search would have been a great idea in the incipient days of the internet, but the very dynamics of networks make it impossible for an expanding node-based network such as the network of hyperlinks of relevant subjects, to be covered by a limited, rule-based semantic system such as the human. It makes much more sense to use raw processing and search power to generate lots of results, which can then be parsed by humans.
This problem is not unlike the so-called Genetic Swarm Optimization algorithm developed for engineering problems, which inherits the best features of the Genetic Algorithm (an evolutionary search method) and the Particle Swarm Optimization Algorithm (another search method based on swarm intelligence).
August 19th, 2008 at 6:56 pm
[…] Related: Human-powered Search Engine […]