Jinni – The Future of Search Is Personal
Somehow, every time a new search engine emerges, we hear a new slogan that predicts the future. Hakia hoped to break the rules as the first ontological semantic and natural language processing (NLP) based search engine, but it lost the war when Powerset, its main strong competitor, sold to Microsoft.
What few know is that all search engines have been using NLP in since the 50s, and all computer programs that deal with language use NLP (including your spell-checker), so bragging about being a NLP-based search engine is not something I’d advise new search engines to do.
Both hakia and Powerset used to call themselves semantic search engines. This has changed, and both make broader and more realistic claims instead. NLP is not synonymous with semantic search – semantic search is just a part of the technology – and we are still very far from having a true semantic search engine, although many are heading in the right direction.
This introduction about NLP and semantic search present a new search engine that has a clearly defined point: Jinni, the new video search engine self-labeled as the “first Taste Engine,” states among its algorithms semantic search technology and personal recommendations, both based on Natural Language Processing and taste profiling. We have to understand from the start that such claims, without access to the algorithms, cannot be sustained by any journalist. As far as the search engines making the claims go, any new innovation needs to capitalize on something, and the idea of semantic search is nebulous enough to allow such PR tweaks.
More interesting about Jinni, however, is their vision of the future where a search engine becomes personal and results are served based on personalized recommendations. These recommendations are made possible by the core of the engine – the Movie Genome (created by movie professionals) that contains several thousand “genes” assigned to each title to describe plot, mood, style, setting, soundtrack and more – a rich alternative to the usual genre language. From an SEO standpoint, the Movie Genome is a step forward: through tagging page elements and video content, to be more accurately identified and visible for search engines, Jinni actually proposes an improvement to the currently existing algorithms.
Another interesting aspect of Jinni is the “pulse” – a feature that shows live streams of users’ actions and opinions. An ongoing monitoring of social networks like Twitter, Facebook and personal blogs is also part of the “pulse” as Jinni wants to be part of the conversation, without actually becoming a social network itself.
Last but not least, the personalized recommendations are indeed personal. There is no “people who like this also like…” as you find on Amazon, Netflix and other similar networks. Jinni’s model is unique. With integration to Netflix, Hulu and other leading content providers, Jinni aims to be the personalized starting point for choosing what to watch next, and according to CNET, “Jinni is the best movie recommendation engine on the Web. Period.” I tend to agree. What’s your opinion?