For those of us not willing to watch the five minute video (I generally fall into this category), an entity is kind of like an ID for discrete things: people, places, or objects. So whenever people ID things, it adds a bit to that entity. So, for example, the University of Illinois is known as UIUC, UoI, U of I, all sorts of names, but everyone is talking about one university. Metaweb would still, however, attach everything to the “University of Illinois” entity. So theoretically, an entity contains every mention of that person, place, or object, every photo, and every relevant piece of data. Entities are an abstraction of the confusing mess of data that search engines deal with every day, and allow relationships to be mapped out much more easily.
Anyway…on a more relevant note, if you had a search engine that was actively constructing you as an entity, gathering photos, comments on websites, blog posts, and whatnot…isn’t that a lifestream? Furthermore, wouldn’t you want control over it?
I’m glad that Google is keeping FreeBase, Metaweb’s open structured database, free to the public and open. That way, Metaweb’s entities are open for everyone. But I’m still uneasy about having entities under one company’s control. Are there any alternatives? Preferably free and opensource.