The Art Genome Project
Like its musical sister, Pandora (of which I am a big user; happy enough though I wish there were more different offerings but it has been growing), the idea is that you input the name of an artist you like and the genome will link you with others who are similar. The theory argues that there are defining characteristics of art that, like the instruments/beat/vocal quality/lyrics of music, if you quantify enough of the characteristics of enough of the input items you can create links.
Of course, the idea is not to leave it just in the theoretical "What kind of art do you like?" stage but to draw more folks in to the art world. Just as with Pandora, the idea is to sell the work of the artists by expanding familiarity and expanding access.
I like the art of Roy Lichtenstein. I "follow" him, as I would on Twitter (or choose him as a root as I do with Pandora). In addition to being able to browse some of the works of his listed (and there's a good interface for basic info and zooming), I get linked to other artists. The link now gives me updates for Jim Dine, William Eggleston, and Steve Lambert; I can see upcoming shows, thumbnails, and if there are works for sale, I can get in touch with MY DEDICATED SPECIALIST (yup, an art history major all my own who will help me begin my collecting.)
How does it make money? They're partnered with galleries directly (not with artists) and receive 3% of sales that come through their specialists/site. (I don't know if they get that $ on the listing so that even if I walk in to the gallery and fall in love with a work and buy it on sight. It would make some sense...) Theoretically, the larger they grow, with connections to more and more galleries, they could make enough money to do well.
What do I like about it?
* They seem to have purchased/legally obtained their works on the site. You cannot download the images. Copyright of images is a tough legal area (and one with things I would like to change but for now, compliance with the law is good).
* I like the ability to both search and browse as I build my roots--there's a lot of paths one could take to building a personalized collection.
* I don't really know enough about computer engineering to get this but I am intrigued with the amount of open software they're using; interested folks should check out their blog.
What do I not like about it?
* I realize they're in invitation-only Beta mode but when I requested an invitation, the first e-mail I received was "Sign up 15 of your friends to get faster access". I'm not going to spam my friends, folks. (though if you want an invite, ask me.)
* The interface is not exactly intuitive. It can be difficult to navigate from one form to another, to go from search mode, to browse mode, to update page.
* I have limited access to changing my genetic offspring. With Pandora, I may input "The Smiths" and get "Echo and the Bunnymen" as an offspring; I am allowed to vote thumbs-up/thumbs-down on each individual song so that I don't have to listen to a song by Echo that I hate but may still get others by them and may get offspring related to them as well as the Smiths. Art.sy does not let me look at the connections and say "I hate Jim Dine (for example, not for true)" and block him from my update feed. So currently: by inputting Ai Weiwei, I now have a calligraphic traditional Asian art show listed on my feed of upcoming shows. The project gets better with more opportunities to refine it.
I find it an interesting project theoretically. I find this logarithm of "If you like X, you might also like Y" a really fascinating aspect of modern consumerism. I'm interested in the idea of traits that might help make connections between artists, especially when one goes circuitously from Jan Van Eyck to Gregory Crewdson (a connection made for me which has me scratching my head a little).
As Wired magazine pointed out, there's a lot of money potentially to be made here.
