In their recent blog entry they say that it is, but in other countries you can sync up with your friends. In the US you can share rss feeds of what you have watched and me and my friends use that.
Netflix used to have friends features in the US too, you could see what they'd watched, and take quizzes where you guessed what movies they liked. It was pretty fun.
It seems quite a stretch to call a database of movie and product ratings a "social graph." Amazon is building something extremely valuable; they are doing it above the radar, not under it; and it doesn't strike me as a social graph.
So it's last.fm for books? How it that a social graph? I don't mean to discount this data entirely, but what some reads (or what music someone listens too) is not anywhere a comprehensive picture of a user; for this reason, niche "social" services like these hypotheticals won't be able to compete with a monster like Facebook. If the service could be a Facebook app, why will users choose this new standalone network over their existing network that already has all of their friends?
A couple commenters point out that this isn't really a social graph, which is totally fair objection to the way the author is talking about this stuff. But my main concern is that this doesn't present any evidence that Amazon is doing this (and I'm not saying they aren't), but just suggests why it might be in Amazon's best interest to do so. They put together a couple graphs, one of which appears to be based on publicly visible "customers who bought X bought Y" data, but that data has been around and visible, and scrape-able for years, and they don't present any reasons for believing that this data is being used by Amazon in a different way now than it was 5 years ago. So, in other words, reckless speculation.
I never care about other readers' notes. In fact i find it annoying that you can't turn them off on the kindle. I don't care what other readers wanted to take note of, maybe because people read a book for different reasons and keep notes for different reasons (one may be doing research on a subject, another may just be collecting jokes). If that was a general case, Library books would be sought after for their side notes. There's little information there. OTOH the "people who bought this also bought" is sometimes useful, sometimes a recipe for very average reading.
To be fair the "friend network" recommendation are not much better either, but the trust factor involved makes us perceive them as more important.
Are you sure you can't turn them off? I thought with my Kindle 2nd Gen I was able to- I recently replaced it with the most recent Kindle Keyboard and unfortunately do to an accident it is out of commission right now so I can't check, but I was fairly certain with the 2nd Gen (note I probably hadn't updated the firmware in a long time) I could turn off other people's notes and highlights.
As it hasn't been said yet - I really hope that Amazon does not do this. We trust Amazon to deliver products, not strangers. Apple's Ping failed for good reason.
I think the point being missed here is that although knowing what people might like is indeed a marketers dream, most computer-generated recommendations systems don't do a very good job at it.
Perhaps some people use recommendations like those at Amazon, but I think generally most people avoid them because (at least at present) the signal to noise ratio is ridiculously bad.
Pandora's recommendation system is not based on a social graph, nor is it particularly "computer generated" compared to most other automated recommendation systems.
It's based on the Music Genome Project. For over 8 years, they've been having highly trained musicologists listen to each song individually, then manually categorize them on ~400 attributes (e.g. is it syncopated? major or minor key? does it have vocal harmonies? etc). Pandora uses this data to find similar songs to the ones you like.
You'd think it wouldn't scale but Pandora seems to be doing it.
"The Music Genome Project's database is built using a methodology that includes the use of precisely defined terminology, a consistent frame of reference, redundant analysis, and ongoing quality control to ensure that data integrity remains reliably high. Pandora does not use machine-listening or other forms of automated data extraction."
http://www.pandora.com/about/mgp
At ROFLCon a couple years back, the Three Wolf Moon meme was described as being started by a guy looking for law books on Amazon, and it said "people who bought xxxx (law books) also bought.... the three wolf moon shirt" ... so he had to buy it. http://en.wikipedia.org/wiki/Three_Wolf_Moon
18 comments
[ 9.4 ms ] story [ 79.1 ms ] threadTo be fair the "friend network" recommendation are not much better either, but the trust factor involved makes us perceive them as more important.
Perhaps some people use recommendations like those at Amazon, but I think generally most people avoid them because (at least at present) the signal to noise ratio is ridiculously bad.
It's based on the Music Genome Project. For over 8 years, they've been having highly trained musicologists listen to each song individually, then manually categorize them on ~400 attributes (e.g. is it syncopated? major or minor key? does it have vocal harmonies? etc). Pandora uses this data to find similar songs to the ones you like.
You'd think it wouldn't scale but Pandora seems to be doing it.
"The Music Genome Project's database is built using a methodology that includes the use of precisely defined terminology, a consistent frame of reference, redundant analysis, and ongoing quality control to ensure that data integrity remains reliably high. Pandora does not use machine-listening or other forms of automated data extraction." http://www.pandora.com/about/mgp