[monkeymace]

How great would it be if Stumble Upon was pre-seeded with a ton of content automatically mined from your online behavior. Manually curating a relevant stream of content requires alot of energy and effort. Right now, Facebook is the easiest, lowest effort source of content. I imagine most Hacker News readers are pretty advanced when it comes to content discovery. But the average user is not as savy. By giving them a few tools, and a clear rules about how their behavior on facebook, (liking, sharing, reading, viewing, etc) contributes to the ads they are served, it could actually become a great resource that they pay attention to

[drharris]

The best forms of advertising don't give a user what he is already looking for, they give him what he doesn't know he needs. By manually curating what ads you like, chances are you'll miss stuff that is actually relevant. I'm a developer, but I have no desire to "learn javascript", because I know it already. Someone who is really into vintage cars will ignore car ads all the time; they have an existing network of sources and information and aren't as interested in changing that unless the story is very compelling.

Contrast that with the idea that FB knows about a 55-year old man that makes ~90k and has few hobbies, but has recently started reconnecting with all his buddies from high school; that type of person is now a profile for someone in whom an advertiser might create a NEW interest for vintage cars. The best part is you don't have to use forensic psychologists to create these profiles; simple statistics would be enough.

[freehunter]

Exactly! Facebook already knows everything about you (if you're using it right). They pre-populate what they think you'd like, and you can fine-tune it manually.

It might even lead to more engaging ads, and a new market for companies who can't figure out how to create gripping ads on Facebook at the moment.