[pcf] Google, Yahoo, AOL
[sketchy notes]
Eric Schmidt is being interviewed. He says that Google has lots of ways to get better. He points to two. First, Google doesn’t always put the right link in the first slot. Second, there’s what others — not Eric — call the “deep Web.”
Orkut is part of their strategy to learn how to collect information [from the social networks], as well as building their ours. The privacy issues will only get worse. For example, Google is getting sued by people for making available public documents that show they were convicted of crimes. But when push comes to shove, the company’s policy is “Don’t be evil.” He says people generallly agree about what is evil. [Say wha’? Could the opposite be any clearer? The fact that it doesn’t seem that way to Google is an artifact of their homogeneity. Which also means that they just haven’t happened to hit an issue that rends that homogeneity. I hope they have a set of back-up policies stored in a “In case of ambiguity break glass” container.]
Dan Rosensweig of Yahoo (265M users) says that social networks can help people synthesize information and create an affinity group of people who may have never met but who can share knowledge and make searches more precise.
Jon Miller says that AOL traps about 3B spams a day. Esther suggests that the user should pay the ISP: you get, say, 100 free emails a month. Yahoo says spammers would still find it worthwhile. Eric suggests that we’ll have public and unlisted emails.
Q: Thanks for Google News. And will you do Google for the home?
Eric: Google News has had an ever bigger effect outside the US. And we’ve done some work on Google for the home, but we haven’t solved that problem.
Q: [Tim O’Reilly] I tend to think technology advances through hacks. Social networks are currently bad hacks that tell me we really need to add P2P protocols to address books so you can visualize your real social network instead of building a faux social network.
Eric: It’d have to be connected to some sort of server to manage all the different devices you’re using. The current social networks are simply trees of information that computers could construct on their own if we simply gave them permission. It’s more a permissions issue. The current SN’s will probably evolve into being more than simply introduction services.
Esther: The problem is “friend inflation.” And, also, these SN’s require you to make social relationships explicit. [Right on.]
Eric: Social networks will get better as we figure out what problem they’re intended to solve.
Categories: Uncategorized dw