Line-art vs. Color: Thoughts on a high-resolution social graph

scotchi.net

26 pointswheels3 comments

Right, except Facebook is also the largest photo-sharing site on the web right now, and one of the most used iPhone apps. So, they've already got all of that data (edit: and momentum) that Color will need to somehow get, and I'm not quite sure how Color will disrupt THAT.

I do agree about the part that Sequoia probably knows what they're doing but the current Color product is shoddy and underwhelming, with that amount of cash and no pressure to launch (SxSW is gone!) - why would you launch a shitty product?

Launching right after SXSW is in itself disruptive ;)

I don't take photos on my phone to share on Facebook, so I genuinely don't know: do Facebook photos have location data? I think what's really interesting about Color is the notion of an elastic/ad-hoc/inferred social network based on location and meaningful action. The issue is you need Color to be the default photo application before that really works.

Seems like the type of thing Facebook could do much more easily with machine learning. Even some simple graph analysis should be able to extract your various clique-like structures. From there you have a dense feature set to figure out what each cluster probably means to you.

Reply on news.ycombinator.com