Curation: Humans v. Machines
Post from the old site, but updated with some new content:
I’ve come to rely on many of the sites I often visit to curate content for me. On Facebook, the News Feed curates news about the friends it thinks I’m most interested in. On Twitter, I rely on certain people I follow to discuss topics of their expertise. I use Netflix and Amazon’s suggestion features to point me towards things I might want to buy. With so much information available on the Internet, curation is becoming increasingly important and necessary. Web curation is based on a few simple premises:
1. The internet has lots of information on it.
2. Users don’t have enough time to sift through all the information themselves.
3. Algorithms can search the web faster than humans and through their social graphs people have already exposed their preferences in one way or another online (your RSS feeds, what you like on Facebook, who you follow on Twitter, what Google search results you +1).
4. You sign up for a site that promises to curate a certain topic for you (I’ve seen news curation as the most common one).
5. In addition to dedicated curation sites, users on networks are curators. Content you post on Twitter, Facebook, Pinterest…it’s all content that you suggest and bears your seal of approval.
This is undeniably a good thing. I’m a voracious consumer and am constantly looking for recommendations or information relevant to my interests. I use Twitter as a proxy now. I follow people who I’m interested in and they post links to things that interest them. I follow people on Google Reader and read whatever they share. I’ve tried Summify, but haven’t been too happy with the results. Summify sends you the top five stories of the day that you should read based on your Google Reader feeds, Twitter follow, etc. The problem is that it weighted the results too heavily towards feeds that post a lot. Also, because I tend to try and click through everything on my Google Reader, Summify was very repetitive for me because it would resend me content I had already seen. What’s the point in that? You’re supposed to make my time more, not less, efficient.
Regardless, this curation movement is very beneficial. I think it is missing a human component though (which actually may undo the efficiency gains of using an algorithm). Noah Brier recently wrote a post about how the algorithms are created to err on the side of being more conservative – the Netflix algorithm, for example, is more likely to show you a movie that you will give 4 stars rather than a movie which has a 50% chance of 1 star and a 50% chance of 5 stars. It make sense that in creating an algorithm, companies feel like they need to take a safer approach so that their recommendations are respected and used. But the missing risk component does diminish the serendipity moments of finding something great. And that’s where human curation through networks can really play a big role. You can even see Google and Bing starting to understand the human component that can complement an algorithms results as they start to use +1 and integrate Facebook/Quora/etc.
The truth is that we trust our friends and people with similar interests when it comes to recommendations. For news, I’m fine with an algorithm. As for entertainment? It’s just too subjective and emotional to rely on equations.







This is me, digitally.