A low-key experiment to find high-key interest
Google’s Area 120 team, an internal incubator that creates experimental apps and services, has launched Keen: a would-be Pinterest rival that draws on the search giant’s machine learning expertise to curate topics. Available today on the web and Android, co-founder CJ Adams says Keen aims to be an alternative to “mindlessly” browsing online feeds.
“On Keen […] you say what you want to spend more time on, and then curate content from the web and people you trust to help make that happen,” writes Adams in a blog post. “You make a ‘keen,’ which can be about any topic, whether it’s baking delicious bread at home, getting into birding or researching typography. Keen lets you curate the content you love, share your collection with others and find new content based on what you have saved.”
This is obviously not a particularly revelatory pitch. Just about every social media feed you browse is trying to personalize its content to your interests in one way or another. And Pinterest has already captured the hobby-focused side of this market with its pinboard-style visual design — two characteristics that Keen is trying to imitate
So what does Keen have that Pinterest doesn’t? Well, for one it has Google’s expertise in machine learning, which Adams says will surface “helpful content related to your interests.”
“Even if you’re not an expert on a topic, you can start curating a keen and save a few interesting ‘gems’ or links that you find helpful,” says Adams. “These bits of content act like seeds and help keen discover more and more related content over time.”
But it’s not like Pinterest doesn’t invest heavily in AI as well. And while machine learning’s ability to find patterns in data outstrips that of humans in many areas, when it comes to niche hobbies and interests, I’d wager that the collective intuitions of a big and engaged (dare I say, keen?) userbase will outstrip those of the machines for the time being.
At any rate, it’s interesting to see Google push its machine learning systems into more varied applications. Especially those that seem like they’re trying to foster users’ interests in rewarding hobbies, rather than algorithms that drive people to greater engagement without caring what it is they’re actually engaging with.