1 Comment

Summary:

Lumi is a new app that turns your internet browsing history into relevant suggestions for what to read next. It comes from the Last.fm founders, who are now tackling smart suggestions.

Co-founders of Lumi.

There are no shortage of social news readers out there to tell you what you should be reading. From Flipboard to Prismatic, Pulse to News360, Thirst to even-maybe-Facebook, it’s a hot market right now.

But most of those services require you to download an app and provide it with a good amount of information so it can serve you relevant content. Some require you to select topics you want to read about, or read a lot of articles on the app so it can learn what you like. Even the ones that just require you to connect your social media accounts assume that you have social media accounts at all, and that you’ve built up those networks there to reflect your interests.

So as an avid news consumer but a pretty lazy person when it comes to investing time in new apps, I was immediately intrigued by Lumi, which is launching Thursday. All the product requires? That you’ve browsed things on the internet before. Lumi imports your browsing history on public websites (so excluding sites like Facebook or Pinterest), and then gives you recommendations based on what you’ve read before. Pretty easy.

Lumi comes from founders Felix Miller and Martin Stiksel who are based in London and previously co-founded Last.fm, which sold to CBS for $280 million in 2007.

I met with Miller and Stiksel this week in San Francisco, who walked me through their ideas behind Lumi and showed me how it actually works. It’s pretty simple — you download a Chrome, Firefox or Safari extension to your browser, and then import your browsing history. Lumi processes your basic interests and then shows you articles corresponding to those interests (which is further informed if you click on articles on the site). My initial results for topics it thought I was interested in were a little odd — I wouldn’t say that Betaworks represents a quarter of my online reading, for instance:

Screen Shot 2013-07-10 at 2.30.45 PM

But it doesn’t particularly seem to matter, since once I went to my Lumi page, the articles it surfaced were all compelling and tailored to my random interests. For example, it knew that I’d be interested in Lena Dunham, the SFO flight crash, Kanye West, Kevin Systrom, Beats, and weird New York behavior, all at the same time. Consider me impressed:

Lumi screenshot newsSo what are the downsides? There’s no mobile app right now, although the website is optimized for mobile, and the founders said it’s next on the list. But overall, they said they’re excited to see how people use the data.

“Usually the only interaction people have with their browsing history is deleting it,” Miller said. But he and Stiksel said they hope Lumi changes that. “Browsing history gives us a great picture of what people like, without them having to do anything,” Stiksel said.

  1. If you are looking for a reader that really learns and continuously adapts to your interests, look no further than NOOWIT. Collaborative filtering combined with NLP just don’t do the job.

    Share

Comments have been disabled for this post