Skip to Content
Uncategorized

Lumi: a Useful Web Recommendation Engine

The service, from two founders of Last.fm, uses your browser history to suggest new pages to visit.
December 12, 2012

Lumi is the latest venture from two of the three founders of Last.fm, a popular Internet radio service that keeps track of what you listen to and then recommends new tunes based on your tastes. Similarly, Lumi is a browser plug-in that tracks the websites you visit and suggests other sites it believes you’ll enjoy.

This service is now in a private alpha test with around 1,000 users, including me. It doesn’t require the user to do anything beyond installing the plug-in (it works on Firefox, Safari, and Chrome). You can then navigate the Web normally—and when you want to see Lumi’s suggestions, you simply log in on its site and take a look.

Lumi records only the URLs you visit, not the content you view. The data that’s collected is saved on Lumi’s servers, where it is encrypted, anonymized, and secured. Of course, the information you provide is still valuable—especially to advertisers—though Lumi’s privacy policy states that this data is “only ever used” to improve user recommendations. (The founders say Lumi still hasn’t settled on a business model.)

I tried out Lumi by installing the plug-in in Firefox. After Lumi’s software analyzed my browsing history, it took me to a Web page showing a seemingly endless array of colorful squares, each pertaining to a page it thought I might like to visit. The site also shows popular pages, determining popularity by tracking pages that lots of people visit, says Martin Stiksel, who created Lumi with fellow Last.fm cofounder Felix Miller. Both left Last.fm in 2009.

While social recommendations through Facebook and Twitter are a common way to discover new content, Lumi’s creators are betting that an automated recommendation service can also prove useful.

In the few hours I played with Lumi over several days, it did make some good suggestions based on the kinds of things I look at throughout my day (tech news, startups’ websites, Facebook, Twitter, and pop-culture blogs, to name a few). I checked on it several times a day and saw things such as a link to a Kickstarter campaign for a file-sharing service, a design company’s neat portfolio site, and a review of the new iOS Gmail app.

Once you’ve clicked on a Lumi suggestion, it will be grayed out on your page of squares to indicate you’ve already taken a look. Older suggestions seem to move down the page over time.

If you’re interested in being a more active Lumi user, you can star pages you like, either on your Lumi profile page or out on the wider Web. Other users can follow you and see the stars that you’ve made publicly viewable, and of course, you can follow other users to see pages they’ve starred. You can also gather similar starred pages into groups.

Some users might be a little nervous about having their URLs tracked, but I liked Lumi’s approach. It’s simple to integrate, and unlike the somewhat similar social search tool Wajam (see “Review: Wajam, a Tool for Searching Socially”), it doesn’t literally follow you around on the Web, which can get annoying. And while it shares the basic premise of content discovery with StumbleUpon, I liked that it didn’t rely on me to seed it with my interests or take any action on the Web to get better recommendations.

“We don’t want you to change your behavior as such,” says Stiksel. “We want you to carry on doing what you’re doing: browsing.”

One thing that seemed a little odd: Lumi seemed to get both better and worse the more I used it. After a few days, I noticed it was giving me more interesting links that fit in with the kinds of things I was checking out on my own online. Yet it was also giving me more links to tech news that I’d either already read about or even written about. As the “private alpha” designation suggests, it’s still very early days for Lumi. Its creators still have much work to do, including (eventually) convincing a lot of people to use it. But given its purpose and the success of Last.fm, there’s a good chance that Lumi will win fans, too. 

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

How scientists traced a mysterious covid case back to six toilets

When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.