Welcome to Learned, a short, weekly look at language, education, and everything else under the sun. I’m Joel, linguist, teacher, slacker. This week, we're looking at the web in a whole new way.
subscribe | unsubscribe | comments | twitter | about
This week's issue is a little different; I don't want Learned to become another link-aggregation newsletter, but sometimes the list-of-links format lends itself to a moving target of an idea. In this case, I wanted to write about how the interfaces we use to access the internet are changing. I know, big topic. But it's one I think can best be explored by studying the examples and uses of new technologies rather than the technology itself and that's what I propose to do this week. So, here's the clickbait: 7 Ways A.I. Is Changing How We Use the Internet! And here we go.
Deep Nostalgia by MyHeritage.com
You may have seen this one in the news recently. Using deepfake technology, My Heritage allows you to upload an old photo and have the A.I. give it limited animation, letting you see how your ancestors might have looked when they were alive and in motion. Honestly, it's quite a bit like the animated photos in the Harry Potter universe, which doesn't make it any less cool for that.
Same Energy by Jacob Jackson
Visual search engines are limited by the need for the user to explain what kind of image they're looking for. The most expedient and common way to do this is by typing into a search field. Same Energy aims to do away with that interface and allow the user to begin searching directly by uploading an image. Using deep learning, the algorithm finds and returns images that are similar. There have been similar projects before (Google visual search being a prominent one) but this is remarkable for it's ease of use, clear purpose, and incredible results.
Flim takes the idea that "images are everywhere" and applies a little bit of data science to provide the user with a new, ever-growing and ever-evolving library of stills from hundreds of sources. Although it functions like a traditional search engine (e.g. you type in an example of what you want to find) the results it returns are something else entirely.
Live Music Analyst by Kat Wilson and Kevin Litman-Navarro
One of the best parts of any concert is when the artist takes a familiar song and plays it in a new way that elevates and refines the original. Think of Robert Plant adding "does anyone remember laughter" to Stairway to Heaven or the way Foo Fighters have turned "Everlong" into a nine-minute, half-acoustic, call-and-response with the audience. This tool helps users find the best live performances, or at least, the live performances most different from the studio version of favorite songs.
Rewrapped (powered) by Anderson A.I.
Although the website does not provide much information, the technology on display here is nevertheless impressive. All it does is take your Spotify playlist, scan it, analyze it, and then present you with new artists similar to the ones you already like. The difference between this and radio-replacement services like Pandora, is that they rely on algorithms derived from your behavior, while Rewrapped asks the A.I. to analyze the music.
Never Been Seen by the Science Museum Group
Perhaps my favorite out of all these tools, Never Been Seen is a tool that reaches deep into the archives of several museum collections and returns something that has not been seen on the internet before. This speaks to just how vast some of these catalogs are and to how fast they're growing as well as giving us a chance to participate, at least via proxy, in discovering historical treasures.
Answer Socrates by Yury and Tobias
Google has a lot of specialized search tools, but only if you know where to look. Answer Socrates attempts to be both research tool and discovery engine by mining the data returned from Google's databases to find the questions that people are asking. Designed to help researchers find questions related to their own searches, this tool has the potential to help both armchair and professional journalists alike.
And that's seven. There are many, many more new tools out there, but these seven caught my interest because I think they are each, with the exception of Deep Nostalgia, an example of how the ways in which we search and the web are evolving. Much as the natural language searches allowed by Google were a huge improvement over paring down your search to its most basic keywords for Altavista or Ask Jeeves, each of these tools shows just how interesting A.I. and algorithm-directed searches can be. It's a brave new world. I wonder what we'll find next.
subscribe | unsubscribe | comments | twitter | about
New issue of The Glossary on Friday. Until then, stay strong, stay curious. Learn something.
Joel
Bonus: For a look at just how we were finding things on the internet ten years ago, look here.