Cylon precursor lives to help you discover new music

Next Story

Internet traffic drops by 33 percent in Sweden after anti-piracy law passes

Meet MuZoid. She’s a very clever bot created by some even cleverer people; she’s all about killing all humans helping you find more music to love, CDs to buy and gigs to attend, and she does it all through Twitter, naturally.

muzoidTwitter users sending a tweet with the name of a band or artist to MuZoid receive a reply in less than 2.5 seconds with a link to a bespoke webpage containing gig listings, a discography, a recommendation of similar artists that the user might like. If MuZoid doesn’t know an artist, she makes a note of it and gets busy learning. Once enough information’s available, she’ll tweet back to the original querant with an update.

The app speaks to peoples’ love of instant gratification; it works with the fact that Twitter is a platform built around real-time messaging, and serves up a slew of aggregated relevant content a sight more zippily than the disparate results a simple Googling would return.

MuZoid’s proud parents, MusicMetric, can be found in London’s ‘Silicon Roundabout‘. The angel-backed startup is in the trends analysis business, using machine learning methods combined with text mining to track trends in the music industry and perform in-depth analysis of music fans’ online preferences. The company is currently working toward rolling out subscription-based analytics and trend forecasting software for the music industry.

In the meantime, they launched MuZoid, who co-founder Greg Mead says was born as a by-product of this process, i.e. large amounts of data such as gigs, genres and releases indexed into the MusicMetric data warehouse. The data is also used to infer similarities between artists, but what sets it apart from Amazon or Last.fm, which use consumer preferences and crowdsourcing respectively, MusicMetric’s inferences are all down to machine learning classification — which means that the information becomes more accurate as more sources of data are found.

It’s a fiendishly clever idea, so it’s probably no surprise that the folks behind it are all, well, wickedly bright. Greg’s background is in physics and computational physics. The other founders are Matthew Jeffery, Marie-Alicia Chang, Jameel Syed and Andrew Walker. Jameel has a PhD and 10 years experience in analytics and text mining for the life sciences industry. Matthew is a computer scientist, Marie-Alicia’s background is in financial software marketing and Andrew is co-owner of creative agency Thin Martian and co-founder of Tweetminster.

MusicMetric plans to release a suite of widgets, social network apps and free public APIs making the data available for people to build their own apps. They’ll also be adding more music stores in the near future, which will allow for price comparisons on albums across the board. There aren’t any standards for music or ecommerce APIs, Greg says, so the more problematic ones (like HMV) will take longer than others.

I asked Greg why they decided to assign a gender to MuZoid, idly speculating on the possibility of her being a frontrunner to the type of AI that will one day lead to Cylons walking the earth. He said:

Developing this technology is a creative journey, and it helps to give the AI more personality if it has a gender.  There’s so much male-dominated language associated with technology that we thought it would make a nice change to anthropomorphise this software bot as female.  If we called it a “cute little guy” that would be way too predictable. The Cylons are cool, I saw one walking the earth down Great Eastern Street the other day, Shoreditch is full of them.

blog comments powered by Disqus