The Tech - Online EditionMIT's oldest and largest
newspaper & the first
newspaper published
on the web
Boston Weather: 55.0°F | Fog

Some of the TR35

EmTech was not only about what technologies currently exist, but also about the technologies that would come to be. Technology Review has picked 35 of the top innovators under the age of 35 (also known as the TR35) that will likely change the world in the future. All of them were invited to the conference to give a 4-minute elevator pitch about the technologies they were developing. Their work targets up-and-coming and new industries ranging from ubiquitous computing, to cloud technologies, to personalized healthcare and beyond.

Gert Lanckriet, an associate professor at UC San Diego, is developing a technology that will automate music recommendations — a task that the company Pandora was forced to do manually. Lanckriet’s technology, called “Herd it,” uses advanced algorithms that intelligently tag music through waveform analysis. Lanckriet envisions a future in which users will not choose music; instead, music will be chosen for users depending on their mood and activity level. The algorithms will decide what you want to hear.

Brian Gerkey, director of Open Source Development at Willow Garage, envisions a future of open-source robotics. Just as open source software has propagated throughout the internet like wildfire, Gerkey wants to spread the availability of tools and software that will enable anyone to build and program their own robots. His company has already started implementing his vision and has released an open-source platform called ROS — it’s already being used at institutions like MIT, UC Berkeley, and Stanford.

Jesse Robbins, Opscode founder (and former firefighter) is developing a product that enables small companies to leverage the power of cloud infrastructure in order to scale. The product, called Chef, is an open source platform that is currently used by over 6,000 organizations, including IGN, Admeld, and Cycle Computing.

Umar Saif, who previously worked at MIT’s Computer Science and Artificial Intelligence Laboratory, has moved back to his native Pakistan to develop network infrastructures for the “other four billion.” He said that today’s technologies make assumptions that do not apply to the developing world, including broadband communication and uninterrupted power. Saif could not personally attend EmTech, but he submitted a video in which he briefly outlined his BitMate project, a peer-to-peer system that is compatible with BitTorrent. Currently, it gives service to 30,000 users in 174 countries, and it has brought SMS to Pakistan, which Saif calls “the Twitter of Pakistan.”

Dan Berkenstock, co-founder and chief product officer of Skybox Imaging, sees a problem in the current commercial satellite industry: “Satellites are too big for their own good.” According to Berkenstock, each satellite costs about a billion dollars to build and launch into space. Satellite imagery is also inefficient — a single satellite can only see a small fraction of the earth’s surface at a given time. Despite this, Berkenstock sees an incredible potential for satellite imagery as a big data source. At Skybox Imaging, he is working on reducing the complexity of imaging satellites in order to “dramatically reduce the price point to put these satellites in orbit.” Berkenstock also plans to transform the collected data into structured content, making it accessible for broader applications. No satellites are in space yet, but the Skybox team is in the process of building their first.

Pieter Abbeel, assistant professor at UC Berkeley, is teaching robots how to learn. “When a robot sees the world it doesn’t see people or chairs, it sees pixels. When a robot acts in the world you can’t ask it to throw a ball or pick up a mug, you have to ask it to move its motors around.” The technique through which Abbeel teaches robots, called apprenticeship learning, doesn’t teach the robots specific tasks. Instead, it teaches them general learning techniques. So far, Abbeel has trained robots how to learn to perform advanced helicopter aerobatics, laundry folding, and the arranging, sorting, and bunching together of socks.