Amazon has created a 1/8 scale self-driving race car powered by an Intel Atom processor and deep learning software, called DeepRacer, and the company wants you to buy one and learn how to program it for the pre-order price of $249. If you fail to reserve your own unit before the March 6, 2019 release date, you can expect to pay $399. The machine runs a custom OS called Robot OS, built atop Ubuntu 16.04, alongside Intel's own OpenVino computer vision toolkit. As for I/O and other hardware that you'll be using to teach the little bot how to do its job, it's got 802.11ac Wi-Fi and a bevy of USB ports. The camera that it uses to see the world around it is no match for the likes of Waymo's LiDar tech, but it is a pretty decent unit for its purpose, weighing in at 4 megapixels with output resolutions up to 1080p. On a full charge, the autonomous racer can zoom about for up to 2 hours. Developers who get their hands on one of these can go on to compete in the DeepRacer League, which will be holding a number of real-world racing events through 2019 and beyond.
Background: This bot comes with an arguably more useful resource for developers; a fully configured and loaded cloud environment geared toward reinforcement learning. The primary use case here is obviously for DeepRacer, but it can be used as a blueprint of sorts to make just about any kind of reinforcement learning environment you may want. Within that environment, all you have to do is craft your algorithm and watch it go. You can even design simulation courses for it and monitor its progress, growth, and codebase changes in real time as it racks up laps around the virtual course. In order to build and maintain a community around these diminutive speedsters, Amazon will be holding competitions at re:Invent where developers can craft a killer routine to win their very own DeepRacer kit, among other prizes.
Impact: The impact here is extremely simple; more developers, and even non-developers looking to get into the computer sciences, will have a tangible and relatively simple point of entry for self-driving, AI, and Amazon Web Services technologies. This will all result in those areas getting an influx of new talent, which will in turn mean more growth and more frequent breakthroughs in those areas. To put it simply, it's a win-win for everybody; tech recruiters will have less of a hard time finding qualified AI developers, current talent will have bright new workers to brainstorm alongside, and consumers will reap the benefits as those new ideas and vast improvements on existing systems trickle their way into the consumer market. Amazon will also win out in the end with the eventual fruits of this product; more developers will be in on its AI conventions and deeply entrenched in the AWS ecosystem, giving the company a competitive edge in its fight to become the de facto standard bearer for everything from web hosting to AI projects.