Voice Command Tech Will Require Industrial Overhaul in Computing

Sensory Announces Multi-Language Support For Speaker RecognitionVoice command technology is going to fuel a major industrial reorganization in computing, according to a Wired article by Cade Metz. Because of the unique demands such technology places on computer processors, the many large companies getting involved in this field are going to need to embrace different kinds of processor chips.

It’s the emergence of voice-controlled artificial intelligence systems that is going necessitate the change, Metz writes. AI systems like Microsoft’s Cortana and Apple’s Siri require a huge amount of computing to operate – computing done out of sight of the user, at the companies’ enormous data centres. It takes a lot of space, energy, and, of course, money. The current model of using traditional CPUs from the likes of Intel and AMD is “going to hit a wall,” according to one computer research scientist, and so these are going to need to be replaced by GPU chips, which are simpler and use less energy, and FPGA chips, which take up less space, use less energy, and are programmable for specific tasks. The AI systems in question tend to require their processors to go through enormous amounts of information very quickly, and these alternate processor chips are going to be able to do that work much more efficiently.

Forthcoming connected devices like the Apple Watch, smart car systems and pretty much everything that can be encompassed in the Internet of Things are going to utilize this kind of voice recognition technology. As such the many companies wading into the voice command field will need to adjust the computational arrangement on the back end. FPGAs and GPUs require more engineering work, but they offer a viable solution. That could in effect mean more jobs (as companies will need to hire engineers to program these chips), less energy use, and better technology for users. A bright future for voice command tech, in other words.