Google announced today it will make its artificial intelligence software, TensorFlow, open source. Programmers from anywhere in the world will have free access to use and modify the A.I. code as they see fit.
Machine learning will be one of the most powerful forces of technology in the future. The ability for artificial intelligent systems to essentially teach themselves what they need to know by sifting through more data than we can possibly imagine will create programs with an almost limitless potential.
Google knows this, of course. Many of Google’s tools already use some version of A.I. to work, such as voice-to-text, language translation, and web search.
TensorFlow is touted to be five times faster than the company’s predecessor A.I. Distbelief. While the older system was specifically tailored for Google’s internal network of data centers, TensorFlow is a general tool that can be used “wherever researchers are trying to make sense of very complex data — everything from protein folding to crunching astronomy data,” according to a Google blog post. The software can be used in as narrow a setting as a single smartphone, or on a network-wide scale across thousands of computers.
It shouldn’t come as much of a surprise that Google would open up its A.I. software like this. Android, the company’s open source mobile operating system, is the most widely used smartphone OS in the world. Sure, the company doesn’t make any money directly from Android, but every Android phone comes equipped with Google’s search-advertising services — generating a hearty profit.
TensorFlow would simply be an extension of that ever-growing revenue beast. And because of how increasingly vital machine learning is to so many different applications and tools, the company could see its claws digging into exponentially more industries around the world — whether it’s the startup trying to develop a new messaging app that’s more intuitive, or state-of-the-art labs looking to develop deep learning networks.
For now, only a portion of TensorFlow will be open source. And Google certainly isn’t sharing any information on the hardware that powers the A.I. engine. It’s still, however, a huge new development.
By making TensorFlow open source, the company is actively and enthusiastically courting the innovations and insight of outside programmers to improve on TensorFlow and use it as a part of any new machine learning architecture.