TensorFlow is Google’s open source software library that makes it easier for coders and developers to design, build, train and execute neural networks and deep learning models. It’s basically a set of codes that you can integrate into an application so it can learn tasks like identifying faces and objects in pictures, recognizing verbal commands, and translating words and statements from one language to another.
Originally, the tool was exclusively available to Google’s own developers. On November 9, 2015, the search giant released it to the world as a free tool under an Apache 2 license, which essentially meant anybody could use the code.
While many eyebrows were raised at this seemingly inane move, Google believed it was a strategic way to ensure that AI would evolve faster. With more people (from within and outside the company) using the open-source tool, there was an increased likelihood that they would be able to help improve on Google’s technology faster. Besides, it’s not like they gave everything away, in a manner of speaking. What they shared were just the algorithms running atop their engine. Access to the infrastructure that runs the engine remained within their boundaries.
Google’s giveaway seems to be paying off now as it isn’t hard to see that more programmers are relying on TensorFlow to build stuff with machine learning. And it’s not just because it’s easy to use and technically superior in more aspects than one. More importantly, anything built using TensorFlow can be run anywhere — on your smartphone, your desktop or your laptop – and on Google’s cloud platform, of course. It is because of this that Google now has the chance to catch up with Amazon and Microsoft on the cloud computing front.
Diane Greene, Google’s cloud business head, previously said they are expecting to take the lead in cloud computing within five years. And encouraging interest in AI will be part of their strategy to achieve this objective.
While TensorFlow isn’t the only tool for creating machine-learning software, it’s clearly a favorite among researchers, programmers, coders and even students. The tool is practically helping revolutionize the field; what took months or even years to do before, can now be done in just days.
As more companies invest in AI, and more will have to unless they want to be left behind, cloud services will become more indispensable. And for those who build apps using TensorFlow, it will make sense to use Google’s cloud service for easier transition and a kind of continuity.
So with the growth of AI, the growth of Google’s cloud computing service will not be far behind. Did we forget to say that Alphabet Inc’s (NASDAQ:GOOGL, GOOG) future and consequently its stock keeps looking rosier by the minute?
TensorFlow (TF) on GitHub has over 63k stars! It is adding stars at a rate of 8x the next best AI framework. TF is already the canonical AI framework. Tutorials, podcasts, taught at University, etc. The AI framework is basically owning the AI OS.
Honestly the battle is already over and Google won.