3 reasons why Facebook, Google and now IBM made their AI free

Today IBM joined Facebook and Google in making its machine learning project free. So now Torch, TensorFlow, and SystemML are all available to developers under open source licenses.

Why would tech companies give out their Artificial Intelligence tech?

There are 3 reasons why. And they all have to do with the fact that this tech is, in reality, not what makes the secret sauce. What does is:

1. Access to proprietary data. As the saying goes: "In theory there is no difference between theory and practice, but in practice there is". Practical, differentiated applications of ML and AI require proprietary data. Even as Google, Facebook, and IBM give away their libraries, they retain both some of the most valuable data sets in the world and their huge data purchasing power (remember IBM buying Weather Channel recently?)
2. Control of the last-mile. ML and AI tech is not very useful if it is not integrated. Facebook, Google, and IBM hold the keys to some very important channels - their consumer and enterprise platforms and their salesforces.
3. Basis of competition. By publicizing the fact that all products can be AI and ML-enabled, Facebook, Google, and IBM decrease the value of non-AI products. This drives the lock-in of AI - which happens to be the axis of their investments - as a key product differentitor in the mind of the marketplace.

Question is - where is Microsoft in all this?

P.S. Thanks for the comments!

"Microsoft open sources Distributed Machine Learning Toolkit for more efficient big data research"

Originally published on LinkedIn.