How different is the work of Google, Facebook, Apple and Amazon in AI ? Who is going to win the AI race?

The last few weeks I spent time interacting with the AI teams of these different tech giants. Here is my quick summary.

Amazon and Microsoft control the cloud market [through which AI is going to be delivered], but they don’t have frameworks like Tensorflow [Google] or Caffe/Torch [Facebook] to give them a strong leg up.

Amazon and Facebook have the key channels through which AI is mostly accessed by the public [Alexa or Facebook messenger].

Amazon and Google have the best speech APIs and NLP.

Microsoft and Amazon provide the best computer vision APIs.

Microsoft and IBM have the best sales teams in this space and work with the widest range of partners to build the AI ecosystem.

Microsoft and Google provide ways to train models through services without worrying about the underlying ML frameworks.

Facebook provides support to the widest range of opensource AI projects, but don’t play the services game. Thus, they might not dominate the AI market.

IBM Watson is the oldest and perhaps the most complete of AI tools/services, but don’t engage well with small developers and thus their applications are limited. Their focus is mostly on the enterprise.

Google’s enterprise sales is weak, but it has perhaps the best of AI technology available both inside and outside. The question is just how well can they interact with the ecosystem and help build mission-critical applications.

Apple has a good AI team inside but unlike other companies, they don’t publish a lot of talk outside their company. No one knows what they do and from what is available public they are perhaps the weakest of the majors in this segment. Not surprisingly Siri has lost out to its competitors in terms of usefulness.

In short, it is a game where no one company really dominates. But, Google perhaps has a slight leg up over the others if everything is taken into consideration.