(Source: www.businessinsider.com)

  • Microsoft and Amazon on Thursday announced a surprise partnership in releasing a free new software tool for developers, called Gluon.
  • Gluon makes it easier for developers to build artificial-intelligence and machine-learning systems — aka apps that can learn.
  • But there’s another, more important reason this partnership is interesting: It challenges Google in its one big area of dominance.

Microsoft and Amazon have joined forces to help spread artificial intelligence across apps.

The two companies released a new tool for developers, called Gluon, as an open-source project, meaning anyone can use it, work on it, or contribute to it for free.

Releasing new programming tools like this as open source helps make sure they will do what people want them to do — because the users are the ones contributing features.

But there’s another, more important reason that this partnership is interesting.

Machine learning and AI are the next big things in cloud computing, with the potential to significantly change the cloud business that Amazon and Microsoft have long dominated.

And when it comes to AI and machine learning, Google seems to have an edge. It has bet its cloud-computing strategy on AI. It believes AI will become so important to the cloud, and that it will be so good at AI, that it will leapfrog Amazon and Microsoft and win the next cloud war.

An enemy in common

Google already has a head start, thanks to a tool it created called TensorFlow, which is also free and open source and aimed at helping developers build machine-learning apps. TensorFlow is immensely popular with developers — it’s the fifth-most-popular project (by stars) on GitHub, a site that hosts open-source projects, out of the over 2 million hosted there.

Although a programmer does not need to use Google’s cloud to use TensorFlow — it’s so popular that Amazon’s cloud also supports it — developers naturally lean toward Google’s if TensorFlow is important to them.

Naturally, Amazon has a TensorFlow competitor, called MXNet. That’s Amazon’s modus operandi: support the technologies its customers use, while also offering versions it wholly controls. That way, it makes money either way.

Likewise, Microsoft also has a TensorFlow competitor, called CNTK.

Gluon’s role is to add a layer that makes MXNet and CNTK easier to use, work with, and program. Only the MXNet version was released on Thursday, but the CNTX version of Gluon is promised to come soon.

Because of the immense popularity of TensorFlow, it “already has a layer like this called Keras that has become extremely popular, so this might start to bring MXNet and CNTK to parity with TensorFlow,” said Lukas Biewald, the founder of CrowdFlower, a startup that helps companies build apps that use machine learning to know when a task can be automated and when a human needs to get involved.

So when it comes to Gluon, the partnership between Amazon and Microsoft is a case of “the enemy of my enemy is my friend.”

“Gluon is an attempt by Microsoft and Amazon to regain some influence in AI tools,” one programmer wrote on Hacker News. “They’re ganging up against Google’s tools.”

‘We love TensorFlow’

Matt Wood, the Amazon tech evangelist, insists Amazon isn’t trying to squash Google’s tool.

“We love TensorFlow (and have a ton of developers using it on AWS),” he wrote on Hacker News.

He continued: “Just like databases we’ll support a wide range of engines on AWS; some of our own like Gluon, along side others from the community like PyTorch and TensorFlow. They’re all first class citizens. We even fund separable (competing!) teams internally to focus on making sure AWS is the best place to run each of these popular engines.”

Still, there’s a reason Amazon creates homegrown competitors to popular tech: to give the company more control and its own stake.

And Microsoft and Amazon have been known to cuddle up on other AI tech. In August, the two announced they would partner to make their voice assistants, Microsoft’s Cortana and Amazon’s Alexa, work better together.

In any case, the competition to create more easy-to-use AI tools for developers gets a thumbs-up from Biewald.

“Making neural nets easier to use is really good for CrowdFlower and, frankly, everyone,” he said. “Right now, they’re a huge pain to build.”

More Info: www.businessinsider.com

Advertisements