It isn’t exactly a secret that Microsoft, Google and Amazon (along with some others as well) all want to do one thing:
Dominate the market and the business of offering artificial-intelligence premium services via cloud computing.
But why is it so important?
It is important because the winner of this race may have the standard operating system of the future.
Legend has it that for many years, the wife of Swami Sivasubramanian wanted to take a deep look at all the bears on summer nights that came out of the woods in order to plunder trash cans that the couple had at their suburban home in Seattle.
So what did Swami Sivasubramanian do?
He took some time off over one Christmas break and started to rig up a special system which would allow his wife to do exactly that.
Being the head of the AI division at Amazon, of course, helped.
Swami Sivasubramanian, so far, has managed to design a unique computer model which can actually train itself in order to identify and track bears.
This computer model is smart enough to ignore late-night joggers along with dogs and raccoons.
How did Swami Sivasubramanian do it?
He made use of a specific Amazon cloud service.
That cloud service goes by the name of SageMaker.
The SageMaker is actually a machine-learning product.
Amazon has designed this product to app developers who don’t know much about topics such as machine learning.
After designing the computer model to identify bears, the next step will involve Swami Sivasubramanian installing DeepLens, another one of Amazon’s new products.
The DeepLens is actually a wireless video camera.
Swami will install it ON his garage.
The DeepLens wireless video camera is actually a $250 device.
Amazon will put it up for sale to the general public sometime in June of this year.
The DeepLens wireless video camera from Amazon contains a new deep-learning software.
That software will help the DeepLens to put the computer model’s intelligence directly into action.
As a result of all this, the deep-learning software application within DeepLens will send a notification message to Swami’s wife smartphone device whenever the software recognizes that it has seen one of those ursine visitors.
Now, Swami’s AI-enabled bear detector system has a long way to go before it can count itself as a killer application for artificial intelligence.
But that’s not the point.
The mere existence of such a system is a significant sign that machine learning products have practical capabilities that are fast becoming accessible to common day problems.
Perhaps that is the reason why for the past couple of years (or more) the likes of Microsoft, Google and Amazon have pushed their AI teams to fold advanced features like face recognition in online photographs along with speech-related language translation directly into their respective online cloud services.
Amazon has AWS while Google and Microsoft have Google Cloud and Azure respectively.
But that is going to change now.
Now all three of these companies are in a bit of a headlong rush.
That rush is towards building on the already existing basic capabilities in order to create artificial-intelligence based online platforms.
Any and all types of companies would have the opportunity to use these platforms regardless of the company’s technical sophistication and size.
Sivasubramanian recently said that machine learning had reached the same point that relational database reached in the early 1990s.
According to Sivasubramanian, back then everyone knew that relational database would become very useful for fundamentally every company but a very few number of companies had the foresight and the ability to take advantage of relational databases.
Companies such as Microsoft, Google and Amazon and of course, to a lesser extent other technology companies such as SAP, Salesforce, Oracle, IBM, and Apple do happen to have massive amounts of computing resources.
All of these companies also have armies of talented people who can help these companies to build the previously-mentioned AI utility.
Not only that, these companies also have the required business imperative.
All of them want to get in on what many are calling the most interesting and lucrative mega-trend in technology in years.
According to an analyst at CCS Insight, Nick McQuire, ultimately most companies would leverage the power of cloud computing to make use of artificial intelligence.
McQuire further added that cloud computing would also enable technology suppliers to make money off AI.
Of course, currently, it is difficult to quantify all the (potentially huge) financial rewards.
However, for the companies that are leading AI cloud services, these currently unknown financial rewards to reach unprecedented levels.
At the present moment, the cloud computing market is worth about $260 billion.
Experts like the senior director of product management at Google Cloud AI unit, Rajen Sheth, predict that AI could potentially double that current worth in the next few years.
The situation is further made more interesting by the way machine learning works.
The nature of machine learning dictates that the more data its systems are able to consume, the better decisions they will make for the customers.
Hence, it is quite likely that customers would prefer to lock into one specific vendor.
That vendor would be their initial AI vendor.
To put it another way, whichever company manages to get out of the blocks early and takes the lead will gain such an advantage that other companies would find it very difficult to actually unseat that company.
Arun Sundararajan, who analyzes the way digital technologies manage to affect a related economy at the New York University’s Stern School of Business recently said that the price for these companies would come in the form of becoming the default operating system of the coming era of technology.
The president of Avendus Capital US (which is an investment bank), Puneet Shivam, recently noted that the leaders in AI cloud technology will have everything they need to become the most powerful and robust companies in human history.
Perhaps that is the reason why it isn’t just about Microsoft, Google, and Amazon.
Even though they have a big lead over other companies but that hasn’t stopped lesser companies from trying to pursue dominance in the AI-enabled cloud services market.
Heading over to China, one can see technology giants such as Baidu and Alibaba trying to become major forces in AI-enabled cloud computing services.
This is particularly true as far as the Asian markets are concerned.
Apart from big technology companies, some of the leading enterprise software application companies such as SAP, Salesforce, and Oracle have also begun to embed machine learning capabilities into all of their applications.
The point is, hundreds and thousands of artificial intelligence-related startup companies have shown ambitions to evolve into tomorrow’s artificial-intelligence leaders.
Who Are Likely To Win?
Microsoft, Google, and Amazon all provide online services for recognizing objects and faces in videos and photos.
They also use the same capabilities to turn speech directly into text and vice versa.
Moreover, they also offer services for accomplishing tasks such as natural-language processing.
However, so far, almost none of the activity related to machine learning and AI-enabled products has resulted in technologies companies generating more numbers in the way of revenue.
The AI field has quite a lot of big players and none of them bother to actually break out sale numbers regarding each of their commercial AI online services when it’s time for their earning calls.
With that said, that scenario is very likely to change very quickly for those companies which will create the essential underlying developer tools and technologies to support the mass-scale commercialization of AI-enabled and machine learning products.
This is exactly what Microsoft managed to do for personal computers.
Back in the day, Microsoft created the Windows platform.
And this Windows platform served millions of developers all around the world to build different and interesting PC programs and/or applications.
In many ways, its competitor, Apple, also managed to do the same with the company’s iOS platform.
The iOS platform successfully spawned the era of easy-to-use and intuitive mobile applications.
In the year 2015, Google jumped out of nowhere to take an early lead.
The company wooed developers when it made the decision of making TensorFlow open-source.
TensorFlow is the same software framework that Google’s very own AI experts make use of in order to create various machine-learning products and tools.
But Microsoft along with Amazon have themselves created similar technology products since 2015.
In fact, in order to cut down Google’s lead over the rest, they even joined hands last year in 2017 and created Gluon.
Gluon is another open-source interface.
Amazon and Microsoft designed the interface in order to make machine learning a bit easier to utilize with or even without Google’s TensorFlow.
The trio of Google, Microsoft and Amazon have continued their work to find new ways to make AI-enabled and machine learning products accessible even to those who are total AI novices.
It is clear that Amazon had exactly this idea behind the company’s SageMaker.
As mentioned before as well, Amazon designed SageMaker in order to make it more convenient to build machine-learning applications.
The company wanted the process of creating machine learning applications as easy (or as complicated depending on one’s technical skills) as creating a standard website.
Just a couple of weeks after Amazon announced its SageMaker in November of last year, Google came to the scene by introducing the company’s own Cloud AutoML.
The thing about machine learning is that a company has the opportunity to feed the machine learning technology its own collection of data (which is mostly unique) and then witnessing how that machine learning technology would generate machine-learning models fully capable of enhancing any business automatically.
Technology giants such as Google have recently stated that it has received requests to try out its Cloud AutoML from over 13000 companies.
The head of Google Brain, Jeff Dean, recently told reporters that the company believed about 20 million organizations existed in the world which could greatly benefit from the advantages of machine learning products.
However, these companies don’t have the luxury of hiring people who have the necessary technical backgrounds.
Hence, according to Jeff Dean, in order to reach out to even 10 million of these organizations and start making them use machine learning products, Google would have to make all this machine learning stuff a lot easier to understand and use.
That raises the obvious question:
Which of the above-mentioned big three technology companies has best positioned itself to win that first-mover advantage (which some consider very important)?
There is no doubt about the fact that all of these three technology companies have immense strengths.
Not only that, one would have to work hard to find out any obvious weaknesses in them.
Let’s take a look at Microsoft as an example.
The company has kept on doing breakthrough work on various important AI problems.
It has managed problems such as natural-language processing and computer vision for over two decades now.
Moreover, the company also has access to the ton of very valuable data to feed and inform its Azure cloud service.
The company can also rely on huge amounts of content that it gathers from its other services such as,
One also cannot ignore the fact that there are over 1 billion people who make use of the company’s Microsoft Office software suite.
To put it in simpler terms, there is no other company in the world who knows more about things what one needs to do in order to sell software to businesses along with other organization.
Or even help out developers to sell the same things to their own clients.
All of that sounds pretty great.
But only until one gives a read to Google’s resume.
Almost all industry insiders consider Google as the Research and Development leader in the field of artificial intelligence.
The company has found a lot of success in leading the way to apply AI capabilities to genuinely ambitious problems.
The most notable of all its efforts is building those self-driving autonomous cars.
Apart from that, Google has also developed its very own line of AI chips.
The company will use those chips in order to run all its machine-learning infrastructure.
Many consider Google as the champion of open-source applications especially after the company handed over TensorFlow.
Google can also thank its search engine (still miles ahead of anything out there) for having access to a massive amount of data.
In fact, the company probably has access to more amount of data than any other technology company.
Combined with its search engine (and the fact that it also owns YouTube, which is the de-facto second biggest search engine in the world), the data that Google has provides the company with a rather detailed picture of the society’s collective desires and interests.
One 20-year-old founder of Scale (an AI startup), Alexander Wong, thinks that Google has placed itself in the best possible position and that too by a long shot.
And then we come to Amazon.
Amazon has managed Apple-esque levels of secrecy in the way the company works.
But many industry insiders considered the company as an also-ran as far as the AI-enabled products and tools went.
That was a year ago.
Now, reports say that the company’s secrecy has apparently managed to mask its sweeping corporate ambitions.
If one looks at the past seven or eight years, according to Sivasubramanian, each and every company business planning document has had to to through the trouble of explaining of how it’s related unit would manage to utilize new developments such as machine learning.
In fact, the requirement of how the unit would integrate machine learning products actually appeared on the company’s boilerplate forms.
These were the forms that managers made use of for all such planning documents.
The forms also came with a parenthetical clause.
That clause clearly read that the answer “None” was not good enough.
And while Amazon can’t really compete with Google when it comes to publishing machine learning related research papers, it still has managed to attain a 40 percent market share as far as the cloud market is concerned.
But the company has started to move ferociously to use its position as a market leader in cloud services in order to dominate the AI cloud era as well.
The company has recently introduced a slew of new online services.
Amazon once used these new services only internally.
As far as acquiring AI startups go, Amazon has gone ahead with the most aggression.
The company has spent double than what Google has spent in acquiring AI startups.
It has also spent four times the money that Microsoft has spent in the last couple of years in buying out AI startups.
This is what the CEO of Iterate.ai (a company that provides AI services), Jon Nordmark, said recently.
Amazon is also well on its way in developing Alexa to dominate the coming phase of a great consumer interface:
On the other hand, Google has busied itself in making headlines on how the company used AI players to destroy Go champions all over the world.
All the while, Amazon has used its expertise in the field of factory logistics and robotics to deliver many millions of packages each day.
It has essentially established itself at the pole position for AI projects which meld data collected from various real-world sensors with digital information.
According to Nordmark, some other technologies companies are publishing more research papers than Amazon.
But Amazon is the only technology company that is taking practical steps to put boots on the ground.
Perhaps that is the reason why it is moving ahead so fast.