A string of acquisitions and announcements points to big changes ahead.
A recent spate of acquisitions and announcements in AI and machine learning is setting the stage for a colossal showdown across the tech industry.
Among those vying for top spots are Samsung, Google, Apple, Microsoft and Amazon, each with a large enough revenue stream to support an M&A feeding frenzy and the sustained investments required to remain competitive. Consider the most recent announcements:
• Samsung said it will buy Viv Labs, whose founders created Apple’s Siri;
• Google launched its Google Home initiative and said it will build its own smart phones;
• Microsoft announced it has invested in CrowdFlower and CognitiveScale, and it is hiring an additional 5,000 engineers and computer scientists for its new AI and Research Group.
• Apple added Siri to desktop computers with its MacOS Sierra release;
• Amazon created a $2.5 million prize to advance conversational artificial intelligence, and according to published reports, the company is investing heavily in this space.
Nor, is it just the big companies digging into this market. The Université de Montréal was awarded $93.5 million to research deep learning and optimization for the knowledge revolution.
Individually these moves are interesting. Looked at together, they point to a concerted effort to change how people interact with machines and how machines interact with the physical world. That includes an emphasis on natural language interfaces with a variety of connected devices, as well as more machine learning, M2M communication, and increased interaction with the outside world. Behind all of this is infrastructure for better and less intrusive security, as well, including multi-factor authentication.
Put in perspective, after nearly two decades of being mothballed, all of the pieces are in place to make AI and machine learning (ML) a reality. That includes faster chips, denser and cheaper memory, better interconnects, and significant advances in software. And it includes enough progress in end markets that are ready to adopt this technology, from smart devices in the home to self-driving cars to industrial, commercial, and mil/aero applications. Even the business justification for why this technology will be adopted is being worked out, with a pipeline of new products that will connect into these systems.
How this will play out isn’t entirely clear, but there is little doubt that initially it will be a battle of the giants, each promoting their proprietary environment. No one will object in the short-term because there are enough opportunities on the sidelines to keep everyone happy and enough competitors to silence the regulatory bodies. But even though the underlying technology is based on standardized connectivity and commodity hardware, it will be these tech giants that will define how and where they will engage with other companies.
There are some benefits to this proprietary push. Maintaining seamless functionality across new technology areas on a grand scale is complicated and expensive. It’s hard to imagine Google Home functioning efficiently with a third-party implementation. Apple doesn’t make better phones, but it has done a better job vetting applications that can be downloaded onto its phones than the Android community and preventing security breaches across its proprietary iTunes network.
In the long term, there also are some good arguments against this. Proprietary environments tend to stifle competitiveness. As one industry executive quipped, if Nokia still dominated the phone market, we would have smaller cell phones. And if Blackberry had won the race, we’d all be typing on our phones.
But putting that aside for the moment, it’s clear that AI and machine learning are very much back in vogue. Both have the potential to transform technology on a grand scale, and along the way they also could transform a broad swath of markets, fragmenting some and creating links to others that today show little affinity. The battle has begun, and given the stakes and the size of the players, it looks as if it’s going to last for quite a while.
Building Chips That Can Learn
Machine learning, AI, require more than just power and performance.
Plugging Holes In Machine Learning Part 2
Short- and long-term solutions to make sure machines behave as expected.
What’s Missing From Machine Learning Part 1
Teaching a machine how to behave is one thing. Understanding possible flaws after that is quite another.
Neural Net Computing Explodes
Deep-pocket companies begin customizing this approach for specific applications—and spend huge amounts of money to acquire startups.