In markets and many other aspects of technology, inequalities are a permanent feature in the overall mechanism. When inequality goes out of hand, there are only a few winners left. When inequality becomes the rule of the market, over time there is only one winner remaining. Monopolies inherently aren't bad until they turn into rent-seeking in nature. The market for AI also seems to be heading in the same direction.
Technology markets are notorious for turning out to be winner take all markets. This is because of network effects and other factors that come into play once a company achieves market dominance. IBM dominated computing for decades. Microsoft dominated PC markets. Amazon still single-handedly dominates the e-commerce sector and is now eyeing the logistics sector as well. As you can see, growing companies with good products and dominance are hard to defeat and they often stifle innovation in the ecosystem.
There are enough lessons in history that hint that AI would also be a winner take all market. AI markets very readily lend themselves to act like broader technology markets. Andrew Ng, one of the leading machine learning researchers introduced something known as the virtuous cycle of AI. The initial AI products at the beginning are built using limited data. These products when in contact with users gather more and more data every day. This data has valuable insights and patterns for machine learning algorithms to leverage. The product gets better every time more data has been fed. This leads to a virtuous cycle of AI and this property will be an important factor in making AI the mother of all winner take all markets.
Why AI Is The Winner Take All Market
- Economies of scale: As companies build AI and it gets better with data, there is almost zero marginal cost to produce and distribute the same AI product leading to higher profits and dominance once the scale increases.
- Vision and mission by founders: Most of AI founders today are very ambitious and would like to take over the whole market because of the mission they are on.
- Best Minds: Almost all the best minds in the world working in various sciences are converging on AI and data sciences. With more brilliant minds working at the same place (like Google), there will be chances that a smaller number of companies will dominate.
- Network effects and ecosystems: Most of the companies today are focussing on building ecosystems rather than individual products which will lead to great network effects and as a result dominance.
Ecosystems and Not Individual Products
We are seeing companies after companies launching multiple products which work on multiple levels and platforms working on together. This is the trick. The trick to getting dominance in AI is to cover various aspects of the market as much as you can.
Take Intel for example, where on one hand they have launched their own distribution of Python programming language. That's not all. Going even further they have also decided to build AI specialised software. They have launched OPEN VINO their own version of computer vision library that works well with the Python distribution. All this is an effort to create and maintain an ecosystem from where developers and users find all their needs fulfilled. Ideally, for Intel, an AI developer of the future will be using Intel hardware with Intel software running Intel's version of his/her favourite programming language.
Is this a good strategy, remains to be seen. Take Nvidia for example, which is going through its own version of AI dominance playbook. With already a strong foothold in the GPU market, it is poised nicely to take over large chunks of the developer and consumer markets in the AI-powered world. Nvidia wants to work on automated factories, kitchens, and robotics where they see the maximum impact in the near future.
Many large tech companies are building their own chips (TESLA, GOOGLE, and others). This is also a part of companies in non-processor AI businesses to take more control of various components of the workflow.
We are in a phase where more and more companies working in the peripheries of AI are looking to expand into adjacent markets with an aim to get the most of AI industry pie. Who takes the market in the next decade, needs to be seen.
Register for our upcoming events:
- WEBINAR: HOW TO BEGIN A CAREER IN DATA SCIENCE | 24th Oct
- Machine Learning Developers Summit 2020: 22-23rd Jan, Bangalore | 30-31st Jan, Hyderabad
Enjoyed this story? Join our Telegram group. And be part of an engaging community.
Our annual ranking of Artificial Intelligence Programs in India for 2019 is out. Check here.
Provide your comments below
What's Your Reaction?
As a thorough data geek, most of Abhijeet's day is spent in building and writing about intelligent systems. He also has deep interests in philosophy, economics and literature.