Keith Speights reports in The Motley Fool:
Data. Artificial intelligence simply can't do anything without data. Lots of data. Unlike traditional computer programming, AI doesn't require smart people to think through every step that the computer should follow. Instead, AI demands that those smart people develop algorithms and feed a lot of data into those algorithms. (And) The best kind of data of all is the data that no one else has. Companies with this kind of data will be able to develop AI systems that are best in class.
Data.
So much for creating suspense. The absolute top thing to look for when investing in artificial intelligence (AI) is data. I'd go so far as stating that data will be one of the most precious commodities -- probably the most precious commodity -- impacting practically every industry under the sun throughout the rest of the 21st century and beyond.
But why is data so important for investing in AI? And how can investors assess which companies will be the kings of data in the future?
Fuel for the revolution
Every business revolution has its fuel. For the computing revolution, that fuel was silicon chips. For the AI revolution, the fuel is data.
Artificial intelligence simply can't do anything without data. Lots of data. Unlike traditional computer programming, AI doesn't require smart people to think through every step that the computer should follow. Instead, AI demands that those smart people develop algorithms and feed a lot of data into those algorithms.
Two terms that you hear much about with respect to AI are machine learning and deep learning. Machine learning is an application of AI that enables computer systems to learn from experience, in a similar way that humans learn. Deep learning is machine learning on steroids.
The use of the word "deep" in the name refers to multiple layers in an artificial neural network. The important thing to know is that learning, whether it's done by a computer or a person, requires a tremendous amount of data.
One of the most publicized uses of AI in learning today is in self-driving cars. Alphabet (NASDAQ:GOOG) (NASDAQ:GOOGL) and Tesla (NASDAQ:TSLA) train their autonomous car systems by running huge amounts of data through their machine-learning networks. Tesla reportedly had accumulated 1.3 billion miles of Autopilot data to use in feeding its AI systems as of November 2016. The company no doubt has much more data now.
Both Alphabet and Tesla use data captured from human and autonomous driving to help their systems learn when to turn, when to slow down, when to speed up, and when to return control to the human driver. With enough data, over time, the AI systems will rarely, if ever, have to relinquish control of the steering wheel.
The best kind of data of all
Nearly every kind of data is valuable to AI systems. The best kind of data of all, though, is the data that no one else has. Companies with this kind of data will be able to develop AI systems that are best in class.
Alphabet, for example, has more data about what individuals across the world search for online than any other company. The more people use the Google search engine, the more powerful Alphabet's AI systems will become. That's true for other Alphabet products such as Google Pixel Buds, which translate 40 spoken languages in close to real time. As the Google Pixel Buds become more widely used, Alphabet will have more data to make its translation AI systems even better.
Which company knows more than any other about people's shopping patterns? Amazon (NASDAQ:AMZN) is probably at the top of the list. The more data that Amazon captures, the more effective it will be at meeting customers' needs in its current areas of focus -- and expanding into new areas.
Think about Facebook (NASDAQ:FB). The social-media giant knows who most of the friends and family are for 2 billion users. Facebook also has data about what interests those users. The company can (and does) use that data in its AI systems to determine how to keep users on social media longer -- which means they view more ads, therefore making Facebook more money.
Three kinds of stocks to buy
My view is that there are three kinds of stocks to buy to take advantage of the enormous growth potential that AI offers. First are the stocks of companies that already have data that others don't have. Alphabet, Amazon, and Facebook are great examples of that category.
Second are the stocks of companies that can get more unique data than others, even if they don't have the data yet. The best places to look for these stocks are in niche markets. For example, Veeva Systems (NYSE:VEEV) provides cloud-based applications for the life-sciences industry.
Veeva has over 550 customers from big pharma companies to small biotechs. I can't think of another company that has as much access to niche data for the life-sciences industry as Veeva does. There's significant potential for Veeva to use its data to power AI systems that allow it to further dominate its market.
What's the third kind of AI stock to buy? Those that enable other companies to accumulate lots of data and process that data using AI systems. I think this could be the trickiest area in which to invest over the long run, though, because technology changes could result in different companies becoming leaders.
All three of these kinds of stocks have a common denominator, of course. They all center around data. Find the stocks of companies that master data, and you'll find success in investing in AI.
0 comments:
Post a Comment