An Investor’s guide to Artificial Intelligence
We’ve been hearing about artificial intelligence for years… but it never seemed to amount to much.
But in the past year or so we’ve seen an explosion of new applications from some of the biggest names in a huge range of industries.
Here’s just a tiny fraction of the recent developments:
- Facebook uses the technology to automatically identify you and your friends in images.
- Google Translate has just taken an enormous leap thanks to AI advancements.
- Boeing is using the technology to streamline airplane maintenance.
- IBM is using the technology to read medical images like X-rays and MRI scans.
And that really is just the tip of the iceberg.
Microsoft, Apple, Uber, Twitter, and Intel (just to name a few) have all made huge investments in artificial intelligence in the past year.
So what changed?
Why are we seeing all of these breakthroughs in artificial intelligence right now?
The short answer is it’s all due to a revolutionary breakthrough in the way that we train machines, called deep learning.
For investors, the exciting stuff is the two key ingredients that deep learning needs… and I promise I’ll get to that in just a second.
But first, let me explain what deep learning is real quick.
What is deep learning?
This revolutionary method tries to teach machines in the same way that we teach humans. Think about when you drive your car, as an example.
When you drive down the road you can easily tell the difference between a car and a truck. And you know that there are all kinds of different trucks, and that you should treat different trucks differently.
For example, if a fire truck is behind you with the siren on, you know you need to get out of its way. We see these differences so easily that we barely even notice we are doing it…
But it’s traditionally been incredibly difficult to program a machine to perceive these differences. With deep learning, computer scientists try to teach computers in the same way humans learn. One of NVIDIA’s engineers compares deep learning to teaching his child what a truck is.
If you have ever been in the car with a curious child who just learned the word “truck”… you will understand the basic idea behind deep learning.
At first, the child will point at every vehicle and say “Truck!”, and if he’s pointing at a truck when he says “Truck!” you cheer him on, and maybe give him a little
bit more information. Maybe you say, “That’s right! That is a fire truck.” And if he says “Truck!” while he’s pointing at a car… you tell him that he’s close, but that’s actually a car. And eventually, the child learns the difference between a car and a truck. And learns the difference between a fire truck and a dump truck.
It turns out we can use a very similar technique to teach machines. At this point we have billions of images of trucks. So, a programmer can start by giving a computer the basics about what a truck is. And then the programmer can feed the truck images to the computer. And as the computer “sees” more and more truck images, the computer gets better and better at learning what a truck is …
Until eventually, the computer can tell the difference between a car and a truck just as well as a human can.
And as you can probably imagine… this has incredible applications.
For instance, you probably saw in the news recently that a Budweiser truck, fully loaded with beer… recently drove ITSELF 120 miles from Fort Collins to Colorado Springs… making a beer delivery with NO DRIVER.
It’s possible for trucks to drive themselves because they are beginning to be able to “see” as well as humans.
Now, self-driving cars probably get the most media attention… but that is really just scratching the surface of what deep learning can do.
Here’s another example from the world of medicine.
There is a type of diabetes that can cause blindness. And, thankfully, we have a drug that can prevent the blindness…
But here’s the thing… it only works if we can detect the warning signs really early. A computer is now able to look at pictures of diabetics’ eyes and figure out which people have the early warning signs of blindness… this means they can be treated early and give themselves the best chance to avoid going blind.
Deep learning’s two key ingredients
Deep learning is powering these amazing recent innovations… but it turns out that deep learning actually isn’t a new technique either. The method has actually been around since the 1960s. So why is it just starting to power so many innovations today?
There are two key ingredients that make deep learning work.
Remember, the computers learn through “seeing” millions or even hundreds of millions of examples. So the first thing you need is a mountain of data to feed into the machines. Thanks to the Internet and the explosion of easily accessible data, we have plenty of examples to show machines. And, as you’d expect, many of the biggest names in tech have access to an enormous amount of data. And that makes them particularly well positioned for the coming AI boom.
The second key ingredient you need is processing speed.
It takes a lot of horsepower to show all those examples to computers. And it turns out that the central processing units (CPUs) you are probably familiar with from your home computer are much too slow to do the job effectively. Andrew Ng and his team at Stanford made a huge breakthrough when they realized that you could
use graphics processing units to do the job instead. It turned out that 12 NVIDIA GPUs could do the work of 2,000 CPUs!
And this incredible speed advantage has led to the deep learning explosion we are seeing today.
Our three favourite stocks for AI:
ALPHABET
This will come as no surprise: Alphabet has been building deep learning tools for years, and recently those investments have been bearing fruit in a multitude of ways. In fact, these developments are now so critical that, on the most recent earnings call, Sundar Pichai, CEO of Google, said…
FACEBOOK
Facebook, the social media king, is working its own angle when it comes to AI. Even though the company’s capabilities may not trump Alphabet’s today, that doesn’t mean that CEO Mark Zuckerberg and his team are sitting idly by. He recently said:
“So the biggest thing that we’re focused on with artificial intelligence is building computer services that have better perception than people. So in the basic human senses, seeing, hearing, language, core things that we do, I think it’s possible to get to the point in the next five to 10 years where we have computer systems that are better than people at each of those things. That doesn’t mean that the computers will be thinking, or be generally better, but that is useful for a number of things.”
NVIDIA
NVIDIA is the king of the mountain when it comes to producing graphics processing units. And the discovery that GPUs can be used to power deep learning has been huge for the company.
NVIDIA’s CEO, Jen-Hsun Huang, summed up the situation in a recent interview:
“At no time in the history of our company have we been at the center of such large markets. This can be attributed to the fact that we do one thing incredibly well–it’s called GPU computing.”
And this dominance in the area of deep learning has some investors really excited. Here’s what famed venture capitalist Marc Andreessen had to say:
“We’ve been investing in a lot of start-ups applying deep learning to many areas, and every single one effectively comes in building on NVIDIA’s platform. It’s like when people were all building on Windows in the ’90s or all building on the iPhone in the late 2000s.”