Why is AI an attractive investment area today?

My name is Jeff Herbst. I'm the founding general partner of GFT Ventures.  We believe that every business model in the world is going to be disrupted by some of the leading-edge technologies that we're involved in, and those are artificial intelligence, data science, and blockchain. So we believe these are fundamental, groundbreaking things that are going to change the way businesses conduct their business.

AI, over the last several years, has changed dramatically. The problem was there wasn't enough computing horsepower to actually train what's called a convolutional neural network in somebody's lifetime or in multiple people's lifetimes. The big innovation that happened was the computing; the horsepower got deep enough and strong enough that you could actually train these networks and all they were trying to do was recognize the difference between dogs and cats. And it may sound pretty simple, but by using hand-coded computer vision techniques, the accuracy rates were pretty bad. In walks a group of people using an AI power, like graphics processors, training deep neural networks, and they were able to get speed and accuracy that far exceeded everyone else. And that was the big bang of AI that unlocked everything that we're seeing today.

And on top of that, not only was the computer horsepower strong enough but there were massive pools of data, some of it labeled, some of it unlabeled, that people can tap into. Essentially at the end of the day, AI is pattern recognition. And it looks at multiple things, whether it be pictures, or voice clips, or packets of numbers, and it looks for commonalities and patterns within those. And if humans do it, it could take us years or lifetimes to look at all the data and figure out the patterns. By using convolutional neural nets and computers that are trained, you can look at all the data in the world that all the people in the world would have looked at and consolidate it into one answer and one neural network, and one model. And that's really the innovation here. 

Imagine you are a doctor and you worked in the middle of rural Mississippi. You know, maybe you don't have an MRI machine. You know, you're the right doctor for the whole town. You've seen a case of lupus maybe once in your lifetime. But what if you can leverage all the collective knowledge of all the other doctors in the world? If you saw something on somebody's body, on their skin, or their eyes were dilated in a weird way. This is kind of the power of AI. Take the collective knowledge of all of us, train a network to recognize situations that we may never have seen before. 

I think the average investor should want and need exposure to frontier technologies, specifically artificial intelligence, data science, and blockchain. The reason is artificial intelligence, in particular, is probably the most disruptive technology that we've ever seen in our lifetime. This is the first time that we can actually say that there's software that can write software. And this fundamentally changes everything that we've ever thought about computer programing and building technology businesses. Because previously, in order to write software, in order to code, you actually had to type it into the computer. So you just had more and more people going later and later at night all through the night, but what if, you know, computers can learn by themselves. What if, by recognizing patterns, of pictures, of images, of video, of voice, of data, they could extract insights on their own without having something to tell them what they were looking for? This is a breakthrough innovation. It's going to change the world. It's going to change every business model in every industry and every vertical.

So this is so groundbreaking that you want to have exposure to it in some way. And the best way to do it is at the early stage, before the ideas have come to the general public, before anybody knows about it, these are when the best companies are born. So, as an investor, this is the area I want to be in. It's on the bleeding edge, and I hope that others will see this as well. 

You need to be very careful investing in any startup technology, especially AI, because there are so many companies out there that are now doc AIs. And so when you look at these companies, you have to look at what is protectable and what do these companies do that nobody else could do? So I look for companies that are solving hard problems in large markets that have to have something unique, something differentiated, either access to data that nobody else has to train an AI model, which is extremely valuable, or they have to have extremely good vertical expertise in the field that they're going after. So as an investor, this is what you have to look for. You have to look for people with the experience, the knowledge to take these models, to train them in a particular vertical that they have, extensive domain knowledge in, and ideally with some proprietary access to data that no one else has.