Once a technology becomes digitized, it moves on an exponential curve | Pascal Finette

So here's your first lesson, and this is really important. Once a technology becomes digitized, it moves on an exponential curve. It's kind of simple, but here's the biggest takeaway for you as an entrepreneur. If you find a technology which is analog today and you turn it into something digital, it's the biggest business opportunity bar none because you're taking something which moves on a linear curve and you turn it into something on an exponential curve. Massive, massive opportunity. So this is your first one. So let's talk a little bit about like, how do you make sense of this change? So we've seen what happens, we've seen, by now we understand like, hey, there's like this exponential stuff happening in technology and it's really affecting us. So how do you make sense of this? This is Ray Kurzweil, one of our founders, one of the leading scientists in the field of artificial intelligence. And he formulated this once as the law of accelerating returns. At its core, the law of accelerating returns says nothing else but once a technology becomes digital, it moves on an exponential curve. But he did something interesting. He took actual data for something, for computers, because he was curious about a question. The question is, if you think about Moore's Law, Moore's Law needs to come to an end because there's physics. At one point, if you think about it, right, like the number of transistors per square inch. At one point, the number of transistors per square inch becomes so small that you come into subatomic and sub-subatomic, right? It doesn't work anymore. So actually, what these exponential curves are, they're S-curves. They're flattening out at one point. So he was curious to see, if I look at different technologies, how do they pan out over time? So he looked at basically all the computing technologies we have, like from mechanical computing to tubes to transistors and integrated circuits, and he used a different number, which is calculations per second for $1,000, which is kind of like if you take Moore's Law, it's the economic expression of Moore's Law. How many calculations can you perform per 1 second if you were to spend $1,000? And then he took actual data over a 110-year period of time and saw And this is really remarkable that effectively Moore's Law stays stable over a 110-year period of time. This is a logarithmic plot, which is the reason why you see it as a linear graph. Now, you can do a couple of things with this. The first is understanding that these sigmoids or S-curves, how they actually work. So what happens is once one technology flattens out, another kicks in, and in some they become a perfect exponential. So tubes came to their end, were replaced by transistors. Transistors came to their end, were replaced by integrated circuits. Integrated circuits will come to their end, will get replaced by quantum or biocomputing. The second interesting thing is, if you're familiar with something called Clayton Christensen's work as the Innovator's Dilemma, he describes the same thing from a different angle. What he is describing is that a company which dominates an era is typically never the company which dominates the next. So you take this in computing: mainframe computing, Honeywell and GE; mini-computing, dominated by Digital Equipment and a few others; personal computing, dominated by Microsoft and the OEM manufacturers; Handheld computing, so phones and tablets, dominated by Google and Apple. The next phase of that is Internet of Things, integrated little devices. I can guarantee you it's not going to be dominated by Apple nor by Google. Every time you have a break in these, every time you're at this point, it's the most exciting time to build a company because it's most turmoil, least certainty, most opportunity, and these curves get stacked quicker and quicker and quicker. So really fascinating world we're moving into. The second thing you can do with this graph is you can extrapolate it out. You can actually, if you believe this graph to be true in the future, and there's nothing which should keep you from thinking that because it's true for the last 110 years, you can actually predict the moment when you have a computer which has the raw compute power of a human brain, which happens in the next 15 years. So in the next 15 years, someone will build a computer which has the raw compute power of a human brain. It gets a little crazier because you can actually determine the moment when you have the raw compute power of 7 billion brains in one machine. Skynet? Which happens 2050 to 2060. So very soon, we will have computers which are, purely in terms of their raw compute power, significantly— will significantly outsmart humans. And then you can ask yourself, like, what do you do with this? Again, like, as an entrepreneur, like, always ask yourself, like, it's interesting, what do I do with it? Peter Diamandis, who's our other co-founder, developed a model called the 6 Ds of Disruption. It's a very nice model to make sense of the change we are seeing. And I want to explain this by talking about Stephen. Stephen is the inventor of the digital camera. What you're seeing behind me is The digital camera, the first digital camera. This camera had 100 by 100 pixels, 0.01 megapixels. It's terrible, horrible. He also worked at a company called Kodak. So here's how this works. The first D postulates, once a technology becomes digitized, you heard that before, it becomes exponential. So you take something which is analog, film, it's a chemical process, And you turn it into a digital good. Now it becomes digitized. In the beginning, these trends are deceptive. The first camera had 100 by 100 megapixels, 0.01 megapixels. A year later, the camera had 0.02 megapixels. It doubled in pixel count. A year later, it doubled again to 0.04. A year later, it doubled again to 0.08. And then the top people at Kodak looked at this and said, this is not going anywhere. We are not funding this project anymore. Because all they could see were zeros, right? Like you go from 0.01 to 0.08, it's all zero. Eventually this stuff becomes disruptive. This is when you had a digital camera which had 2 megapixels. 2 megapixels is enough to do a 9 by 13 print in centimeters. It looks great on the screen. This was the moment, 2 megapixels, when I stopped buying film. The last time I bought film. Eventually then, technology becomes dematerialized. So instead of buying a digital camera, it's now a $1 tiny little chip and a little bit of optics in your phone, and everything else is software. Even the shutter sound of your camera is now software. Once that happens, you demonetize it. So taking a picture is now effectively free. And then it becomes democratized. So today there's about 3 billion pictures uploaded to photo-sharing sites such as Facebook and Flickr every single day. 3 billion pictures. That's democratization of photos. You can play this with other technologies as well. Let's take Uber. Uber digitized the rider information. Where's the car? Where's the rider? Put it on a digital map. It was deceptive in the beginning because Uber's first product was black cars. These big town cars, like a luxury product. It became disruptive when Uber copied Lyft and created UberX. They dematerialized car ownership. If you live in San Francisco, there's a really good chance that you don't buy a car anymore because it's actually easier, cheaper, and better to have an Uber. You demonetize it. Uber is significantly cheaper than a cab. And you democratize it. Their big game is to put a rider into every seat, into every car. So this is a really interesting model because it predicts where an industry is going. So if you take an industry, you can actually see where on that axis it is and where will it go next.