According to futurists, Moore’s Law means that computers will eventually become smarter than us. And that moment is not far away.
According to Nvidia CEO Jensen Huang, Moore’s Law, an informal rule that the capacity of computers doubles every few years, has expired, according to his colleague from major competitor Intel. But computers will get faster and better in the coming years. When will they surpass man?
Our brain, a giant parallel computer
The nerve cells in our brain, the neurons, are very slow compared to computer chips. But there are many of them and they all work at the same time. Therefore, we can do certain things with our brains that require a lot of computing power in computers. Or that they cannot match themselves.
In fact, this is why the level of human thinking has not yet been reached. But at the moment, at the end of 2022, we are very close to the moment when the information processing capacity of computers is equal to that of humans.
Since the 1950s, when the invention of the transistor made the modern computer possible, technology has developed rapidly. What Silicon Valley, at the time the place where computers were developed, essentially did was cram more and more transistors onto the same surface.
Because technology has continued to improve, every 18 months we manage to install twice as many transistors. This relationship is called Moore’s Law after its discoverer. Since the 1950s and up until about 2020, this law has been shown to have great predictive value. The smaller transistors become, the cheaper and more economical. The result is that every few years you have a computer twice as good for the same money.
Nevertheless, the chip manufacturers will soon run into problems because we do not yet have femto technology. So we are bound by the size of an atom and it becomes difficult to build transistors smaller than 1 nanometer.
Why is Moore’s Law so relevant?
Computers are machines for processing information, just like our brains do. The faster and better the computer, and the better the program, the better a computer is able to take over tasks from the human brain. Brain games like checkers and chess were once thought to be forever beyond the reach of computers.
Currently, supercomputers are capable of wiping the best players off the board with ease. Computers are also beginning to approach or even surpass human skills in tasks such as facial recognition, translation, and even art creation.
The magic ingredient here is doubling computing power every few years. In other words, Moore’s Law. Every job is a matter of processing information. If the calculation capacity is doubled, this means that the task can also be completed in half the time. And that complex tasks are also within reach.
What can we do that computers can’t yet?
There is a list of things that humans can do, but computers can’t yet. Multitasking, explaining why they do something, making moral judgments, feeling empathy, taking initiative. That list is getting shorter and shorter. In 1997, chess grandmaster Kasparov’s pieces flew off the board, followed by 9-dan world champion Lee Sedol’s groups of go pieces in 2016. Both players retired.
Kasparov is now a leading Russian dissident, Lee retired from professional work in 2019. We don’t hear much from him, this kind and humble man may have collapsed. A victim of Moore’s Law that is. In 2022, an AI entry won an art competition, much to the anger of artists. All these things were called impossible in the years before. There is a clear lesson to be learned from this: don’t underestimate artificial intelligence.
Neural chips are making smartphones and computers ever smarter
Cameras are currently available that can recognize faces without the help of a supercomputer. This also applies to Face ID in the latest phones from Apple, Google and Samsung. They use neural chips for this, for example Google’s Tensor chip. This goes pretty far. With more and more powerful neural chips becoming available, the day is not far off when our smartphones come close to the capabilities of our brains.
By extrapolating Moore’s Law, Ray Kurzweil predicted that this point would be at the end of this decade, around 2029. While Moore’s Law is slowing, it is still accelerating. So he wasn’t far away. The software is also getting better and better, allowing engineers to get higher performance from the same hardware.
Nor did Kurzweil foresee at the time the emergence of neural network processors that could mimic THE major advantage of the human brain, a hundred billion neurons working simultaneously. For that reason, 2025 to 2035 seems a realistic estimate.
Find out what all this will mean in this in-depth article from WaitButWhy. Recommended. And think about how you will live in the world of the future, where we are no longer the only intelligence. Like Gary Kasparov or like Lee Sedol?