How much is a nail? A square centimeter, maybe an inch and a half? Either way, the millimeter is not important here. Imagine that billions of tiny dots could be lodged on a surface of this size and they were also connected to conductive tracks. Their length totals kilometers. Now you have a first idea of the super tiny dimensions that most of us simply take for granted: computer chips.
A chip like the one in the laptop that this column is written on is not just a bunch of dots, that is, transistors. It consists of several complex units that can be applied to the silicon substrate in this way or in a completely different way. Even planning what goes where takes several weeks. Chip developers with many years of experience and a creative mind are therefore in demand for what is called floor planning.
But now they could at least get some help. Google scientists have trained artificial intelligence (AI) to solve the problem. According to her, she now knows at least as well, but often better than humans, to arrange the different elements and connect them to each other in a space-efficient and energy-efficient way.
Ten thousand chip designs
The training was somewhat tedious, after all, the AI was presented with around ten thousand chip designs, each of which had to be graded beforehand. The task of the AI then was to create designs with the best possible marks. It’s no surprise that an AI works well here, after all, there are almost countless ways in which conductive paths can work – to name just one example.
Anyone who has ever dealt with futurists like Ray Kurzweil and others can now hear the bell: Isn’t that exactly the development that people like Kurzweil have always predicted: that machines will eventually build ever better machines? It seems plausible, because if chips get faster, better, and more energy efficient, AI will benefit as well. And he can make even better crisps again …
Because machines do it at a different rate than human cramps, the evolution of technology (which is already much faster than normal) is entering a kind of turbo mode. Oh, what does turbo mean here, we’re talking about exponential growth, and the assumption would be that the point is now reached where the curve starts to go up and steeper and steeper. Among other things, Boris Johnson’s insane experience will likely provide a vivid example of this when the number of corona infections in Britain skyrockets.
But there are also arguments against the fact that fleas are constantly growing without control. Previous designs are gradually improved on the basis of proven models. If an AI internalizes this, some truly creative approaches could possibly be blocked. In addition, AI will also have to contend with the physical limitations that the technology is already struggling with today. At some point, it just can’t get smaller when the boundaries between transistors are barely larger than a single atom. Or is an AI really developing a completely new creative idea? Rather unlikely.
Because that’s exactly what AI lacks: out-of-the-box creative thinking isn’t its strength. Leaked through huge mountains of data in a very short time, yes. Recognize connections that a person might not have seen, yes. Complete a symphony if he has been given enough examples beforehand of a musician’s way of composing, including this one, but he cannot really invent something new, because that is the essence of art, of the human being in general. Which is with some probability because AI achieves its results with completely different methods than humans do with their 20-watt bio-calculator.
But they cannot be ignored. The fundamental rejection is just as harmful as the technocratic hurray of patriotism. After all, AI is also just a technology, just a tool that can be used for both good and bad purposes.