Wright's Law - why Moore's Law is outdated

Moore's law states that the complexity of integrated circuits doubles approximately every 18 months without an increase in cost. This statement is useful and has served us well over the last few decades. Nevertheless, we know that at some point it will no longer be valid, if only because of the physical circumstances. Furthermore, the link to time is problematic. A better alternative is Wright's Law, which is consistent with Moore's Law but has many more areas of application.

The following is about what exactly Wright's Law is and how it helps us in systems engineering.

Wright's Law: The learning curve effect

The core idea behind this law is the dependence of cost reduction on the frequency with which an activity has already been carried out. This can be the activity of a person, but also a production process.

Practice makes perfect

Old proverb

We are all familiar with this effect when we learn something. It doesn't matter whether it's something motoric, such as juggling. Or whether it is mental work, such as solving quadratic equations. Or a production activity, such as making a chair or an integrated circuit.

The Engineer Theodore Paul Wright He observed that in aircraft production, the amount of work required for each additional aircraft produced was reduced by approximately 20%. He also observed this in the production of other goods and accordingly formulated the law named after him as follows:

For every doubling of the total quantity of products produced, the costs fall by a fixed proportion.

Wright's Law

Wright's Law is everywhere

Now it is certainly understandable that this law also applies to the Moore's law applies. Instead of referring to a time interval, we can also refer to the number of products produced, i.e. the number of integrated circuits.

The law can be applied to many industrial sectors and often fits better than Moore's Law. A concrete example: solar cells. These have been subject to a rapid fall in price in recent decades. However, the price decline stagnated several times between 1990 and 2005, as can be seen in the graph on the left. However, if we plot the installed capacity of solar cells on the x-axis instead of time, we suddenly see the exponential development predicted by Wright's Law (right):

What Wright's Law means for systems engineering

The implications of this law are twofold. Firstly, we can include Wright's Law in a cost calculation. This is particularly relevant in industries with large quantities, such as smartphones or Vehicles. For example Elon Musk called lidar systems a "crutch" a few years ago. At that time, lidar sensors were also quite expensive. With this statement alone, Musk has certainly contributed to reservations about lidar, which has probably slowed down the fall in the price of lidar sensors. Nevertheless, if a development team from a major OEM were to decide in favor of lidar, a corresponding drop in prices would be expected when calculating the costs.

Also worth reading: Langdon Morris: We urgently need innovation in systems engineering

SE trends

The second implication is even more important: we are doing more and more and therefore learning more and more. It is becoming easier and easier to make use of this by combining new products. One example is Artificial intelligence (AI). The number of new libraries and frameworks is staggering, as I recently wrote in my guest article at HOOD. Those who take a correspondingly lightweight approach in order to discover and exploit the relevant developments at an early stage have a clear advantage.

We live in a world of exponential growth

Whether we like it or not, we have only seen the tip of the iceberg with Moore's Law. Wright's Law is suddenly popping up everywhere and causing a mutual acceleration. Anyone who fails to take this into account in systems engineering will be left behind.

Photo by Jakob Owens on Unsplash