The worst thing that happened to science in the last fifty years is the invention of the super-computer, while the great computational power of these machines made it possible to solve some problems using the latest scientific theory it has also been a barrier to the development of newer and more efficient theories.
Almost all resources have been put towards developing better computational implementations and bigger computational capacity; while very little has been devoted to developing theories that need less computation.
People hail as great achievements ‘advances’ that in reality has returned us to the curious phenomenon of the building-computer, last seen in the age of the vacuum tube.
Although such computers achieve many results and help in the advancement of the scientific knowledge they also impede scientific progress by prolonging the usage of inefficient theories, still worse they have helped to form a false belief in the generation of scientists that grew up with super-computers that the solution to all scientific questions is large-scale computations and if current computers can’t handle it then the only way forward is bigger computers.
This mistaken view is, of course, the result of a complex social dynamic in which commercial interest, research funding needs, national prestige and many other factors come into play; one of the most important factors is the slow deterioration in the quality of scientists since the end of WWII, who, unfortunately, have spent ten times the resources of their predecessors while only advancing science by a tenth as much (keeping in mind that applying and advancing science are two different things).
Modern science has a very shaky philosophical foundation; I view it as a big tree with many overhanging branches that are loaded with fruit, but the roots of this tree are dead and the rot is beginning to spread upward.
One of the most serious philosophical challenges to face the modern would-be scientists is to understand the need for the correct scale of abstraction. A phenomenon in nature has scale and if one wants to study it then he should do so at the appropriate scale.
For example, one can spend a lifetime studying a single H2O compound without ever obtaining a single property of hydrodynamics, because the phenomenon of liquids has a very different scale to that of chemical compounds.
The high computational requirements of the theories in usage today are the result of scale mismatch; low-scale theories are used to solve high-scale problems. If one would model water as discreet chemical compounds instead of a continuum then to obtain any resemblance with how water behaves in real life the mathematical model would involve billions if not trillions of individual compounds, which would need computational power on a galactic scale, fortunately equations that describe behaviour of liquids have been derived in times when scientists were of much better quality.
The jump from atoms to liquids is a mighty jump that is unbridgeable by any computer, but one can make a smaller jump in scale, lets say from liquid in a closed system to the weather, such a jump is and has been bridged by super-computers.
Although great advances have been made in the science of weather-prediction, there has been almost none in the science of weather (if not worse!); one wonders if, given the huge investments already made, it is still acceptable to bridge the scale-gap scientifically?
Let us take the case of Hurricanes, this phenomenon has a scale: a large size, a substantial amount of energy, high speeds, etc., while the theory used to describe it was developed for a totally different scale: a small size, much less energy, low speeds, etc., hence the need for high computational power to bridge the scale-gap.
We should all ask the scientists who study hurricanes and demand bigger computers: Why can’t you develop a theory that describes hurricanes on the scale of hurricanes? With the available technology the scientists can track the phenomenon of a hurricanes from birth to death, tracking it in every sense: tracking its path, its intensity, its interaction with the environment, etc., they have more measurements on one single instance of this phenomenon than all the measurements obtained on the phenomenon of planets when Newton presented his three laws of dynamics; yet they can only demand bigger computers!
After hurricane Katrina I was briefly interested in the science of hurricanes, without any effort I was able to list few statement that would describe the behaviour of hurricanes, they were the following:
The Laws of Cyclone-Dynamics
- A cyclone moves in the direction of the maximum heat differential.
- The intensity of a cyclone is inversely proportional to the diameter of its centre.
- The diameter of a cyclone is proportional to the amount of water vapour in it.
- The change in total energy of a cyclone is proportional to the area of the cyclone multiplied by the heat differential with the water surface.
I don’t claim that they are correct or specific/general enough to be used usefully in predicting the development of a hurricane over time, but I do, proudly, claim that these statements are on the correct scale for the phenomenon under investigation!
PS: The lack of scale appreciation is also present in sciences that do not use super-computers, most modern economic theories are based on the individual scale instead of the social scale, hence the total failure of modern economic theories.
