,
27 tweets,
11 min read
Read on Twitter
There's an unexpected connection to planetary science here. The exponential growth of Moore's Law has been consistent for so many decades, it demands an explanation. As John Nash was represented in "A Beautiful Mind" saying, "I need to look through to the governing dynamics." /1 
2/ *Something* must be at work to cause this to have the *same* exponential rate for so many decades. To be this consistent, it can't be random. But what is at work? The mathematically obvious (but inadequate) answer is this equation. It predicts exponential growth whenever...
3/...whenever something reinforces itself. When the rate of growth of something is proportional to how much of it already exists, then it will grow exponentially. Like bacteria multiplying in sugar water. The more there are, the faster they reproduce. This is simple, but....
4/...it can't be enough to explain Moore's Law. Moore's Law says the number of transistors that we can squeeze into a computer chip keeps growing exponentially. But the rate we squeeze transistors into a computer chip *doesn't* simply depend on how many are already squeezed in.🤷♂️
5/ A while back I searched the science & engineering literature desperate to find papers that "look through to the governing dynamics" to explain *what* is behind Moore's Law, because this simple equation isn't enough to explain it. I found only one published paper on this topic.
6/ That paper used math to show it is the growth of complexity of our entire civilization that is self-reinforcing. Moore's Law is just one indicator of the *overall* growth of complexity, just going along for the ride with everything else. This is a very satisfying explanation.
7/ For one thing, we *do* see exponential growth in many things related to civilization's overall complexity, especially in information systems. (Three of these are semilog plots so exponential is a straight line; one is a linear plots so exponential is a curve.)
8/ It is believable that as technology becomes more complex, then it is capable of creating complex technology at an every increasing rate. Therefore, the overall system can be (believably) described by this simple equation predicting exponential growth. But...
9/...it is super amazing how this *overall* self-reinforcing growth of complexity projects into a metric like Moore's Law, the way your car's roadspeed projects into the speedometer needle on your dashboard, so it looks as if transistor density is a standalone exponential growth.
10/ There's much discussion how Moore's Law can't continue forever. Eventually it hits the limit of physics & then it will end. This is true, but why pick on Moore's Law? Countless technologies hit their limits every day, yet the overall system keeps growing exponentially.🤷♂️🤯
11/ The answer to this is that every time a technology hits its limit, then it is replaced by a successor technology so the overall exponential growth of complexity in civilization continues. An example from power density in transmission systems. (Credit: spectrum.ieee.org/semiconductors…)
12/ If this succession of technologies is what enabled exponential growth of complexity for so many decades, such that Moore's Law & every other tech are just tiny pieces riding along, then it's amazing how there are always successor technologies appearing when they are needed...
13/...and it's amazing how a succession of such diverse technologies manages to keep a *consistent* rate for so long. Well, the rate isn't perfectly consistent, so that makes it more believable. E.g., here are successor technologies *before* Moore's Law. This is >100 years!
14/ This is super speculative, but it has to be because there's only 1 published paper so far; it's near the leading edge of thought. But here it is: it's been claimed the overall organization of civilization is also a series of successions analogous to successor technologies.
15/ OK, so here's the connection to planetary science. Because of @Google, we now have ability to do bibliometrics studies to learn how scientists have used terminology over the centuries. I've been studying the evolution of taxonomy, and how taxonomy function within science.
@Google 16/ The idea is that taxonomy is actually part of the scientific process, because words are the way we organize our thinking and they are the way we formulate theories about nature. Taxonomical categories like "mammal" or "planet" are the nouns, and theories are the verbs.
@Google 17/ Again, this is an area where there are almost no published papers. Philosophers of science are unfortunately stuck on obscure questions about "ontology" in taxonomy, rather than practical questions how taxonomy is a tool in the way science functions. But I found a few papers.
@Google 18/ This lack of prior work makes the research more difficult because I need to do a lot of the primary research myself. So I have spent days and days using Google Scholar to produce datasets to show how taxonomy evolves, and how it relates to practical science questions.
@Google 19/ And in the course of this research my mind is being BLOWN by the exponential growth that occurred in science. Here is a plot of the words Astronomy, Planet, and Satellite counted year-by-year in the literature. This is a semilog plot, so it shows amazing exponential growth.
@Google 20/ In the late 1600s, there were only a few papers each year using the word "planet". Now, it is in the range of 100,000 every year. (The dip at the end is because recent papers are still being uploaded and counted.) There are some cool things we notice in this plot!
@Google 21/ First, note there's no possible way that any scientist can keep up with this flood of information. It is beyond human ability to read this much and have enough any time left over to do science. This means the role of the individual human in the scientific process is evolving.
@Google 22/ We might be tempted to say that human brains are no longer able to keep up with this "Moore's Law" of scientific publication, and therefore our individual brains have already been replaced by a successor technology (involving groups of people & computers).
