Science works in mysterious ways. Since we read and write scientific publications, we at least hope that people will read and cite what we have to say. But will that really matter? Hard to say. For more reading check here, here, here, here, here, here, and here.
The nature of scientific evolution
There is a persistent understanding that human ideas evolve just like living things, only the basic core of these changes are not genes but memes. This idea predates internet. The memes involved are the basic concepts and hidden patterns that form knowledge. Funny cats and epic practical jokes might be popular, but they convey very little knowledge, so we are talking about some different kinds of memes.
Different ideas may change, compete, collaborate and even cross-breed. For example, is light a wave or a particle? It is both in different ways. This principle of duality is a result of forced marriage between ideas that competed for centuries. And then the idea of the light evolves and mutates when the light touches the event horizon of a black hole, into something we do not yet fully understand, alternatively the light may break into a pair or a particle and antiparticle.
The paragraph above builds upon centuries of evolving scientific work and several scientific revolutions:
- Newton’s work with prisms,
- Maxwell’s equations,
- Einstein’s work in quantum physics for which he got a Nobel price,
- Einstein’s relativity theory for which he is remembered,
- Feinman’s work in quantum physics,
- Hawking’s equations of the black hole event horizon,
- and several more recent works.
There were literally millions of other contributions huge and small to the theory of light, which I cannot even mention.
Did we evolve from apes?
We can learn a lot from biological evolution. Meet Ida: a complete primate skeleton dated around 47 million years ago. She is a common ancestor of man, apes, monkeys, and lemurs. We did not evolve from apes, which would be a simplistic understanding of evolution, but rather we evolve parallel to other primates from Ida. Aristotle‘s works are to science something very similar to Ida for mankind. Not everything that came from his works is science, but many sciences can be traced to Aristo.
By the time we get to Lucy, or maybe even Ardi, we have a primate who lived ~3-4 millions ago side by side with chimps, but is somewhat very similar to humans. Similar, but not quite human. Modern humans appeared only several hundreds of thousands years ago. Between Ardi and us there were many different kinds of humanoids that disappeared. At least three kinds of humanoids used to breed with modern humans: Europeans have ~5% of Neanderthal genes, and some aborigines have ~5% Denisovan genes. The third kind of humanoids to breed with us was not yet discovered and can be suggested by some genetic studies. All of these humanoids are now extinct, and we also were very closed to extinction a couple of times.
We are not more evolved than chimps. Chimps split from bonobo approximately when modern humans appeared. Having common ancestors makes us more like close cousins, especially since we lost our humanoid siblings to endless waves of extinction.
Does a similar destiny apply to science? Are most scientific theories extinct or close to extinction? Will they rule the science world someday? Did they breed with predominant paradigms?
Is science even relevant?
There are huge revolutions in science from time to time. Darwin’s “On the On the Origin of Species” from 1859 was a revolution in the scientific world. It was being written and rewritten by the author for 25 years because his grandfather was ridiculed for similar ideas. A similar revolution happened in physics, in the beginning of the 20th century with the introduction of quantum physics and theory of relativity. Most of the time we do not see huge paradigm shifts.
We often see small paradigm shifts caused by better equipment and computational capabilities. Neural networks were very popular in 1980s, but only very small networks could be supported. So in 1990s the research in neural networks was almost extinct. In 2012 Alexnet over NVidia GPU caused an explosion of neural networks. We would not see this explosion if the Alexnet did not win an international scientific competition. If you will write yet another publication of yet another neural network it will be lost in the noise, unless you win one of the biggest international challenges. Even then your popularity may well be short-lived.
Most publications get cited by a very narrow group of scientists who cite each other. To become a significant member of a serious scientific breakthrough one needs not only intelligence, knowledge, creativity and political skills, but also a lot of luck.
Diminishing returns
As the technology required for scientific experiments gets more expensive, and the number of mediocre publications increases, the returns on investment in science also decrease. To win a competition with AI network, you will probably need expensive GPUs, super-fast internet access, a lot of electricity and a very good team. Your team will need to scrape the data, improve the network architecture, fine-tune the training procedure and analyze the results better than teams from Google, Facebook, Microsoft, Samsung, Amazon, and other industrial giants. The chances against you will be staggering, so you will need to innovate. Very few scientific establishments can do that.
Thus, a lot of scientific research is done by industrial giants who have very deep pockets to fund the best teams available: either directly or through grants. If successful, the minority of the research results will be shared with the scientific community, and the majority will be used by the sponsor. DARPA-sponsored research is somewhat similar, only you share your findings with the US military rather than a specific company.
You can do a lot in less popular fields of knowledge with very little equipment, but then the returns are also not stellar:
- Your entire field of knowledge is likely to go extinct
- There is an increasing number of people working in similar fields in poor countries
- Even if everything goes well, very few people will ever benefit from your work
I quote: “The problem is that at any time the number of scientific openings, of fruitful questions – questions that lead to new insights, not dead ends – is limited. It may not have kept pace with the demand. ”
A scientist cannot afford to fail repeatedly, so after trying something risky he will eventually generate less risky mediocre publications that is compliant with the ruling paradigms. These mediocre publications are not driving the science forward, but instead, increase the noise through which new discoveries have to cut. Many scientists fail to reproduce even their own experiments. Diminishing returns become a vicious cycle.
Diversity in science
We would expect to have a similar number of men and women, as well as people of all races and ages in science. This is not the case.
New discoveries in mathematical subjects are made by young people. Those who age above 40, usually run laboratories and raise funds but rarely generate new discoveries. In humanitarian subjects, just the opposite: those below 40 struggle, and most discoveries are made by aging professors.
Women in science are under-represented. Having children during career-building years is very detrimental in highly competitive fields. There are few part-time scientific positions, and very few people can work part-time and still have a competitive advantage over full-time employees. The situation is worse because not many women want to see themselves as scientists.
Also, not all ethnicities are equally motivated to pursue scientific careers, which results in a very uneven ethnic landscape. Jews and Asians are culturally biased to pursue a scientific career, while others often prefer to compete over business and marketing positions or work in a family business of more traditional nature.
Science is diverse, but there is a lot of place for improvement.
One possible way to advance science
A disproportionally large body of scientific progress has been done by people that crossed the chasm between scientific disciplines, for example of physicists in biology. One of these revolutions was generated by people who built the atomic bomb and were shocked by ethical issues to become biologists.
Crossing the chasm between scientific disciplines is like changing a biological niche: very hard initially with unexpected benefits. Consider penguins: birds fly through the air, but penguins adapted to fly through the water. Or dolphins: mammal metabolism proves to be advantageous for marine animals.
Interdisciplinary studies fill in the missing ecological gaps and increase diversity in the scientific community. Of cause, this is difficult, because it requires the scientist to become expert in several fields of knowledge, but the competition for those who succeed is less fierce.
[If you know to read fast and remember a lot, crossing between scientific fields is relatively easy]
The greater risk
We all are built to survive, both gene-wise and meme-wise. As much as we want to have children and grandchildren, we want others to use and develop our work. We do not want to be one of many evolutionary dead-ends, and we do not want to be locked in a competition over diminishing resources.
The best way to succeed and flourish is discovering new niches, and this requires crossing complex boundaries. I think this is still less risky than the alternatives, especially if you have the accelerated learning skillset.
