Does Innovation Increase with Population? No, but Noise Does
A common assumption in society, especially in economics and futurism, is simple: More people mean more potential inventors, therefore more innovation.
The first premise is almost certainly true. The conclusion does not follow.
Innovation requires not merely the creation of an idea, but its recognition. Every additional person can produce useful ideas, useless ideas, mistakes, frauds, repetition, and everything in between. Production can scale, while evaluation needs not scale proportionally. This creates a simple asymmetry:
More people → potentially more signal.
More people → necessarily more noise.
Is Innovation Increasing?
If population automatically produced greater innovative capacity, the enormous expansion of modern research should provide an extraordinary natural experiment. It has not produced the expected result. Bloom et al. (2020) found that research effort has substantially risen while research productivity has fallen sharply. Maintaining Moore’s Law, for example, requires more than 18 times as many researchers as it did in the early 1970. Park et al. (2023) examined 45 million scientific papers and 3,9 million patents across 6 decades and found that both have become progressively less likely to disrupt existing knowledge. Remarkably, the absolute number of highly disruptive works remained relatively stable while total scientific output exploded.
Note: The colors in the inset represent fields of study; the colors in the main graph represent levels of disruptiveness. Yes, the same colors are being used for different variables.
The graph nevertheless shows a striking pattern: While weakly disruptive papers and patents have increased enormously, the absolute number of highly disruptive works changed little. More research is being produced, but that does not mean proportionally more progress being made. The most societally convenient hypothesis is that there is simply little more to revolutionize. The most societally realistic hypothesis is that people do not like perceived disruption after a certain point, artificially maintaining an upper bound, visible in graphs. This of course does not necessitate a grand master plan. People simply refusing to acknowledge some transformative ideas is sufficient.
But Noise Does Increase
Suppose a city contains 10.000.000 people and only 0,01% possess some rare quality we seek.
There are 1.000 desirable candidates.
Now suppose an extraordinarily accurate screening method produces only 1% false positives.
Among the remaining 9.999.000 people, that still produces approximately 99.990 false positives.
If one candidate received merely one hour of careful evaluation, eliminating those false positives would take one person approximately 11 years of continuous work, or approximately 48 years under a normal human work schedule. However, the problem is more fundamental than finding good candidates: What if the false positives become more successful than the true positives?
Scientific publishing already provides an extreme demonstration (Richardson et al., 2025). Paper mills now produce fraudulent research at scale, accompanied by networks of brokers, authors and editors capable of exploiting the mechanisms intended to distinguish legitimate research from fraud.
Get Andrei Ranete’s stories in your inbox
Join Medium for free to get updates from this writer.
Moreover, what happens if people appreciate bad ideas over good ideas? Indeed, how exactly do people differentiate a good idea from a bad one?
The Forgotten Variable: Time
As explained more generally in “From Geopolitics to Loneliness, There is Only One Metric: Time”:
Evaluation has a non-zero time cost.
Time is finite.
Therefore, evaluation is finite.
Population increases the number of possible innovators, but it also increases the number of people and ideas through which society must search. There may consequently be an optimum population for innovation, or at least a point of diminishing and eventually negative returns, depending upon how efficiently ideas can be identified. This raises an uncomfortable historical possibility. Perhaps past scientific revolutions did not occur despite having vastly fewer researchers. Perhaps having vastly fewer researchers helped make them possible. A revolutionary idea produced within a small intellectual community has fewer competing claims through which it must travel. As the population of researchers and publishers grows, revolutionary ideas need not stop occurring. They merely become progressively easier to lose. The celebrated undiscovered genius therefore represents more than romantic mythology. It represents a noise problem.
More people may mean more geniuses.
It certainly means more people among whom the geniuses must be found.
Population increases possibility.
Noise determines whether we ever find it.
References
Bloom, N., Jones, C. I., Van Reenen, J., Webb, M. (2020). Are Ideas Getting Harder to Find? American Economic Review, 110(4), 1104–1144. https://doi.org/10.1257/aer.20180338
Park, M., Leahey, E., Funk, R. J. (2023). Papers and patents are becoming less disruptive over time. Nature, 613, 138–144. https://doi.org/10.1038/s41586-022-05543-x
Park, M., Leahey, E., Funk, R. J. (2022). The decline of disruptive science and technology. arXiv. https://arxiv.org/abs/2106.11184
Richardson, R. A. K., Hong, S. S., Byrne, J. A., Stoeger, T., Amaral, L. A. N. (2025). The entities enabling scientific fraud at scale are large, resilient, and growing rapidly. Proceedings of the National Academy of Sciences, 122(32), e2420092122. https://doi.org/10.1073/pnas.2420092122
This text is also archived here: https://doi.org/10.5281/zenodo.22232018
This text is part of a series on human consistency. Previous text: https://aranete.medium.com/from-geopolitics-to-loneliness-there-is-only-one-metric-time-749e2f3e9e5a
