Yesterday I read a scientific article from Nature (2023) that said scientific papers and patents are becoming less groundbreaking — the authors analyzed 45 million works over 60 years and proved it quantitatively.

I got interested because it directly relates to my dissertation on venture investments, here's what I found:


A breakthrough is when your work makes previous ones irrelevant.
The authors divided all scientific works into two types: some develop and strengthen existing knowledge, others reset it — subsequent researchers stop citing what came before.
In short:
the real contribution to science is measured not by the number of citations, but by how much your work changed the direction of the entire scientific field.


The volume of publications has grown, but the number of breakthroughs has not.
Despite the exponential growth in the number of articles and patents, the absolute number of groundbreaking works has remained constant for decades (see picture).


Using the same sources of information is one reason for the decline in innovativeness.
Scientists and inventors have begun to rely on an increasingly narrow range of sources: they cite the same top works, increasingly reference themselves, and use older sources.
In short:
reading outside one's specialization is a condition for creating groundbreaking work.


The title of the work reveals its type.
The authors found a shift in verbs:
In the 1950s, words like "discovery," "determination," "creation" prevailed — today they have been replaced by "improvement," "application," "evaluation."
In short:
if the title of your work is dominated by words like "improvement" and "application," you are likely consolidating existing works rather than creating something groundbreaking.


Self-citation is an indicator of intellectual stagnation.
In short: The more a scientist cites their own past works, the less groundbreaking their new ones become. Statistically, this is one of the strongest predictors of a decline in groundbreaking works.
In short: if you only build on what you have already done yourself, you are deepening a rut, not paving a new path.


Thus, when evaluating any scientific work or patent, it is worth paying attention to the list of references and the vocabulary of the title.