>>17077734
Specialization is the way.

>>17077734
Generalization lives at the two extremes.

On the one hand, generalization can very easily to lead to towards never developing any real research expertise. That's obviously not good because you spent a ton of time focusing only on the surface area of different fields instead of properly gaining an understanding of any one of them.

On the other hand, some of the most important breakthroughs in the history of science were a result of generalists using tools which found success in one domain to solve problems in an area where those techniques had never been explored.

Being a generalist can give you both better job security as a researcher, and a better chance at making breakthroughs other researchers might miss. But you have to work 10x as hard to get there because you'll need to get 70% of the way to subject matter expert in a dozen different fields instead of getting 90% of the way there in one.

>>17077745
>>17077741
Both of you are completely off the mark. I'm talking about generalizing mathematical concepts. I've seen quite a few papers on the arXiv (and, regrettably, some in journals) whose merit solely rests in "generalizing" existing concepts. One notable example of a slop-paper like this is 2206 13331

>>17077751
Specialization is the only way. You cannot generalize in math. There are simply too many numbers to check.
Focus on a few to become expert in.
Specialize.

>>17077751
Worse yet it's from my own field of research, which is analytic number theory and class field theory. 2206 13331 is notable because of how offensively shit it is. It was likely written by an amateur, in fact it could've been written by anyone. Specifically, an amateur who only watched some pop-sci explanations of my line of research and decided to "collaborate" with an LLM to write this garbage.

>>17077734
all generalisations are bad

>>17077762
You can specialize in generalization though.

>>17077734
its just low level physics