In theory, researchers publish their work in order to make their knowledge available to others. In theory, the more people who read a published paper, the better, regardless of whether or not the researcher who wrote it is even alive. In theory, the only reason researchers put their name, affiliation, and contact information on published papers is so that they can answer questions other people might have after reading the paper. It's a great theory...

In practice, researchers publish their work because their jobs and livelihoods depend on it. In practice, researchers publish the same paper with minor tweaks in multiple journals because they need to pad their CV. In practice, most of humanity cannot read journal articles anyway so researchers are only really communicating with each other when they publish -- and they already know how to contact each other.

So that is the value researchers derive from publishing their work: failure to publish means failure to advance one's career. It makes no difference what a researcher publishes, and so researchers tend to gravitate towards problems that do not seem very hard (but they make it sound very exciting in the titles and abstracts of their papers, because the people with hiring/firing power usually can't get through the introduction).

Whose theory are you describing in the first paragraph? Yours? In practice, the needs of society -- new freely available knowledge -- match up imperfectly with the needs of the people who can or should do the research: money, recognition by their peers, time to spend on fun problems, and general satisfaction (which can and almost always does come from a large degree from learning something new that other people will want to learn too).