Yes, they do; they form compounds, most of which are "meta-stable" --- they persist until some new atom enters the environment and reacts with them, breaking them up.
Oh, there are many more blocs than those, and many factions within those blocs.For human "atoms" this is the formation of power blocs - be they companies, religions or industry groups.
That's true. The trick is to devise a structure for government that constrains it's scope and powers, regardless of who dominates.With dominance comes governance.
That a system has some of the features of an organism doesn't make it an organism. Societies are complex adaptive systems, not organisms. The definitive property of an organism is that it adheres to a design. Both CAS's and organisms have multiple interacting parts, but in an organism the parts are designed for specific tasks and are locked into their relationships with the other parts. A liver cell in an animal cannot decide it is bored with its role and decide to become a nerve cell. A kidney cannot rearrange itself into a lung. A human heart cell cannot decide it would be happier in a walrus. Their roles, properties, and relationships are fixed, per the design specified in their DNA. The elements of CAS's, in contrast, are generalists --- they can play many roles, and their tasks and relationships are dynamic --- variable, transitory, and unpredictable. While they exhibit a definite pattern at any given moment, that pattern is not dictated by any pre-existing design, and is unpredictable except in the very short term.I would say that societies are more like simple organisms. They have now effectively evolved "nerve nets" and perhaps a very weak and small central brain. That would make human societies more like plants, sponges or cnidarians than brained animals.
And of course, the "organic"-"atomistic" analogy is just an analogy, and like most analogies, can't be pressed too far.