A few questions:
Besides looking really cool, what are the practical uses of "game of life" or "langton's Ant"? I understand how agent-based modeling itself is a potentially useful methodoly, not how the 2D game-of-life representation is useful itself?
When trying to develop interesting complicated designs in the game of life, (e.g. https://www.youtube.com/watch?v=C2vgICfQawE ) is it safe to assume that there are certain "building blocks" that can be positioned to produce a predictable pattern?
does research in this field generally follow the pattern: assign some rules, simulate some data, see how the data matches up with real world stuff, change the rules and repeat?
The Game of Life is but one example of: Cellular Automata, which are very useful for modelling complicated, non-linear, systems in physics, chemistry, biology, meteorology, cosmology, computational science, engineering, .... the whole gamut.
Such stable patterns will arise, but are very difficult to predict from the basic laws. Cellular automata are composed of simple deterministic atoms, but higher order behaviour that emerges from the structure (arrangement) of atoms, isn't simple at all.
Something like that, yes. That's the scientific principle at work: observe data, model system, run model, compare result to new data, tweak model, repeat.