The key scientific insight from studying cellular automata and simple computational rules is that even with extremely simple rules (like Rule 30 with just one black cell as initial condition), very complicated and seemingly random behavior emerges, contradicting the intuition that simple rules produce simple behavior.