**Abstract** : The intended goal of this manuscript is to build bridges between two definitions of complexity. One of them, called the algorithmic complexity is well-known to any computer scientist as the difficulty of performing some task such as sorting or optimizing the outcome of some system. The other one, etymologically closer from the word "complexity" is about what happens when many parts of a system are interacting together. Just as cells in a living body, producers and consumers in some non-planned economies or mathematicians exchanging ideas to prove theorems. On the algorithmic side, the main objects that we are going to use are two models of computations, one called communication protocols, and the other one circuits. Communication protocols are found everywhere in our world, they are the basic stone of almost any human collaboration and achievement. The definition we are going to use of communication reflects exactly this idea of collaboration. Our other model, circuits, are basically combinations of logical gates put together with electrical wires carrying binary values, They are ubiquitous in our everyday life, they are how computers compute, how cell phones make calls, yet the most basic questions about them remain widely open, how to build the most efficient circuits computing a given function, How to prove that some function does not have a circuit of a given size, For all but the most basic computations, the question of whether they can be computed by a very small circuit is still open. On the other hand, our main object of study, cellular automata, is a prototype of our second definition of complexity. What "does" a cellular automaton is exactly this definition, making simple agents evolve with interaction with a small neighborhood. The theory of cellular automata is related to other fields of mathematics, such as dynamical systems, symbolic dynamics, and topology. Several uses of cellular automata have been suggested, ranging from the simple application of them as a model of other biological or physical phenomena, to the more general study in the theory of computation.

**Résumé** : The intended goal of this manuscript is to build bridges between two definitions of complexity. One of them, called the algorithmic complexity is well-known to any computer scientist as the difficulty of performing some task such as sorting or optimizing the outcome of some system. The other one, etymologically closer from the word "complexity" is about what happens when many parts of a system are interacting together. Just as cells in a living body, producers and consumers in some non-planned economies or mathematicians exchanging ideas to prove theorems. On the algorithmic side, the main objects that we are going to use are two models of computations, one called communication protocols, and the other one circuits. Communication protocols are found everywhere in our world, they are the basic stone of almost any human collaboration and achievement. The definition we are going to use of communication reflects exactly this idea of collaboration. Our other model, circuits, are basically combinations of logical gates put together with electrical wires carrying binary values, They are ubiquitous in our everyday life, they are how computers compute, how cell phones make calls, yet the most basic questions about them remain widely open, how to build the most efficient circuits computing a given function, How to prove that some function does not have a circuit of a given size, For all but the most basic computations, the question of whether they can be computed by a very small circuit is still open. On the other hand, our main object of study, cellular automata, is a prototype of our second definition of complexity. What "does" a cellular automaton is exactly this definition, making simple agents evolve with interaction with a small neighborhood. The theory of cellular automata is related to other fields of mathematics�� such as dynamical systems, symbolic dynamics, and topology. Several uses of cellular automata have been suggested, ranging from the simple application of them as a model of other biological or physical phenomena, to the more general study in the theory of computation.