One of the major challenges for the emergence of complex multicellular organisms is to generate an enormous diversity of cell types from a single genomic sequence. In the simplest scenario, the different cells would have the exact same protein complement available during embryo development to achieve their distinct functions and morphologies, often as divergent as those of a neuron or an erythrocyte. Therefore, how could neurons, for instance, tweak the structures and properties of this common set of proteins to optimize their specific and distinct neuronal functions without jeopardizing those in other cell types? A well-known evolutionary mechanism to overcome this challenge is gene duplication and functional subfunctionalization of the copies. However, a less well established – and yet probably more flexible and widespread – mechanism with similar consequences is alternative splicing (AS). Through differential processing of introns and exons, AS can produce cell type-specific protein isoforms that allow optimization of their specific cellular functions. One of the most striking examples of this is provided by microexons in neurons. These tiny exons, which can encode as little as one or two aminoacids, are switched on during neuronal differentiation and show the highest evolutionary conservation of all AS types. They are often located in structured domains of proteins, where they subtly sculpt their interaction surfaces thereby modulating protein-protein interactions in a neuronal-specific manner. Although we are still beginning to unveil their biological functions, we already know they crucial for proper neuritogenesis, axon guidance, and neuronal function. The remarkable example of microexons illustrates how a co-regulated program of cell type-specific AS can diversify proteins sequences to generate novel molecular functions as well as optimize ancestral ones for complex cell type-specific tasks.