We can look at the history of technology as a human-driven, parallel experiment of evolution. Artifacts are (so far) not capable of self-reproduction, but the population-level dynamics of long-term technological innovation nonetheless resemble biological evolution in many ways.  The design of new technologies is strongly influenced by existing technologies, and technological change can be viewed as a process of descent with variation and selection. Both chance and the appropriate context are required for innovations to occur. 

Lineages of design often show rapid change and diversification as well as exaptation. The later is illustrated by Gutenberg's printing press, when an existing technology (the screw press) that was co-opted to serve a completely novel purpose. Extinction and replacement are also common. As soon as a genuinely novel invention appears, it is typically followed by an enormous diversification, followed by the extinction (and turnover) of most competing inventions. Moreover, technological change also displays convergence: similar discoveries are made simultaneously by different inventors such as the more than 20 different patents involving light bulb inventions prior to Edison's success. The view that technological evolution follows similar rules to biological evolution has captured the interest of scientists, historians and engineers alike. 

                                     

Despite the commonalities, technological evolution departs from biological evolution in fundamental ways.For technological change long-term goals and expectations play a leading role in which the designers seek optimality, typically under explicit criteria such as efficiency, cost and speed. Moreover, as pointed by François Jacob, in contrast to artifacts, living structures are largely the result of tinkering, i.e. a widespread reuse and combination of available elements to build new structures. Technology is highly dependent on the combination of preexisting inventions, but unlike biology, the introduction of new simple elements can completely reset the path of future technologies.  In contrast, in biology, once established, solutions to problems are seldom replaced.

Both biological and technological innovations involve cost constraints. Thermodynamic efficiency can also help understanding the origin of some structures. Allometric scaling laws provide a good illustration of how a theory of biological distribution networks (including both vascular and respiratory ones) based on efficient energy dissipation on fractal trees. Efficiency has also been driving technological improvements and marks the development of the steam engine and the bicycle. The evolution of the latter can be traced as a succession of improvement steps towards increasing performance and lower metabolic cost. However, the coupling between energy costs and improvements is not a precondition for technological change to occur. On one hand, many examples illustrate a common pattern of development of a given invention: in early stages, inventions are often overly expensive and not perceived as economically relevant. The barrier to diffusion can only be overcome through the vision of individuals pursuing their views and goals. 

Is it possible to formulate a theory of technological evolution? How much can we take advantage from our theoretical understanding of biological evolution? Recent advances within network theory and a unique availability of the fossil record of human inventions might help in reaching that goal. Such theory needs to consider the existence of universal trends, the economic context and history. We believe that a major effort in this direction would settle the debate on similarities versus differences.