Gérard Berry, a professor at the Collège de France, looks back at the rise of computer science and the role that computers and algorithms will play in our societies. In his view, the brain and the computer are made to work together, not because they are alike, but because they complement each other. There is therefore no point in setting them against one another.
Tangente: How do advances in computer hardware interact with abstract scientific research in theoretical computer science? G.B.: "In theory, all computers have been equivalent since they first appeared. But physical reality has a powerful influence on the development of computer science, and today's technology owes a great deal to Moore's law. Machine learning using deep neural networks, for example, made a major leap forward thanks to advances in graphics processing units (GPUs), while the computerization of science owes much to increases in computing power and data-storage capacity."
Shannon expressed the idea that information has a reality of its own, just like matter or electromagnetic waves. Yet it cannot be observed physically. Is this, then, the first major example of a fundamentally abstract concept that we have managed to embody and put to practical use? "One statement here is false: of course information can be observed through matter! Living organisms continually create information, which can, for example, be stored in sediment: what is a fossil if not a stone bearing information about an animal's shell? Information must be embodied in matter and energy, but it can take many forms, at least at the level of human information. Some physicists also seem to think that, at a very fundamental level, matter is merely a manifestation of information. But for now, that idea can take you anywhere without producing many results. In science, thinking is not enough: above all, ideas must be tested, and we are a long way from that… Personally, I am not very interested in the fanciful speculations that can be spun arbitrarily from a few good ideas…"
Can we study the algorithms of living systems? "We might say that life combats entropy by creating information. It uses algorithms to do so. The way an individual's spinal column develops, for example, can be described very effectively in algorithmic terms. For decades, people have spoken of the Holy Grail of putting biology on a mathematical footing, following in the footsteps of physics and chemistry. One possible way to overcome the difficulties involved may be to adopt an algorithmic approach, which could prove useful for formalizing certain aspects of living systems."