Information

From Tournesol
Jump to navigation Jump to search

In Shannon's information theory, information is what allows to reduce uncertainty. Information, or data, can also be regarded as the object of analysis of algorithms.

ElmhamdiHoang-19FR argue that information is central to all scientific fields, but also to our universe, to life and to societies. As a result, any revolution in information collection, storage, processing and transmission is bound to have major consequences. Evidently, because of the rise of information technology like Internet, and new information techniques like machine learning, we are undergoing such a revolution.

A broad view on information

In physics, information became a central object of study with the introduction of the second law of thermodynamics, which asserts that entropy increases TedEd-17, and its more detailed study in statistical mechanics PBSSpaceTime-18. In particular, a thought experiment called Maxwell's demon shows that (irreversible) information processing has an energy cost SixtySymbols-16 UpAndAtom-17. Information has also played a central role in the physics of black hole Numberphile-12. In particular, Bekenstein's bound defines a physical limit on the information storage capability in a bounded region of space within our universe. Meanwhile, Bremermann's limit defines a physical limit to information processing speed within our universe.

"It from bit". John Archibald Wheeler.

In biology, genetics has become increasingly regarded as an information science. DNA and RNA are particularly fascinating information structures ZettaBytes-18, which have provided a new ability to collect, store, process and transmit genetic information, and can be exploited to engineer, for instance, RNA vaccines BertHubert-20. Life also found multiple additional ways to perform information processing, as it leveraged hormones, immune systems and neural networks.

"Modern biology is becoming very much a branch of information technology". Richard Dawkins.

In societies, living beings have managed to collaborate to achieve complex tasks, by evolving in packs, flocks or banks. Bees and ants are known for their particularly sophisticated organizations Kurzgesagt-19, but trees also have complex networks with subtle interactions Radiolab-16. Such multi-agent structures solve remarkable distributed computing tasks. It is often argued that humans' selective advantage lies more in this ability to cooperate at scale than in their individual cognitive capabilities Sapiens-11.

"Those who control the flow of data in the world control the future, not only of humanity, but also perhaps of life itself." Yuval Noah Harari EPFL-19.

In human societies, new information technologies have been argued to revolutionize their organization. The first great information revolution is sometimes argued to be the invention of language, which allowed much more sophisticated coordination between individuals of a tribe. A second great revolution was the invention of writing and writing technology like clay, papyrus and paper. This allowed for the long-term storage of information, which may have enabled investment and juridiction. A third important revolution was the invention of the printing press, which allows to copy-paste information at scale. This allowed, for instance, to make the Bible accessible to a very wide public. In all such revolutions, crucially, information was externalized from human brains.

"Writing is the first medium external to the human body intended to carry information (…) As soon as this information is outsourced from our body, everything changes in our civilisation." Michel Serres InriaChannel-10.

The new information technology takes this externalization a lot further. They enable the collection, storage and transmission of information on an unprecedented global scale. Perhaps most importantly, they enable the automated processing of information via algorithms executed by computing machines. This opens new doors for customized information processing, and in particular, for targeted content recommendation. Put differently, especially with the rise of machine learning, algorithms can now tailor what information to send to what user to greatly maximize their goals, e.g., to maximize attention.

More details about the importance of information is given in ElmhamdiHoang-19FR.