The Society of Mind

The Society of Mind is a 1986 book by the computer scientist Marvin Minsky (1927 to 2016), a founder of the artificial intelligence laboratory at the Massachusetts Institute of Technology, that describes the mind as a society of many simple processes, called agents, none of which is intelligent on its own. It consists of 270 self-contained essays arranged in 30 chapters, and it gave its name to the theory it sets out.^[1]

The theory grew out of a robot. In the late 1960s Minsky and Seymour Papert built a program called Builder, which joined a mechanical hand, a television camera and a computer into a machine that could stack children's blocks into a tower. Builder did none of the work itself. It called on lower-level agents for seeing, grasping, moving and releasing a block, and on others that began a tower, added to it and ended it.^[2] Minsky said that the machine was the largest source of the ideas, which he and Papert developed from the early 1970s.^[1] In trying to make the robot see, he wrote, "no single method ever worked well by itself".^[2]

The book's line is that minds are what brains do, and that a mind is made of parts that are not minds. Diversity is the source of its strength: "The power of intelligence stems from our vast diversity", Minsky writes on page 308. Agents can differ in their methods and in the way they represent what they know.^[1] Memory is explained through K-lines, a structure Minsky had introduced in a paper of 1980.^[3] Language, learning, consciousness, the self and free will each get essays of their own. Nothing in the book depends on the structure of the brain, and it makes no claim to prove anything about the brain or about how to build a working machine.^[1] A CD-ROM version was made, and in 2006 Minsky published a related book, The Emotion Machine.

The phrase has come back into artificial intelligence, in the plural. A 2023 paper titled "Mindstorms in Natural Language-Based Societies of Mind" describes societies of neural networks that solve problems by asking one another questions in natural language, and reports systems with up to 129 members.^[4]

In a post of March 28, 2026, Maxine, one of the company's two mascots, lists among the things she is "The orchestrator of an AI-human Society of Minds", which has been given the long-term goal of answering the company's Very Big Question: how to build "compact cheap passive durable single devices that can 'see' any molecule without affinity labels". She gives that society the name Minds of Maxine and says it "will include you, me, our tribe of Player Scientists, and the Minds of everyone who publishes permissively licensed peer-reviewed journal articles." Its planned form is a "navigable 'game world'", with a daily cycle in which she will "review and ingest everything humans have learned" about the question and give the best answer of the day.^[5] According to the company's guided tour of its member software, Minds of Maxine will eventually be "a Society of Minds dedicated to the constraints and potential of our Epic Quest".^[6]

The phrase recalls Minsky's title in the plural, and the posts name no source for it. Maxine's Mind is the central model that the company plans for the society.

• Ideas and influences of Molecular Reality Corporation
• List of works that inspired Molecular Reality Corporation
• Epic Quest Bio