Back to timeline

Milestone · July 1984

The Cyc project begins

Lenat set out to write common sense down by hand: millions of assertions about what people know of the world and never say.

Why it matters

The longest bet in the field's history that brittleness is cured by volume of knowledge rather than by a new method.

The plan followed from the brittleness of expert systems, which broke as soon as a question left the described domain, and from the conviction that learning and language understanding would not scale without a base of real-world knowledge. Cyc was to supply the shared ground that texts never state: that an object cannot be in two places at once, that you have to be awake to eat. By 1995 Lenat counted a person-century of effort and about a million hand-entered axioms, and the writing went on for decades. The project neither reached its stated goal nor was shut down; it remains the longest experiment testing that hypothesis.

Event record

Event date
July 1984
Timeline date
Event date
Verification
Sources gathered automatically · September 17, 2026
Lines
ID
evt-0145

The project began in July 1984 at the MCC consortium; the first published account appeared in AI Magazine in 1985.

Sources

Related events

Earlier

Later