Cyc: the 40-year quest to hand-code common sense that AI history forgot
ArtificialOther · x · 2026-11-13
Ian Fisher's long read revisits Cyc, one of the most ambitious and now largely forgotten AI projects.
- Since 1984, Doug Lenat's Cyc has tried to hand-code the common-sense knowledge underpinning human reasoning, aiming at a single human-like reasoning agent
- Its knowledge base holds 1.5M concepts and 25M rules feeding an inference engine with 1,000+ specialized submodules, capable of reasoning chains thousands of steps long; users range from an Ohio research hospital to the NSA
- The effort traces back to a 1983 Stanford meeting on programming common sense; Lenat was already known for AM and other programs with flashes of intelligence
- Despite these achievements, Cyc was eclipsed by the machine-learning generation behind today's chatbots and self-driving cars; the piece examines why the four-decade experiment became a historical footnote
More from AGI Musings
- Mathematician Strogatz Asks on NYT Podcast: What Remains of a Craft When a Machine Does It Better? — stevenstrogatz · 2026-11-13
- Korea's top tax exam scorer accused of wearing AI glasses — with zero evidence — meshive · 2026-11-13
- Attack of the Clouds revisited: what the global Outlook outage says about centralization — technollama · 2026-11-13
- Why AI and Consciousness May Never Meet — marigo · 2026-11-13
- In the Age of AI, Human Capability Is the Real Advantage — uxmag · 2026-11-13
- Robotics researcher: manipulation may no longer be the best area for academic research — AvivTamar1 · 2026-11-13