Obituary for Cyc (yuxi-liu-wired.github.io)

Article summary

The Cyc project, a 40-year effort to build artificial general intelligence through symbolic logic, has failed to achieve its goals despite significant investment. The project, led by Douglas Lenat, aimed to create a vast foundation of common sense knowledge to enable true machine learning. However, the system never achieved general intelligence, and its commercial applications have been limited to standard methods in expert systems and data integration. The project's insular nature and lack of transparency have also been criticized.

Main themes

• Artificial General Intelligence
• Symbolic Logic
• Machine Learning
• Knowledge Representation
• Cognitive Architectures
• AI Research

What commenters say

• The Cyc project's failure is due to its reliance on symbolic logic, which is insufficient for achieving true intelligence.
• The use of large language models and neural networks has made symbolic approaches like Cyc obsolete.
• Combining symbolic reasoning with large language models could potentially lead to breakthroughs in AI research.
• The human mind is a probabilistic computer, and deterministic symbolic logic is unable to capture the complexity of human cognition.
• Probabilistic and fuzzy logic approaches can effectively represent complex concepts and uncertainty.
• The amount of investment in Cyc was insufficient to achieve significant results, and more resources could have led to better outcomes.
• The failure of Cyc does not necessarily mean that symbolic approaches are flawed, but rather that they may require more development and refinement.
• The use of rules and facts to represent knowledge is fundamentally limited and unable to capture the nuances of human understanding.