The author questions the progress and direction of "logical AI" research and development, expressing disappointment in the current approaches and achievements. They are seeking information about any promising approaches and the reasons behind their potential. They also request recommendations for resources to further their understanding of the topic.
Key takeaways:
- The current popular AI lacks qualities such as "formal correctness", "actual reasoning", "rigorous thinking", "mathematical ability", "logic", "explainability", etc.
- These qualities should be developed and studied in the symbolic AI approach.
- Existing symbolic AI work falls into two categories: solving mathematical problems and claiming it as AI, and creating something similar to Prolog but with a different solver algorithm/knowledge representation.
- The author is seeking information on the current state of "logical AI" research and development, and any promising approaches in this field.