Conclusions
Weak Agents can be Powerful
- A Minimal Model is Good
- Minimizes user interaction
- Facilitates system analysis
- Model is based on tasks
- Not theory of reasoning
- The AI can be Behind the Model
- There's a TMS (Shhhhh)
- More important, a model to guide the TMS
- Model is based on application analysis
- ACL's and Ontologies should be problem-based
- Focus on the Problem
- Use Minimal Standards that Solve
a Large Class of Problems
- Avoid Theories of Everything
cjp