The "To Keep Up" Wiki

This outline is designed for a presentation to the Lexington Computer Group, incorporating the topics and suggestions discussed in recent emails. It emphasizes practical demonstrations and a comparison of current AI tools, while also delving into more advanced concepts.

• Briefly introduce the purpose of the presentation: to explore the landscape of current AI engines and discuss their capabilities and implications.
• Acknowledge the group's interest in AI and the collaborative nature of this topic (as suggested by recent email discussions).

• Acknowledge the rapid evolution of AI and the need to understand its various facets.
• Briefly touch upon the presenters' journey into understanding AI.

• To provide a comparison of popular AI chatbots.
• To demonstrate practical applications and advanced features of AI.
• To foster discussion and shared learning within the group.

Note: AI Chatbot = AI Bot = AI Engine

• Comparison paper
• Washington Post summary

• Interactive Demos: For each chatbot, perform a standard set of tasks/questions.
  • Example Tasks:
    • Generating creative text (e.g., a short story, a poem).
      • write a short story about a boy and his dog
      • tell me a joke about a doctor, an engineer, and a scientist
        
        
    • Summarizing a complex article.
    • Answering factual questions.
      
      
    • Performing a simple coding task or debugging.
      • Write a stand-alone program in Java that asks for a number and computes all prime numbers less than that number.
        → How would you test that the program works correctly?
      • Write a stand-alone program in Java that asks for some text and then produces an alphabetized list of the words in the text along with the frequency of use for each of the words.
        → How would you test that the program works correctly?
      • Write a java program that asks for a number and then produces all Fibonacci numbers less that number.
        
        
    • Write or analyze a legal document
      • Write a straightforward purchase-and-sale agreement between me (selling the car) and a buyer (buying my car). It is important that the buyer knows that the car comes as-is with no warranties or guaranties from me.
      • A professor gives a presentation and a videographer records the presentation and later produces a video of the presentation. Write a copyright agreement that allows both to use the result and yet protects the rights of each.
        Followup: please simplify the agreement using non-legal language.
        
        
    • More personal side Questions to ask AI engines
      
      
• Discussion. Attendees impression of Strengths (from experience):
  • Share personal insights on which chatbot excels in specific areas (e.g., creativity, factual accuracy, coding, conversational flow).
  • Discuss user interface and ease of use.
• Sentiment Analysis Comparison:
  • Briefly present findings or a live demonstration of how different AI engines respond to sentiment analysis queries, as Peter Albin explored with Perplexity.

====III. Beyond Basic Chat: Advanced AI Capabilities (20-25 minutes)==== ===A. Retrieval Augmented Generation (RAG) Technology:=== * Concept Explanation: Use visualizations to explain what RAG is and why it's important for overcoming AI hallucinations and providing more accurate, grounded responses. * Demonstration: Show an example of an AI (e.g., Gemini) leveraging RAG to answer a question by citing specific sources or external knowledge. ===B. Deep Research Capability:=== * Concept Explanation: Discuss how AI can be used for in-depth research, going beyond simple search queries. * Demonstration: Showcase a “Deep Research” session (as John Day mentioned using Gemini for the energy consumption report). * Walk through the process of posing a complex query. * Show how the AI gathers and synthesizes information from multiple sources. * Present the results and discuss their utility. ===C. Programming with AI to Overcome Ambiguity:=== * Concept Explanation: Address the challenge of ambiguity in generative AI responses and how programmatic approaches can help refine and control outputs. * Demonstration: Provide a simple example of using AI for programming tasks, perhaps demonstrating how to refine a prompt or use structured output to get a more precise response. ====IV. Q&A and Open Discussion (15-20 minutes)==== ===A. Facilitated Discussion:=== * Open the floor for questions from the group. * Encourage attendees to share their own experiences with AI tools. * Discuss potential future topics for the group related to AI. * ===B. Future Outlook:=== * Briefly touch upon the ongoing advancements in AI and what the future might hold. * Emphasize the importance of continuous learning in this rapidly evolving field. ====V. Conclusion (5 minutes)==== ===A. Key Takeaways:=== * Reiterate the main points: AI tools offer diverse strengths, and advanced techniques like RAG and deep research enhance their utility. * The importance of hands-on experience and critical evaluation. ===B. Thank You and Next Steps:=== * Thank the audience for their participation. * Suggest resources for further exploration (e.g., specific AI platforms, online courses). * Invite continued discussion within the Lexington Computer Group. Visuals & Style Notes: * Use a mixture of high-level visualizations for concepts and live, on-line demonstrations (heavy emphasis on demos). * Keep explanations concise and focus on illustrating capabilities through practical examples. * Maintain an engaging and interactive tone.