IA - Fluency

Introduction

Install Bun

Install OpenCode

bun add -g opencode-ai

Configure

With OpenCode you can use any LLM provider by configuring their API keys.

Run the /connect command in the TUI, and select GitHub Copilot

/connect

Select ‘Claude Sonnet 4.5’

AI Fluency

This video explores what it really means to be “fluent” with AI and why this matters.

We discuss how AI Fluency involves developing practical skills, knowledge, insights, and values that help you interact with AI systems in ways that are effective, efficient, ethical, and safe. We also introduce three ways people engage with AI:

• Automation: The AI completes specific tasks based on your instructions.
• Augmentation: You and AI collaborate as creative thinking and task execution partners.
• Agency: You configure AI to work independently on your behalf, establishing its knowledge and behavior patterns rather than just giving it specific tasks.

The 4D framework

The four core competencies of AI Fluency, or the “4Ds”: Delegation, Description, Discernment, and Diligence.

• Delegation: Thoughtfully deciding what work to do with AI vs. doing yourself
• Description: Communicating clearly with AI systems
• Discernment: Evaluating AI outputs and behavior with a critical eye
• Diligence: Ensuring you interact with AI responsibly

How these competencies work together across different ways of engaging with AI, and why developing these skills prepares you for whatever AI evolution brings next.

Exercises

Apply the 4D’s

Pick one of these collaboration scenarios and consider how you might apply the 4D Framework:

Communication project

You’re working with an AI assistant to draft a series of emails for a marketing campaign.

• Delegation: What aspects of this project would you handle yourself vs. collaborate on with AI?
• Description: How would you communicate your vision for the campaign’s tone, purpose, and success criteria to the AI?
• Discernment: What criteria would help you evaluate whether the AI-drafted emails meet your needs?
• Diligence: What considerations around transparency and responsibility would be important?

Research project

You’re using AI to help analyze a large dataset for a research paper.

• Delegation: How would you divide the analytical work between yourself and the AI?
• Description: What context would the AI need to understand about your research question to do its share of the tasks well?
• Discernment: How would you verify the AI’s analysis for accuracy?
• Diligence: What ethical considerations might arise when publishing AI-assisted research?

Creative project

You’re collaborating with AI to develop character concepts for a story.

• Delegation: What creative elements would you want to explore through AI collaboration vs. develop independently?
• Description: How might you guide the AI to generate characters that fit your story’s world?
• Discernment: How would you decide which AI-suggested elements to keep, modify, or discard?
• Diligence: How would you acknowledge AI’s contribution to your creative work?

Explore something you love

Spend 5-10 minutes chatting with Claude about a topic you’re passionate about and know well. We will be using Claude across exercises for the rest of the course, but you can also do these exercises with another AI. In fact, it might be worthwhile for you to try the exercises across this course with several AI assistants to get a feel for how they differ.

Instructions:

• Choose a topic you know well and enjoy discussing such as a hobby, professional interest, favorite book series, etc.
• Have a natural conversation with Claude about this topic, like you would with someone who shares your interest.
• Try to notice moments where:
  • Claude enhances your thinking
  • You need to clarify or correct the Claude’s understanding
  • Your expertise leads you to evaluate the Claude’s responses

Learn something new

Spend 5-10 minutes asking Claude to teach you about a topic you’re unfamiliar with but interested in exploring.

See how this experience differs from your conversation about something you love and know well.

Instructions:

1. Choose a topic you’d like to learn more about.
2. Engage Claude in a conversation to help you understand the basics of this topic. Don’t worry about prompting the “wrong way” or the “right way.” Just ask Claude to teach you.
3. Try to notice moments where the Claude:
   • Offers helpful explanations
   • Provides examples that make abstract ideas concrete
   • Responds naturally to your questions as they arise
   • Explains things where you would want to double check its explanations

Reflection

Before moving on, take a moment to consider:

• Which of the 4Ds (Delegation, Description, Discernment, Diligence) do you feel most confident in already? Which might need more development?
• Can you recall a recent AI interaction where the framework might have helped?
• What specific skills from the 4D framework would most enhance your work or personal projects?

Generative AI

Generative AI fundamentals

This video introduces the concept of generative AI, focusing on its ability to create new content rather than just analyzing what already exists. We walk through how large language models (LLMs) like Claude actually work and the technological journey that made them possible, from algorithmic breakthroughs like the transformer architecture to vast training datasets and powerful computing. We also explain how these systems learn through pre-training and fine-tuning and discuss concepts like context windows and emergent capabilities.

Capabilities & limitations

This video examines what generative AI can and cannot do effectively at this point in time. We highlight generative AI’s versatility across language tasks, ability to maintain conversational flow, and capacity to switch between diverse tasks without additional training. We also address limitations including knowledge cutoff dates, hallucinations (factually incorrect outputs), context window constraints, and reasoning challenges. We emphasize how the field is evolving rapidly and explain that the most effective applications bring together the complementary strengths of humans and AI working together.

Key takeaways

• Generative AI creates new content (text, images, code) rather than just analyzing existing data

• Modern systems like LLMs were made possible by three key developments:

  • Algorithmic and architectural breakthroughs (especially the transformer architecture)
  • Vast amounts of digital training data
  • Dramatic increases in computational power

• Generative AI learns through two stages: pre-training (analyzing patterns across billions of examples) and fine-tuning (learning to follow instructions and provide helpful responses)

• Current capabilities include versatility across tasks, conversational awareness, and the ability to connect with external tools

• Current limitations include knowledge cutoff dates, potential for hallucinations, context window constraints, and challenges with complex reasoning

• The most effective applications combine human and AI strengths, with humans providing critical thinking, judgment, creativity, and ethical oversight

Delegation

Generative AI Fundamentals

Pending

AI Fluency: Framework & Foundations