Skip to content

Add an AI-agent-instruction case study #45

Description

@jt14den

How could the content be improved?

Computational thinking's core moves — decompose the problem, sequence the steps, spot the pattern, abstract to a reusable procedure — map closely onto instructing an AI coding agent well. You still have to do that same decomposition to prompt an agent effectively; the agent replaces the manual execution step, not the thinking step.

Proposal: add a section or exercise that reframes one of the existing pen-and-paper problems as "now instruct an AI agent to do this," rather than rewriting the lesson's core. Keeps the lesson unplugged and broadly applicable, while setting learners up for lessons like Agentic Research Workflows that go deeper on AI-specific workflow.

Scope note: this should stay a bridge or case study, not a merge of the two lessons. Part of Computational Thinking's value is that it doesn't require code or an AI tool at all.

Next steps

  • Draft one exercise reframing a decomposition problem as an AI-agent prompt
  • Decide where it fits — new episode vs. addendum to an existing episode
  • Get feedback from co-authors before merging

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions