The Scenario
You are trying to process a 3-hour podcast transcript, but it exceeds the model's context window (e.g., an older model with an 8k token limit). You cannot process it all at once.
The Brief
Design a "Rolling Summary" (or Chunking) workflow. Write the prompt that will be used to summarise Chunk 1, and the modified prompt that will be used for Chunk 2 (which must incorporate the summary of Chunk 1).
Deliverables
- An explanation of how you will chunk the text (e.g., by speaker change, by word count, with overlap)
- Prompt A (used for the very first chunk)
- Prompt B (the "Rolling" prompt that takes [Previous Summary] + [New Chunk] and outputs an [Updated Summary])
Submission Guidance
Rolling summaries can suffer from "forgetting" early details. Your Prompt B needs to instruct the AI to keep core themes alive while adding new developments.
Submit Your Work
Your submission is graded against the rubric on the right. If you pass, you get a public Badge URL you can share on LinkedIn. There is no draft save, so work offline first and paste your finished response here.