Question: A company is building a vision system to inspect warehouse bins, but it cannot collect photos that match the lighting and camera angle it expects in production. It considers starting with product-catalog photos taken in a studio. Explain the tradeoff involved in that choice.

Sample answer: Using catalog photos can be a practical way to begin, because the team can train and test an initial model instead of waiting for perfect data that may never arrive. The downside is that those photos may differ in important ways from the real production images, so a model trained on them may not work well once it is deployed.

Key points:

• Proxy data can help the team make progress when production-like data is unavailable
• Starting with a substitute dataset avoids delaying the project
• The main cost is a higher chance of mismatch between training data and real use conditions
• The central tradeoff is faster development versus weaker generalization

Rubric: Full credit: response explains that proxy photos are useful as a fallback, states that they let the team start working, and clearly identifies the risk that the model may not generalize to production conditions. Partial credit: response covers only one side of the tradeoff. No credit: response omits the mismatch risk or implies the substitute data is equally representative.