Glue vs ChatGPT for Product Managers

ChatGPT is a general-purpose AI assistant. Glue is specialized in reading codebases. Both can help product managers, but in different ways.

What ChatGPT Does Well

ChatGPT can:

• Draft specs and documents
• Brainstorm features and ideas
• Explain technical concepts
• Summarize text and feedback
• Answer general questions

Best for: Writing, thinking, and explaining.

What Glue Does

Glue can:

• Answer specific questions about your codebase
• Estimate impact of feature changes
• Identify technical risks
• Understand architecture
• Generate documentation based on actual code

Best for: Understanding your specific system and making better decisions about it.

The Key Difference

ChatGPT knows about programming in general, but doesn't know your codebase.

Glue knows nothing about programming in general, but knows your codebase specifically.

Ask ChatGPT "how long will this feature take?" → It guesses, often wrong. Ask Glue "how long will this feature take?" → It analyzes your code and gives an informed answer.

Example

Product Manager: "Should we add dark mode?"

ChatGPT: "Dark mode is popular. High adoption in comparable apps. Easy to implement. Recommended." (Guessing based on general knowledge)

Glue: "Dark mode would require: theme context in React (new), updating 47 components to use theme values (medium), database migration for user preference (easy). Estimate 2-3 weeks, high confidence." (Based on your actual codebase)

When You Need Both

ChatGPT for: Writing, brainstorming, explaining Glue for: Estimating, understanding, deciding

A workflow:

1. ChatGPT: Draft a spec for dark mode
2. Glue: Analyze implementation cost for your specific system
3. You: Decide if the cost/benefit makes sense