My LLM Codegen Workflow

tl;dr; Brainstorm spec, then plan a plan, then execute using LLM codegen. Discrete loops. Then magic. ✩₊˚.⋆☾⋆⁺₊✧

The Process

Here's my current workflow for using LLMs in software development. It's built upon personal experience and conversations with fellow developers. Note that in the fast-moving world of AI, what works today might need adjustment tomorrow.

Step 1: Idea Refinement

Start with a conversational LLM (like ChatGPT-4) and use this prompt:

Ask me one question at a time so we can develop a thorough, step-by-step spec for this idea. Each question should build on my previous answers, and our end goal is to have a detailed specification I can hand off to a developer. Let's do this iteratively and dig into every relevant detail. Remember, only one question at a time.

Here's the idea: <Your Idea>

After the brainstorming concludes, ask for a comprehensive spec:

Now that we've wrapped up the brainstorming process, can you compile our findings into a comprehensive, developer-ready specification? Include all relevant requirements, architecture choices, data handling details, error handling strategies, and a testing plan.

Save this as spec.md in your repo. This spec can be used for multiple purposes:

Code generation
Idea validation
White paper generation
Business model creation
Deep research foundation

Step 2: Planning

Take your spec to a reasoning model and use one of these prompts based on your needs:

For Test-Driven Development:

Draft a detailed, step-by-step blueprint for building this project. Break it down into small, iterative chunks that build on each other. Make sure the steps are small enough to be implemented safely with strong testing, but big enough to move the project forward.

The output should provide a series of prompts for a code-generation LLM that will implement each step in a test-driven manner. Prioritize best practices, incremental progress, and early testing.

Make sure each prompt builds on previous ones and ends with proper integration. No orphaned code!

For Regular Development:

Draft a detailed, step-by-step blueprint for building this project. Break it down into small, iterative chunks that build on each other. Make sure the steps are small enough to be implemented safely but big enough to move the project forward.

The output should provide a series of prompts for a code-generation LLM that will implement each step. Prioritize best practices and incremental progress.

Make sure each prompt builds on previous ones and ends with proper integration. No orphaned code!

Save the resulting plan as prompt_plan.md and generate a todo.md checklist to track progress.

Time Investment

This entire planning process typically takes about 15 minutes. The investment in proper planning pays off significantly during the implementation phase, leading to more structured and maintainable code.

Why This Works

This approach provides several benefits:

Clear documentation from the start
Well-structured development plan
Manageable, testable steps
Consistent progress tracking
Reduced complexity in each development phase

Remember: The key is in the iterative nature of the process. Each step builds on the previous one, ensuring that you're always moving forward with a clear understanding of what needs to be done next.