Getting the Most from Vibe Coding

Regular refactoring with AI helps keep code clean and manageable.

"Make sure to refactor frequently... You can even ask the LLM to identify parts of your code base that seem repetitive or might be good candidates for refactoring."

• Use AI to identify and refactor repetitive code sections.
• Keep code modular and small for better understandability.
• Maintain high code quality through continuous refactoring.

Voice input can accelerate interaction with AI coding tools.

"I use Aqua, a YC company, and basically I can just talk at my computer and Aqua transcribes whatever I'm saying into the tool I'm using... I actually wrote this entire talk with Aqua."

• Utilize voice transcription tools for faster input.
• Increase productivity by speaking instructions to AI.
• Overcome typing speed limitations with voice.

Different AI models excel at various tasks, impacting product development.

"I try every new model release to see which performs better... Some are better at debugging or long-term planning or implementing features or refactoring."

• Test new AI models regularly for different development tasks.
• Switch models if you encounter limitations or issues.
• Stay updated with AI advancements to optimize productivity.

Selecting a tech stack familiar to AI models can improve coding outcomes.

"I chose to build my project partially in Ruby on Rails... There's a ton of pretty consistent high-quality training data for Rails codebases online."

• Use widely-adopted frameworks to leverage AI's strengths.
• Recognize that AI performs better with languages it knows well.
• Anticipate performance variations with less-common languages.

Writing high-level tests helps maintain code quality when working with AI.

"Make sure you write high level integration tests before you move on to the next feature... So you can git reset and start again."

• Write integration tests to catch regressions from AI's code changes.
• Use tests to identify unnecessary changes made by AI.
• Maintain code stability by testing before advancing.

Using version control ensures stability when AI tools make unexpected changes.

"Version control is your friend. Use Git religiously... So I always make sure I'm starting with a kind of a clean git slate before I start a new feature."

• Commit changes frequently to revert if AI outputs unwanted code.
• Reset to known states before attempting new features or fixes.
• Manage AI-generated code effectively with robust version control.

Viewing AI as programming with language changes how you build products.

"My advice would be to think of the AI as a different kind of programming language... Instead of programming with code, you're programming with language."

• Provide detailed context and information to the AI.
• Learn to "program" by crafting precise language prompts.
• Enhance results by treating AI interactions as coding in language.

AI can quickly diagnose and fix coding errors, streamlining product development.

"The first thing I do when I encounter any bug is just copy paste the error message straight back into the LLM... Often this error message is enough for the AI to identify and fix a problem."

• Feed error messages directly to AI to diagnose issues.
• Let AI suggest fixes based on the errors without additional input.
• Streamline debugging by leveraging AI's problem-solving abilities.

AI can accelerate not just coding but also non-coding tasks in product development.

"I use them for a lot of non-coding work when I'm building these kind of side projects. For example, I had Claude Sonic 3.7 configure my DNS servers... It was a DevOps engineer for me and accelerated my progress at 10X."

• Utilize AI tools for tasks like DevOps configurations.
• Employ AI for design tasks, such as creating and resizing images.
• Expand AI use beyond coding to speed up overall development.

Planning with AI helps you build products efficiently by leveraging AI's ability to structure and guide your development process.

"I would work with the LLM to write a comprehensive plan. Put that in a markdown file inside your project folder."

• Collaborate with AI to create a detailed project plan.
• Implement step by step, testing and committing as you go.
• Avoid one-shot attempts; build incrementally.