To ensure high code quality, Gabor created a specialized 'code maintainability agent.' This AI's sole job is to check for circular references, enforce naming conventions, and ensure high-quality comments—technical details a product manager might overlook but are critical for long-term project health.

An internal OpenAI team maintains a codebase written entirely by AI. By removing the "escape hatch" of manual coding, they are forced to solve fundamental problems in providing better context and documentation to the AI, thus uncovering best practices for agent interaction.

Atlassian found users struggled with prompting, using vague language like 'change logo to JIRA' which caused the AI to pull old assets. They embedded pre-written, copyable commands into their prototyping templates. This guides users to interact with the underlying code correctly, reducing hallucinations and boosting confidence.

Instead of codebases becoming harder to manage over time, use an AI agent to create a "compounding engineering" system. Codify learnings from each feature build—successful plans, bug fixes, tests—back into the agent's prompts and tools, making future development faster and easier.

Even for a simple personal project, starting with a Product Requirements Document (PRD) dramatically improves the output from AI code generation tools. Taking a few minutes to outline goals and features provides the necessary context for the AI to produce more accurate and relevant code, saving time on rework.

When an AI agent was given one large prompt to create a design, it ignored parts of the style guide. Gabor theorizes this is due to 'context compression' where details are lost in a large prompt. The solution is to break tasks into smaller, ticketed items, mirroring human workflows to ensure fidelity.

Instead of fighting for perfect code upfront, accept that AI assistants can generate verbose code. Build a dedicated "refactoring" phase into your process, using AI with specific rules to clean up and restructure the initial output. This allows you to actively manage technical debt created by AI-powered speed.

When an AI-generated app becomes hard to maintain ("vibe coding debt"), the answer isn't manual fixes, but using the AI again. Users should explain the maintenance problems to the tool and prompt it to rethink the solution from a deeper level, effectively using AI to solve AI-created tech debt.

Use AI to manage its own development tasks. After a brain dump of project goals, have the AI create tickets in a tool like Linear. Then, let the AI work through the tickets and update its own statuses, significantly reducing your mental load and freeing you up for higher-level review.