Don't give LLMs full control. Use deterministic code for core logic, validation, and enforcing rules. Delegate only tasks requiring flexibility or understanding of unstructured input to the LLM, treating it as a specialized component, not the entire system.

AI development tools can be "resistant," ignoring change requests. A powerful technique is to prompt the AI to consider multiple options and ask for your choice before building. This prevents it from making incorrect unilateral decisions, such as applying a navigation change to the entire site by mistake.

To maintain the integrity of your "second brain," prohibit the AI from writing directly into your vault. If an agent adds its own notes, its generated patterns can contaminate your own. Enforce a strict separation where you manually integrate AI output to keep the vault a true reflection of your thinking.

AI can generate comprehensive documentation and extensive test suites in an instant. This devalues them as signals of a project's maturity or quality. The new, more reliable indicator of quality is actual usage and battle-testing, as AI-generated code might be technically perfect but practically unproven.

Continuously updating an AI's safety rules based on failures seen in a test set is a dangerous practice. This process effectively turns the test set into a training set, creating a model that appears safe on that specific test but may not generalize, masking the true rate of failure.

As AI generates more code than humans can review, the validation bottleneck emerges. The solution is providing agents with dedicated, sandboxed environments to run tests and verify functionality before a human sees the code, shifting review from process to outcome.

When an AI coding assistant goes off track, it can be hard to undo the damage. Developer Terry Lynn mitigates this risk by programming his AI workflow to make a Git commit before and after each small phase of a task. This creates a trail of "breadcrumbs," allowing him to easily revert to a stable state if the AI makes a mistake.

After an initial analysis, use a "stress-testing" prompt that forces the LLM to verify its own findings, check for contradictions, and correct its mistakes. This verification step is crucial for building confidence in the AI's output and creating bulletproof insights.

When an AI coding assistant asks you to perform a manual task like checking its output, don't just comply. Instead, teach it the commands and tools (like Playwright or linters) to perform those checks itself. This creates more robust, self-correcting automation loops and increases the agent's autonomy.