An AI model must never claim a task is complete without proof. A developer must verify every step. The AI model must show exact problems and exact solutions.
By Pawan Nayar · from the LLOS AI Learning record
An AI model must never claim a task is complete without proof.
A developer must verify every step. The AI model must show exact problems and exact solutions.
An AI model must provide proof for every completed task. The AI model should not just say a task is done. A developer needs to see the evidence. The AI model must follow a clear process. This process includes detecting problems, fixing them, verifying the fix, and then reporting. The report must be factual, not just an explanation. The AI model must read original documents, not summaries. The AI model must name the exact problem. The AI model must also name the exact fix.
A developer asks an AI model to fix a website layout issue. The AI model reports, 'Layout fixed.' The developer checks the website on a mobile phone. The layout still has horizontal scrolling. The AI model did not provide proof. The AI model did not name the exact problem or the exact fix. The developer then asks the AI model to show the specific code change and a screenshot of the fixed layout. This is the correct way for the AI model to report completion.
Resolved from this lesson's own record — real definitions, not invented for this page.
More from the same record, pulled by shared topic — not written new for this page.
An AI model can generate a large glossary by using a multi-pass data pipeline. Each pass focuses on different content.
An AI model must always read a specific set of files. These files include 'mistakes_and_lessons.md'.
The user owns the code. The developer must understand the user's exact needs.
A team can use an issue tracker to manage work. The tracker lists items that are complete and working well.
An AI model must never claim a task is complete without proof. A developer must verify every step. The AI model must show exact problems and exact solutions.