Leveraging AI for Problem-Solving
Recently, I faced an intriguing challenge in a marketing campaign initiative. My campaign manager asked us to create 10 landing pages all the same but with a different video on each. I wanted to save my team time so I suggested we should just create a landing page where the embedded Vimeo video would change dynamically based on a query string parameter.
The challenge boiled down to this: I am busy and did not really want to write the code or search the internet to find solutions others have already tried. But how can I ensure that the right video plays when users visit the page with different query string parameters, all while keeping the page responsive?
This is where ChatGPT came to the rescue. I turned to it to help me generate the JavaScript code required for dynamic video switching. I implemented query string parsing to extract the video parameter, established a default video ID for cases with no parameter present, and integrated the Vimeo Player API to dynamically update the embedded video while preserving the page's responsiveness. The whole thing took about twenty minutes instead of the two hours I would have spent doing it manually.
The Bigger Lesson: AI as a Senior Colleague
That project taught me something bigger than query strings. It showed me that AI is most powerful not when you ask it to think for you, but when you use it to shortcut the tedious parts of execution. I did not need ChatGPT to tell me what to build. I already knew the solution. I needed it to handle the implementation grunt work so I could move on to higher-value decisions.
That distinction matters. The people who get the most out of AI treat it like a senior colleague you can interrupt at any hour, not a magic oracle that replaces your judgment.
How I Use AI for Progressively Harder Problems
Since that Vimeo project, I have leaned into AI for progressively more complex problem-solving. Debugging production issues at 2am by pasting stack traces and getting targeted hypotheses back. Drafting architecture proposals where I describe the constraints and let the model surface trade-offs I had not considered. Writing migration scripts for databases I have not touched in months.
In every case the pattern is the same: I bring the context and the judgment call, AI brings speed and breadth. Neither is useful without the other. The engineers I see struggling with AI are the ones who skip the context step and expect the tool to read their mind.
How to Start Using AI for Real Problem-Solving
If you want to start using AI for real problem-solving, here is what I would do. Pick a task you have been procrastinating on, something where you know the outcome but dread the manual work. Give the AI clear constraints: what you are building, what stack you are using, what edge cases matter. Review the output like you would review a pull request from a junior developer.
That mental model keeps you in the driver seat while still capturing the speed advantage. The goal is not to become dependent on AI. The goal is to free up your brain for the decisions that actually require a human.
Related: From AI Assistant to 10x Engineer and AI: Your Job's Future Isn't What You Think.
