What if you take an idea on a napkin and have a practical app for weeks instead of a long-standing coding class?
That's what LinkedIn's Daniel Roth did with Audio2, a podcast clipping app. His journey provides a blueprint on how AI coding assistants can reconstruct the meaning of “learning coding.”
Ross wasn't aiming to become a professional developer. Instead, he treated the AI as his partner. Using Claude Pro and Cursor, he was able to explain what he wanted, repetitively and compare different solutions quickly.
Pitting Claude against Cursor became a repetitive tactic, with each tool catching what the other tools missed. For design help, Expo handled the deployment while he leaned over Google's Gemini. He spent $807 creating this app.
Not everything went smoothly. Early attempts to build the entire functionality in AI at once led to dead ends. Winning strategy? “Think like a snow fort”: small, test its strength and keep stacking.
Ross also learned to find the “false confidence” of AI. A quick check on Reddit made it clear that it was deprecated when Claude assured him that FFMPegkit was the easiest way to generate video clips. Pivoting to the screen record saved the project.
Github has become his safety net. By actively diverging, he avoided losing several weeks of work into the AI-generated rabbit hole.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
And perhaps the most underrated tip: set strict session limits to resist the temptation of “another feature.”
AI-assisted coding is messy, but powerful. You don't have to be a programmer. You need to be a clear communicator, patient tester and willing to pivot. For builders with ideas, the barriers to creating software are falling apart.
Sign up for BI's Tech Memo Newsletter here. Please contact me by email abarr@businessinsider.com.
