I Wanted AI to Build a Small Tool, Then Got Stuck Immediately
Last Tuesday night, I was at the desk in my Hangzhou rental apartment, trying to use AI to build a client quotation sheet. I wrote three rounds of prompts, and the AI said it needed to call an API. I didn’t even know what an API did. I’ve been stuck like this before: not because I didn’t want to learn, but because opening a course and seeing dozens of pages of lecture notes immediately discouraged me.
Later, I reorganized my learning path and realized those fundamentals hadn’t been wasted. I now know what a tool needs to run, where its data lives, and which layer to check first when something breaks. This isn’t a crash course in becoming a programmer. It’s the lowest-cost path for non-coder founders to gain enough understanding to follow and evaluate what’s happening.
Carnegie Mellon Put Many Computer Science Courses on YouTube
The free courses cover areas including computer science fundamentals, programming, and AI. They suit people who want to build their own workflows, understand some technical principles, or communicate more effectively with outsourced developers. Carnegie Mellon isn’t sending anyone into a computer lab; it’s making classroom recordings and supporting materials publicly available so we can learn on a more flexible schedule.
I know Ajie, a fellow designer who runs his own service business. He wasn’t planning to become a programmer; he just wanted to automate his repetitive quoting process. From a shared office in Shenzhen, he used YouTube to learn about variables, conditions, and error handling, watching for 20 minutes during his lunch break each day. A few weeks later, he could read the flowcharts developers gave him and catch risks earlier, like client information being written to a public spreadsheet.
Starting Today: $0 and About 7 Days
Money: $0. The public videos and course materials are free. Time: Start with 7 days of 30 minutes a day. Technical barrier: You need to know how to search on YouTube, pause videos, and take notes; you don’t need to write code. First step: Open YouTube, search for “Carnegie Mellon computer science course,” enter the CMU course page, and choose a course labeled “Introduction” or “Beginner.” Save the playlist, then create a note titled “Work Problems.”
I made this mistake too: I initially saved more than a dozen courses, assuming that learning more would make me better. I abandoned the plan after a week without finishing anything. Now I follow just one course and connect each lesson to a real problem, such as “How can I turn client forms into quotes automatically?” If a lesson isn’t useful yet, I skip it. This tool isn’t essential for everyone, either, so there’s no pressure to try it right now.
How I’d Arrange It by Stage
If I were just starting out, I’d choose an introductory course and study twice a week, aiming to understand the key terms in AI tutorials rather than build a product independently. If I already had 1–2 clients, I’d choose courses around automation and data processing while asking a collaborating developer to map the workflow, so we wouldn’t end up talking past each other. If I were scaling up, I’d turn the course material into internal training so operations, sales, and support could all understand the tool’s basic limits.
I’d start by saving one lesson, watch for 10 minutes tonight, and then decide whether to continue. Learning isn’t about proving that I understand technology. It’s about reducing the guesswork when I next pay for something, write a request, or make a decision.