I built an entire AI-powered app live, in front of an audience, in just four hours. Did I finish it? Not quite. Did I learn a huge amount? Absolutely. Here is what happened, what I learned, and why I will do it again.
The challenge was simple: could I build and launch a working AI cheatsheet generator, live on stream, using AI first coding and Kaijo1 as my main tool?
Answer: almost! By the end of the session, the app could create editable AI cheatsheets, but it was not yet deployed. A few minutes of post-stream fixes later, it was live for everyone to try. (Next time, I will check deployment on every commit!)
Try the app here: aicheatsheetgenerator.com
What We Achieved in Four Hours
In just four hours, I went from a blank slate to a working product. The app generates AI-powered cheatsheets that you can edit and customise. The core features were working by the end of the stream, but deployment tripped me up. It only took a few more minutes to fix, but it was a reminder: always check your deployment pipeline as you go.
The real story is not just about the code. It is about how much faster and more reliably I could build with AI-first tools. Coding AI made it possible to move at a pace that would have been unthinkable a year ago. Kaijo, my reliability assistant, will make future improvements even smoother.
What I Will Improve Next
Shipping live is only the beginning. Here are the improvements I am planning:
Pay to download: I want to add a way for users to pay to download their cheatsheets, supporting the project and unlocking more features.
Better editing: I will improve the editing experience, including box ordering, styles, and colours, so users can make cheatsheets truly their own.
Automatic Kaijo evaluation: I plan to tell Kaijo about the edits and automatically evaluate them, making the app smarter over time and more tailored to the user.
Lessons Learned from Live Coding
AI-first coding is a force multiplier. I achieved more in four hours than I could have managed in a full day solo.
Building live is exhilarating and exhausting. The pressure to deliver in real time forces you to make decisions quickly and trust your tools. Here is what stood out:
AI-first coding is a force multiplier. I achieved more in four hours than I could have managed in a full day solo. The AI handled boilerplate, suggested fixes, and kept me moving forward.
Deployment is always the bottleneck. No matter how fast you build, if you do not check deployment regularly, you will get caught out. I will automate more checks next time.
Audience feedback is invaluable. Viewers spotted issues, suggested features, and kept the energy high. Live interaction made the process so much more fun and productive.
Watch the Full Recording
If you missed the stream, you can watch the full recording below:
Or watch it directly on YouTube.
What Should I Build Next?
Should I do another live build with a different service? What would you like to see me tackle next? Let me know in the comments or reach out on social media.