Vibe-coding a kid-friendly AI fortune teller for your Halloween festivities | Marco Casalaina

How I AI How I AI Oct 31, 2025

Audio Brief

Show transcript
This episode explores how AI tools like GitHub Spark can rapidly create functional web applications for personal projects, exemplified by a Halloween fortune teller app. Key insights from this demonstration include how AI tools accelerate web app creation, the value of generative AI for creative personal projects, and the importance of iterative prompting for refining AI output. Using GitHub Spark, a host quickly coded a “Mystic Oracle” app to generate kid-friendly fortunes for a community Halloween event. This “vibe coding” approach demonstrates rapid generation of full-stack intelligent apps directly from natural language prompts. The project highlights a practical and creative way to leverage AI for non-work purposes, creating interactive experiences. It shows AI's utility in making engaging content for family and community events. Initial AI-generated fortunes were overly poetic. Through clear, conversational follow-up prompts, such as “make it kid-friendly” and “more concrete,” the app's output was quickly improved, resulting in humorous fortunes suitable for children. This illustrates AI's significant potential for both quick development and creative application, making complex tasks accessible for anyone with an idea.

Episode Overview

  • This special, unplanned Halloween episode of "How I AI" showcases a fun, real-world application of AI for a community event.
  • Host Marco, dressed as Captain Picard, demonstrates how he uses AI to create a fortune-telling app for his role as the "block fortune teller" at his neighborhood's Halloween party.
  • Using GitHub Spark, Marco live-codes a "Mystic Oracle" web app that generates kid-friendly fortunes with the click of a button.
  • The hosts show how to iterate on the initial AI-generated app by providing follow-up prompts to refine the fortunes, making them more humorous, concrete, and suitable for children.

Key Concepts

  • Vibe Coding with AI: The central theme is using natural language prompts to rapidly generate functional code and user interfaces for simple applications, a practice often called "vibe coding."
  • GitHub Spark: The demonstration uses GitHub Spark, a tool that transforms text-based ideas into full-stack intelligent apps, handling everything from the product requirements document (PRD) to the HTML, CSS, and JavaScript.
  • Iterative Prompting for Refinement: The initial version of the app produced overly poetic fortunes. The process shows how providing follow-up instructions (e.g., "make it kid-friendly," "make it more concrete") allows for quick iteration and improvement of the AI's output.
  • Creative AI for Personal Projects: The episode highlights a practical and creative way to use generative AI for non-work purposes, creating a fun and interactive experience for a family and community event.

Quotes

  • At 00:00 - "Welcome to a spooky and unplanned Halloween edition of How I AI." - Claire Vo sets the stage for the impromptu, holiday-themed episode that resulted from technical difficulties with their planned recording.
  • At 01:26 - "By night, at least for Halloween, I do something a little bit different. I am the block fortune teller." - Marco explains his fun Halloween tradition, which provides the context for building a fortune-telling app with AI.
  • At 09:30 - "This week, you will find a shiny rock that looks like a slice of pizza and makes everyone giggle when you show it off!" - An example of a kid-friendly and concrete fortune generated by the app after Marco refined his prompt to make the results more humorous and less abstract.

Takeaways

  • AI tools like GitHub Spark can transform a simple idea into a functional, styled web application in just a few minutes, perfect for rapid prototyping or quick projects.
  • Use generative AI for fun and creative personal projects beyond typical work tasks, such as creating an interactive activity for a party or family event.
  • Iteration is crucial when working with AI code generators. The first output may not be perfect, but you can easily refine it with clear, conversational follow-up prompts.
  • Even with simple prompts, AI can generate impressive user interfaces and designs, making it possible to create visually appealing apps without deep design or front-end coding knowledge.