Developer Alex O'Callaghan wanted to help his son practice phonics for an upcoming test, but faced a familiar parenting struggle: the kid wasn't exactly thrilled about drill work. The twist? His son happens to be obsessed with Minecraft at what O'Callaghan describes as "a professional level." So he did what any engineer parent would do โ€” he opened Claude and described what he wanted.

The Build Process

In a single conversation with Claude, O'Callaghan specified the requirements: a browser-based phonics game for Key Stage 1 (UK curriculum), Minecraft-themed visuals, two modes (picture-to-spelling using letter tiles, word-to-picture multiple choice). What came back was a fully functional React component styled to look like Minecraft. No iteration needed โ€” it worked out of the box.

Taking It Further

Being an engineer, O'Callaghan couldn't leave well enough alone. He pulled the code into Claude Code, set up Vite and TypeScript, added Vitest for testing, and found an npm package with actual Minecraft item sprites to replace the emoji placeholders he initially used. "The real assets helps make the icons much more recognisable, some of the emoji choices were a bit questionable," he admits. He also added difficulty scaling so word pools adjust as his son improves.

The Unexpected Co-Design

When O'Callaghan showed his son the game after school, the reaction was immediate enthusiasm โ€” no prodding required. But what surprised him more was the co-design feedback that followed. His son immediately suggested including netherite "since it makes the strongest armour and would be a tricky word." The child was invested in the project, practicing phonics alongside something he actually wanted to play. There were unintended benefits too: navigating the game provided fine motor practice with a mouse that his son hadn't gotten from touchscreen apps.

What This Means for EdTech

Commercial educational software targets the average student because that's where the economics work. But O'Callaghan argues this creates a gap: "My son is not the average child. He is a specific child who is obsessed with a specific game, sitting a specific test, at a specific moment in his learning." AI changes the equation by dropping the cost of building tailored solutions toward zero โ€” implications that extend beyond parents to tutors building exercises around student interests or teachers adapting materials for their specific classes.

Key Takeaways

  • A single Claude conversation produced a playable Minecraft phonics game with no engineering background required
  • The initial version was functional and engaging; iteration added polish but wasn't necessary for value delivery
  • Child co-design emerged naturally โ€” the son proposed adding netherite as challenging vocabulary
  • EdTech's economics have always favored average users; AI could enable true personalization at scale

The Bottom Line

This isn't just a cute parent hack โ€” it's a glimpse at what happens when building personalized tools becomes cheap enough that imagination is the only constraint. O'Callaghan put it plainly: "The constraint has always been time and technical access, not imagination." That's a big deal for education, where one size has never actually fit all.