In the ‘90s, we were told we could do anything if we just set our minds to it. Want to be an astronaut? Go for it! The world was our oyster.
Then reality hit: astronaut gigs are few and far between, mostly reserved for rocket scientists, military pilots… or celebrities.
But for Generation Alpha? They might actually be able to build anything—at least, anything that fits on a screen.
I have a long train commute, which is how this newsletter got started. When I’m not writing, I spend that time tinkering—building apps, testing tools, exploring what AI can do for solo developers like me.
Originally, I was coding most things by hand and using ChatGPT to fill in the gaps. But in just the past year, AI-assisted software building has gone from “helpful co-pilot” to “just hand it the wheel.”
Last year I wrote about how I was struggling to get my 4-year-old son to follow the rules of board games.
commented on that post suggesting an AI tool that helps explain to kids how to play.I gave it a shot in Replit, and each attempt—spaced out over a few months—got better. Part of that was me improving my prompts. But the bigger shift? The AI itself got smarter.
So one night in March, while I waited for my kids to fall asleep, I grabbed my phone, opened the Replit app, and gave it one more try.
A few prompts later… I had a working checkers game with real-time move suggestions.
Introducing: Checkers Master
An AI-powered checkers trainer you can play in your browser.
Built on a combination of the MiniMax algorithm and an LLM, Checkers Master gives:
Tactical move suggestions at every turn
Strategic tips to help you control the board
Performance tracking over time
It’s not perfect—but considering most of it was built quickly on my phone, it’s kind of amazing.
Just a few years ago, implementing the MiniMax algorithm which selects the computer’s moves took me several hours. Now? I’m getting playable prototypes while sitting in the dark waiting for a toddler to fall asleep.
If that’s not progress, I don’t know what is.
Dad vibe coding!!