Memorizing scripture, one blank at a time

A Goal I Kept Falling Off Of

It's been a goal of mine for a while to build a habit of memorizing Bible verses. Not just for the sake of doing it, but to remind myself of what Scripture says. I believe the Bible is the perfect and inerrant word of God, and it's the root of what I believe as truth. So for everything theological that I believe, there's a passage of Scripture I should be able to point to.

If I'm ever asked about what I believe, I want to recite the passage behind it—word for word, chapter and verse.

I started earlier this year by putting together 100 references I wanted to memorize, specifically related to core Christian doctrine. They covered a variety of themes, like the nature of God, the Trinity, and salvation by grace through faith. I started with the first verse, Genesis 1:1, then the second, Exodus 3:14–15—and then I fell off.

Why Memorization Is Hard

Scripture memorization is hard, especially when you're constantly memorizing new verses. And there are a lot of pain points that make it even more difficult than it needs to be. It takes time to find the verse in the Bible several times a day, and even more if you're working on several at once—constantly flipping between pages and screens. Flashcards help some, but it isn't convenient to carry them, store them, and keep them organized. A flashcard app is even worse; just typing in each verse is tedious and doesn't feel productive.

An Idea at the Coffee Shop

Then one morning before work, at the coffee shop, I had an idea: an app that would hold my hand and guide me through each and every verse.

The app would present one verse at a time, with a couple of words swapped for blanks. A word bank below would let me tap the right word for each blank. Over time, the number of blanks would grow until the entire verse was blank. I'd only focus on three verses at a time, and every day I'd still review the verses I had already memorized.

As a verse climbs from Easy to Medium to Hard, fewer words stay on the page to lean on. Once it graduates, nothing is left: every review starts from a completely blank verse.

A Prototype in a Day

As soon as I had the idea, I opened a new Claude chat and started free-writing. I went back and forth with the model to organize my thoughts into a cohesive spec for a prototype, then had it generate a plan for a PWA and a backend. From there, I started a Claude Code session with Fable 5 and had it build both.

That same day, I had a working prototype.

It was rough around the edges as far as UX goes, but it worked, and I could see the potential as I poked around. There were plenty of usability issues:

1. You had to constantly scroll up and down to see both the verse and the word bank.
2. There was no way to pause a session midway through—you had to finish or start over.
3. It wasn't clear to a new user how the progression system worked.
4. There was only one translation, the World English Bible.
5. Features meant for quality of life ended up being confusing, like showing the first letter of each word in its blank.

But the app worked surprisingly well for something developed so rapidly.

Polished Through Daily Use

Over the next couple of weeks, I spent time polishing. The app was installed on my iPhone, I was using it heavily every day, and I kept noticing things to fix. Each day I'd make those tweaks, redeploy, and see what I thought.

A small group of people used it alongside me. For the most part, their feedback matched mine, but they had several ideas I'd never thought of—like a translation preference, or the ability to rearrange the verse queue.

At a certain point, the changes I was making got smaller and smaller, and it became clear the project was maturing. The app was done, and anything further would be an update. Today, Genesis 1:1—the first verse I tried and failed to keep going past—sits in my library as memorized, coming back for review every two weeks.

How AI Changed the Way I Build

Between this project and my last one, AI agents have gotten really, really good. They were capable enough that I could completely change my development flow to take full advantage of them, and the result was a product made faster, with fewer issues, and with better UX by several factors.

Without AI, I'd have to be far more intentional about where my effort went. Prototyping an idea was a huge investment, and if it turned out sub-par or a dead end, that time was largely wasted. With AI tools, an idea can be prototyped in minutes. Several prototypes can be made. Iteration happens in minutes, not weeks—and this app took several iterations before it became effective.

Prototypes are also where bugs love to creep in, but frontier models rarely generate buggy code these days. When they do, the agents themselves have the resources to take a bug report, debug it, test it, and fix it.

Claude Design

Claude Design may have been the biggest game changer. Design is tedious by hand—not just developing layout, navigation, and hierarchy, but basic groundwork like building a design system. Claude Design let me visualize several ideas at once instead of spending a full weekend developing one concept. It fixed UX issues in minutes and updated mockups across devices.

It let me focus on what the product is and how people use it, rather than the minute details that matter less in a first version.

What Is a Developer For?

This raises the question: what is a software developer good for if AI can do all the work? For me, it frees me to do what I enjoy most—creating something useful as best I can.

AI doesn't know what to create or who the customers are—that's up to me. It doesn't know the future of the product or what we need to do to prepare for it. It's up to me to make sure we build with that future in mind, so we don't engineer ourselves into a corner.

Developers will need a broad understanding of the whole stack in the coming years. Minute code details will matter less. Readability will always be important, but soon developers will hardly be modifying code by hand.

Built With