An app in an hour

There's a lot brewing right now about AI taking over the world.

Maybe it will, maybe it won't.

But stressing out over the future doesn't help. Yes, I want to keep learning all I can and apply the learnings practically; but anxiety isn't a winning formula.

We can just take one step at a time.

So my practical step is to just go build something.

If you're curious about this whole idea of building apps with AI, the following steps are roughly how I approach pushing an idea forward.

Note: Here I must insert a million caveats and disclaimers. You could do something dangerous, you could build something that leaks data, please proceed at your own caution.

Move fast

Alright with that out of the way, let's say you've got an idea, something that you want to ship and show to others.

With your first idea it's crucial to get something out immediately to users for very early feedback. See how people respond, what they think, whether they find any delight in the idea.

Most things that you try won't work. So just get in the reps shipping.

As I write out this post I'm going to walk through the idea I have, share my process, and then share a link at the end to try it.

That's obviously quite limiting to what can be done in such a short time.

That means it can't be an app that requires personal data, can't be an iOS app, and generally should focus on sharing information or offering entertainment.

Constraints are good.

My goal is to get something out before I lose steam.

You may think that your perfect idea takes time to bake, and requires iteration and thought. It should. It does.

But not your first idea. Your first one isn't precious, it's a messed up broken thing, and needs to be shared to the world before you get a moment to think it too perfect. You should throw it out while it's still terrifying and before it's ready.

Pick an idea and just start

So, onto the ideas.

There's a million apps I want to build. But since I'm doing a web based mobile app, and I want it done quickly, I can't afford to worry about anything complicated. That means no storing data between sessions, or managing anything even slightly sensitive.

I keep a list of ideas in a note doc. I just took a look at it and had to ignore just about everything. A few were interesting: flash cards, a color mixer for kids, trivia games, etc.

These ideas aren't great, and I could just as easily start asking ChatGPT to help me flush out a bunch of new ones.

But that's not the point.

The goal right now is to take a bad idea and implement it quickly. That's all that matters.

I've worked on a trivia app before, and it takes quite a while to source data correctly. So I know I don't want to go down that route. I don't want to have a thousand trivia questions that are likely to be hallucinated with the wrong answers.

I also don't want an app that requires custom illustrations or designs.

So taking all those ideas and basically throwing them in the trash, I kind of want to just do something silly.

That makes my mind go in the direction of decision trees. An app like, "Should I run today or not?" that always ends with yes, regardless of what you choose. Or, an app that helps you decide which shoe you should wear for which occasion—and why wool socks with Birkenstocks is probably the right answer for the majority of the time.

But none of those are quite striking my interest, and I can think of problems that will take too long.

I'm going to go with something simpler.

For the last few years I've maintained a few lists of books that I like to recommend.

I've written on multiple occasions about the joy of reading, and how a good book can encourage a habit that will last a lifetime—but that's not the point of this article, or necessarily what the idea has to solve.

What I want instead is to take the data I already have—a list of books I've compiled, and display them to the user based on a few things they choose from.

I want this to be for someone who hasn't read a book since college, and feels worn out by the whole idea. I want it to be a joy, and to have options from fiction and non-fiction alike.

So, a name. I can't spend more than five minutes on this. A few domain searches reveals that readmyfirstbook.com, choosemyfirstread.com, are available, and a bunch of other names I'd prefer are already taken.

I have a rough idea for what the name will be, so that's good enough to proceed.

The basic idea

Now that I have an idea, and maybe a name, I need to quickly decide what is worth shipping in this first version. I want the name of the site, a quick description, and a quiz that will reveal a book based on options that the user chooses.

Now, there's an unlimited number of books, and there's obvious financial incentive at stake in recommending books—maybe even a trillion dollar idea if I dare so say myself—but that's not my interest here. I want a simple tool that I can share with friends, where they can take a quiz, and get a book recommended to them.

And in this case, the trick is that the book has to be one I've read.

There's obvious ways here to bring in network effects, build a community, and have recommended lists from others—wow, sounds like another great billion dollar idea.

But again, that's all been done, and I'm not trying to replicate any of it.

I just want to share my personal taste in books, and replicate a conversation virtually that I've had many times in real life. When that happens, when I bump into someone looking for a book to read, I go through a quick progression of questions.

I want to know if they're interested in fiction or non-fiction, and then go through the sub-genres from there.

Flushing out the details

So practically speaking, as I start to think through the app I'll create, I want the user to choose what they're in the mood for, and ideally get 1-2 books to choose from. That's it. All pre-vetted by me.

This feels like a bad enough idea to get started, so here's the prompt that I'm throwing into Claude Code:

I have an idea for an app that I want to build. It's called woldbooks.com. I'm not particularly sold on that name, but it works for now. The idea is that Joshua (me) has a list of recommended books I like to share with friends in person, and this app can help replicate that experience. Typically this is for someone who is not a big reader, and is looking for that first high to hook them into stories again. This will be a web based app, and you can just whatever code base you need for me to drop this into Netlify or something similar at the end of the day and ship it. I'm going to buy the domain on Cloudflare. We won't need any photos or graphics, just text and icons (maybe) for now. Pull from https://joshuawold.com/so-you-want-to-read/ and https://joshuawold.com/top10/, remove any duplicates, and add in the descriptions for why once they choose an answer. On the site, then, the user should choose fiction or non-fiction, and then categorize the rest of the book options into sub-categories. But no matter what they choose there should be at least one option in each sub-category, based on my lists. I'll add more to it later, but I just want to get out an app immediately. For styling, I want a nice font that feels like something you'd read a book with, and I want you to just go with dark background and light text, large fonts, and simple buttons that are only text based. The reveal, the book you choose, should just show the name, and it can have a link to the book on Amazon (if that's easy to do in a single prompt here), and have a way to go back. I should be able to change my results easily, and though this is a simple filter results of a list, it shouldn't feel like that. It should feel like I'm getting a special choice based on what I selected—because that's actually true, every single one of these books are stories I've read and love to recommend. Note, please remove Ender's Game. Also pull from Dungeon Crawler Carl and Game of Thrones, based on snippets from https://joshuawold.com/lost-in-worlds-found-in-our-world/. I want, at the end of the day, for someone to come to this website, choose their top genre, tap a button or two, and get a good recommendation for their first read. That's it. Give me ideas for styling, but I just kind of want it to be all text based.

That's a lot. And it's a little over the top, but that's mostly how I write my first prompt for an idea these days.

I put that directly into Claude Code, and wait for my first response. While that loads, I go ahead and buy the domain. I prefer Cloudflare Domains now, because they're cheaper than Hover, which I'm slowly moving away from.

I'm tempted to try Cloudflare Workers, and figure that out. But my goal right now is speed, and throwing this up on Netlify is good enough for this first day; after all I want to test the idea right now, not build the perfect infrastructure.

Ship it

The first version out of the box?

Frankly, it's exactly what I imagined, and had a few things I hadn't considered. The Amazon links don't all work, and since I don't want to bother figure those out right now, I'm going to ask Claude Code to strip those out.

I need links, of course. But I'm assuming my friends can just search for the book based on the title alone.

I'm really happy with the idea otherwise. So, a little fiddling around with domains and project settings, and the app is live.

From start to end this took about an hour. Including the time I needed to write this article. There's so much more I could do with this idea, but I'd love to hear from you, get your initial reactions on it, and also please share your own ideas back.

Based on the original premise, this idea has been an absolute success, and I'm inspired enough to want to try again tomorrow.

This is an exciting new world, and building little ideas like this helps make it just a tad more delightful.

Oh yes, and the link? Check it out at Wold Books.