The Impulse to Begin — "What If I Built a Tarot Web App?"

 

Most side projects die in the planning phase. You know the pattern. A spark of inspiration hits, you open your notes app, start outlining features, burn through all your energy organizing ideas, and the whole thing quietly drifts into a "someday" folder.

This is the story of how I broke that pattern. I threw a single sentence at an AI, and within 10 minutes, the project had taken shape.

A Project Born from One Sentence

"I want to build a web page that does tarot readings."

That was it. No elaborate spec document. No PRD. No competitive analysis. I had a casual interest in tarot, and the thought of building it as a web app seemed fun. I had no feature list, no tech stack in mind, nothing.

Normally, this is where one of two things happens. Either you start googling "how to build a tarot app" and fall down a research rabbit hole, or you open Notion and start writing a table of contents for a planning document. Both approaches are fine, but they share a common danger: you run out of steam before you ever write a line of code.

This time, I tried something different. I asked Claude directly. That single decision shaped the entire trajectory of the project.

Why Projects Die in the Planning Phase

Let me take a quick detour and think about why side projects so often fizzle out before they start. In my experience, there are three reasons.

First, the perfectionism trap. The moment you think "I need to plan this properly before I start," you begin writing outlines, researching competitors, and listing features. Before long, the scope balloons, and you arrive at the inevitable conclusion: "This is way more than a weekend project."

Second, the domain knowledge wall. To build a tarot web app, you need to know things about tarot. How many cards are there? What types exist? How does a reading work? Once you start researching, tabs pile up. Information is everywhere, but none of it is organized to the level your project actually needs.

Third, decision paralysis. What tech stack should you use? Do you need a database? Where will you deploy? When choices pile up simultaneously, you freeze. You cannot figure out which decision to make first, so you make none.

I learned through this project that AI can lower all three of these barriers at once.

What Happens When You Throw an Idea at AI

"I want to build a web page that does tarot readings."

When I sent this sentence to Claude, I expected a vague follow-up like "So what exactly do you want to build?" The response was far more structured. Claude immediately began breaking this fuzzy idea down along several axes.

Which tarot deck would it be based on? What reading layouts would it support? How would card interpretations be delivered? Where would the card images come from?

This is the first advantage of an AI conversation. It decomposes a vague idea into concrete decision points. When you plan solo, the answer to "what am I building?" tends to exist as one amorphous blob in your head. When an AI asks explicit questions, that ambiguity transforms into specific choices.

The key point is that the AI was not making decisions for me. It was asking questions. I made the decisions. But having someone organize "what needs to be decided" completely changes the speed of decision-making.

The Courage to Start Without a Perfect Plan

Honestly, throwing a single sentence at an AI was not some strategic move. I did it because it was easy. I did not feel like writing a planning document. I just wanted to talk through my ideas.

But in hindsight, this approach turned out to be a perfect methodology for side projects. A side project is not a company project. You do not need a spec document for stakeholders. You do not need timeline estimates. All you need is a minimal agreement about "what to build" — and when the other party in that agreement is yourself, a single conversation is enough.

The AI conversation was a tool that rapidly brokered that agreement with myself. It pulled vague, floating ideas out of my head and made them concrete through a series of questions. It felt like brainstorming with a colleague over coffee — "Wouldn't it be cool to build something like this?" — except this colleague happened to know about tarot, web development, and design all at once.

Capturing the Energy of an Idea

The most important resource in a side project is not time. It is motivation — energy. That burst of excitement when you think, "This would be so fun to build." That energy fades quickly.

The traditional process of writing a spec, benchmarking competitors, and evaluating tech stacks slowly chips away at that energy. If you reopen "that tarot project" two days later, your initial excitement is already halved. A week later, you start wondering whether the whole thing was too ambitious.

The way the AI conversation solved this was simple. It captured the energy of the moment the idea struck and immediately channeled it into something concrete. Ten minutes after I typed that one sentence, I already had a clear spec: 78 tarot cards, one-card / three-card / Celtic Cross reading modes, pre-written interpretation text, React + Vite. At that point, the project was already alive.

Through this process, I realized something. What truly matters in a side project is not "how well you plan" but "how quickly you get your hands dirty." You do not need a perfect plan. You just need a direction. Everything else can be adjusted as you build.

Naming the Project

The project name emerged naturally during the conversation. I asked, "How about just going with Tarot Master?" and the feedback was that it was clean and intuitive.

It might seem trivial, but naming a side project matters more than you think. The moment it has a name, it gains a sense of reality. It goes from "that project" to "Tarot Master." Once the folder is named and you start referring to it by a proper noun, the project becomes real.

Here, too, the advantage of AI collaboration showed itself. Instead of agonizing over the perfect name for hours, I made a quick decision and moved on. If the name bothers me later, I can always change it. In a side project, the name is a starting gun, not a final verdict.

Starting Right Now

Looking back, the most important moment of this project was not when I wrote the first line of code or when I completed the deployment. It was the moment I typed "I want to build a web page that does tarot readings" to Claude. That was when the project began, and once the wheel started turning, it never stopped.

You probably have an idea floating around in your head too. "It would be so fun to build something like this." Do not transfer that idea into a planning document. Throw it at an AI in a single sentence. See for yourself what happens in 10 minutes.

What I Learned

First, when you ask an AI an open-ended question, you get surprisingly good structuring in return. The vague input "I want to build a tarot web page" came back decomposed into decision axes: reading layouts, interpretation engine, tech stack.

Second, the most dangerous thing in a side project is "getting exhausted before you start." A fast start beats a perfect plan, and AI is the tool that creates that speed.

Third, the energy behind an idea has an expiration date. If you do not capture it the moment it appears, it vanishes. An AI conversation is the fastest tool for instantly converting that energy into something concrete.

Next Up

The decision to start has been made. But I know almost nothing about tarot. I did not even know how many cards were in a deck, and terms like Major Arcana and Minor Arcana were foreign to me. In Part 2, I will share how I used AI to dive into an unfamiliar domain, and how the learning efficiency of an AI conversation compares to traditional googling.