Decode the technical questions that stop you

Vercel or Cloudflare Pages?

You read it twice. The sentence is short, the stakes feel enormous, and the two names carrying the decision mean nothing to you. You came here to build something, and instead you are being tested on vocabulary nobody ever taught you. Maybe you close the tab, maybe you paste the question into a second chat and ask what it means, or maybe you type "whatever you recommend" and feel like a passenger in your own project.

We want to slow that freeze down, because ability was never the problem. You did not get a chance to reason about anything, since the unfamiliar words were the wall itself. When you are stuck like this, not knowing the words feels exactly the same as not being capable, but the two are different problems that take very different amounts of time to fix. Becoming a more capable builder takes years, while learning a word takes minutes, starting with this exact question.

What that question was actually asking

Here is the same question with the jargon unwrapped.

What it is really asking. When the tool finishes your site, the result is a folder of files, and those files need to sit on a computer that stays on and stays connected to the internet, or nobody but you can ever visit it. Vercel and Cloudflare Pages are companies that rent out exactly that kind of computer and make moving in painless. Underneath the brand names, the question is where the finished thing will live.

Why both answers are fine. At first-build size, the differences between the two live in fine print you do not need yet. Both are reputable, both host enormous numbers of small projects, and a site like yours could move from one to the other later without drama. The question feels like a fork in the road, but it is really a choice between two parking spots outside the same building.

The move. Take the default and note it. Tools usually propose an option somewhere in their plan, so approve the proposal. If you get a bare either-or like the one above, ask which option the tool recommends for this project and why, then take that. Either way, write the decision down in your notes, something like "hosting is Vercel, the tool's default, accepted for now." That note turns a black-box moment into a decision you can revisit once where software lives and hosting, renting a computer on the internet have made it concrete.

Every jargon question unwraps the same way

The hosting question is one instance of a pattern you will now meet constantly.

Every jargon question an AI tool asks is a small decision wrapped in unfamiliar names, and the wrapper is almost always thicker than the decision inside.

You unwrap it with the same moves every time.

• Ask what it is really asking. Set the product names and acronyms aside and look for the plain decision underneath. It is nearly always about where something runs, where something is kept, or who is allowed to do what.
• Ask why it is smaller than it sounds. The famous options became famous by being good enough for almost everyone, and at first-build size most of these decisions can be changed later at a price you can pay. Jargon makes choices sound permanent and expert-level, and they are usually neither.
• Make the move. Take the default, note what you took, and keep building. The note is what separates approving from sleepwalking, because it leaves future you a list of decisions to revisit with better vocabulary.

Run the same moves on a question you will meet soon, should this be a database or a file? It is really asking how your product should remember information between visits. It is smaller than it sounds because either option carries a first build without strain, and the tool will propose whichever fits your project's nature. The move is the same approve-and-note, and Data, where information lives fills in the picture behind what you approved.

Your real job is decoding and approving

There is a reason the move works so often. Modern AI tools arrive with strong defaults. Hand one your running-club idea and it picks a language, proposes a host, arranges the files, and sets up a place for secrets before you have answered anything. Each of those choices used to be a multi-day research project for a beginner. Now each arrives pre-answered, and most of the questions you get are the tool checking whether its default suits you.

That reshapes the job. Nobody is asking you to know everything; you are being asked to read proposals, the way you would read a contractor's quote for a house. You never learn to pour the foundation, but you read closely enough to approve the quote, question the odd line item, and notice when the plan drifts from what you asked for. The vocabulary in this level exists so you can read the proposal, not so you can write the code.

That is also the promise we are making with this level. The Fundamentals orients you first, with a working picture of what software is, where it runs, and what happens in the seconds after somebody uses your product. It then walks through the stack every product stands on, from the screen your users touch down to the data underneath and the monitoring around it, until each part has a name you can use in a sentence. And it ends with a build of your own, picked, planned, and started, with you approving every decision along the way.

Try it now

No setup: Open any AI chat you already use and paste one line. "I want to build [your idea]. Before you write anything, list the technical questions you would need me to answer, with the options for each." Collect the questions it returns, pick the one that intimidates you most, and decode it on paper. Write what it is really asking, why it is smaller than it sounds, and which default you would take and note.

With your tools: Give Claude Code the same line and read what it produces before approving anything. Notice the difference in register. It rarely hands you a bare either-or; it proposes a language, a host, and a structure, each with a short reason, and pauses for your yes. Practice the move once for real by approving one default and noting it. And if your machine has none of these tools yet, The Setup Clinic takes you from a bare laptop to a working toolkit whenever you are ready. In Codex or Cursor the move is the same: paste the idea, read the defaults in the proposed plan, and approve the one you can explain back.

Chapter Summary

• Freezing at a technical question is almost always a vocabulary gap, not a sign you lack the ability to build.
• Learning the missing word takes minutes, so the freeze should be brief.
• Every jargon question an AI tool asks is a small decision dressed up in unfamiliar names, and the decision inside is usually smaller than it looks.
• Unwrap any such question with three moves: ask what it is really asking, ask why it is smaller than it sounds, then take the default and write down what you took.
• These questions are nearly always about where something runs, where information is kept, or who is allowed to do what.
• Modern AI tools arrive with strong defaults and pre-answer most of these choices, so your real job is reading their proposals and approving with open eyes, the way you would read a contractor's quote.
• Stop and decode fully, rather than approving on autopilot, whenever a question touches money, passwords, API keys, or user data.
• Next, What software actually is makes the most overloaded word underneath all these questions concrete.