Products BeaconMCPs
Company About Blog
Appearance

WYPIWYG: the FrontPage of our times

In middle school, Microsoft FrontPage opened the door to my career by pointing me at the language underneath. AI is doing the same now. Vibe code without guilt, but read what it writes.

I was in middle school the first time I made a website. It was for a computer science class, in Microsoft FrontPage. You dragged elements around, picked styles from a sidebar, watched the page render on the canvas, and at the end you had a website. I loved it. FrontPage was a good first instrument.

I started googling things I wanted to do (more animated gifs, mostly; this was the era) and figured out that FrontPage was writing code. The code was bad: bloated, malformed, table-soup HTML that wouldn't render the same in two browsers. But it was code. A language I could read. So I started reading.

Almost 30 years later, I'm still working in a design career that began with a website built by Microsoft FrontPage. The "bad code" I'd noticed was HTML, and HTML was a language I could learn, and learning it compounded into a hobby, then a career, then a worldview. Dreamweaver was "better" by degrees, but it suffered the same ailment and pointed somewhere similar. I moved to a plain text editor because the people on the forums told me that's what you did once you understood what was actually happening on the page.

Two arguments collided then. WYSIWYG let regular people make websites, and democratizing web presence was the right goal. But the code those tools generated was a mess that someone always ended up having to clean up, and the argument was structural: a tool that hides the language it's writing only hides it until something breaks, and when something breaks you need someone who can read the language.

In 2026, the old debates are new again, this time with AI. Tools like v0, Lovable, Bolt, Cursor's chat mode, and GitHub Spark take different approaches to AI but share the same promise to would-be developers: describe what you want, get an app or website, no "coding" required. Vibe code if you want to. You should. It's currently the cheapest, most accessible path from idea to working software we've ever had, and refusing on principle puts you behind the people who don't refuse.

But when you prompt the AI, it writes code. Real code, in real languages: JavaScript, CSS, HTML, Python, SQL, depending on the tool. The output is uneven, sometimes subtly broken, often fine until it isn't, and when it isn't, you need to read the language to fix what's there. The tool is generating a language for you. That part doesn't go away because the interface is a chat box instead of a sidebar.

We've entered the era of WYPIWYG: what you prompt is what you get. The people I see generating the cleanest output from AI tools are the ones who learned the languages (JavaScript, React, CSS) before they reached for a prompt. They use the AI the way an experienced developer uses autocomplete that occasionally hallucinates: gratefully, intentionally, but always with a bias toward verification, not blind trust. They prompt better because they understand what they're asking for, and they debug better because they understand what they got.

Applied to today, the FrontPage lesson is about the language underneath. You're writing one whether you realize it or not, so spend some time learning it. Most of the modern web is built using tools and frameworks that hide the gnarly parts for you. People who refuse to use any of them rarely ship as much as people who pick their tools carefully. But the day a tool you depend on misbehaves is the day you start wishing you'd spent some time on the language underneath.

What FrontPage did for me, AI is doing for a new generation. It opens a door. Some kids will walk through that door, get a website made for class, and never look back. That's fine. Others will notice that something is happening behind the prompt and want to understand it. To those kids: read the code that gets generated. Open up the files, look at the JavaScript and the CSS, start asking why. The languages will reward you, the way they rewarded me.