From AI Skeptic to AI Practitioner: A 30-Year Technologist’s Perspective

Why I Was Skeptical

After 30+ years in IT—from mainframes to cloud—I’ve learned to be cautious with any technology that promises to 'revolutionize' or 'disrupt' the world. AI was no exception. I was curious, but unconvinced. Beyond chat, I wasn’t sure how much practical value I’d get from it.

That changed when I started working with AI agents—especially through Google’s antigravity, a tool comparable to GitHub Copilot. I’m not recommending one over the other, but I want to be transparent about what I used. I also tried all the LLMs available out of the box in antigravity and had equally strong results for the kinds of tasks I threw at them.

Project #1: A Chrome Extension That Finally Got Built

One of my long-standing personal projects was a Chrome extension I needed for my own workflow. It had been sitting in the 'someday' pile for years—partly because I didn’t want to spend time learning EPUB standards, Chrome extension APIs, or the nuances of Google Play Books integration.

With antigravity, I finally built it. It wasn’t instant. It took over 100 revisions, countless iterations, and a lot of back-and-forth. But the agent kept researching, learning, and adapting—pulling documentation, checking standards, and redesigning approaches when it hit dead ends.

Once it worked reliably, I open-sourced it. It’s currently pending review in the Chrome Web Store.

Project #2: A Professional Website in Under an Hour

My old website was a template-based, half-finished placeholder. I wanted something clean, professional, and easy to maintain—but I didn’t want to spend days designing layouts or generating assets.

With antigravity, the core site was done in about 30 minutes. Most of that time was spent generating images. The real work—testing across devices, refining layouts, and building an easily updatable blog—took half a day. This article is the first post in that new blog.

Project #3: A Remote LLM Development Environment

In an earlier experiment (linked in my previous article), I used AI purely as a conversational guide to help me set up a remote LLM development environment. Even in that 'chat-only' mode, it accelerated the process dramatically. It wasn’t writing code for me—it was acting as a knowledgeable partner, helping me navigate decisions and avoid pitfalls.

What Changed My Mind

These agents don’t just generate code—they learn. They look up documentation. They read standards. They adapt their approach. They fill in the gaps you don’t want to spend your life learning.

That freedom—the ability to skip the one-time, low-ROI learning curve of obscure standards, APIs, or formats—is not just convenient. It’s transformative.

The Real Impact

• Time saved was substantial.
• Mental load reduced was even more meaningful.
• Ideas that had been stuck in limbo for years finally became real.

That last point is the one that changed my perspective. AI didn’t just make me faster. It helped me finish projects that had been stuck in limbo for years.

Where I Go From Here

I’m still cautious. I still test everything. I still don’t recommend tools lightly. But I’m no longer on the sidelines. I’m an AI practitioner now—and this is just the beginning.