A couple weeks ago someone sent me Matt Shumer’s post on AI.

Instead of immediately dismissing it as hype, I decided to test the idea myself and see if it was even plausibly true.

The first week was rough.

I spent most of my time fighting with the development environment:

âť“ Can I get this thing connected to Git?
❓ Why doesn’t this work?
âť“ Why am I getting conflicting instructions?
âť“ Why? Why? Why?

By the end of the week, I was ready to write the whole thing off. Fraud. Charlatan. Just another overhyped AI claim.

Then I tried a different AI platform: Claude

Within two hours, it was:

⚙️ Syncing with a Git branch
đź“ť Documenting my code
🔧 Suggesting improvements
🎨 Redesigning parts of the UI when prompted to “better organize the interface according to modern design patterns”

The redesign was shockingly good.

Does it still make mistakes? Of course.

In one experiment, I thought it had completely ignored the requirements I gave it. So I changed my workflow. Instead of saying “fix this,” I asked a different question: “Why doesn’t this work?”

The issue turned out to be completely unrelated to the code — it was a permissions problem.

And the other “missing” requirements?

They were actually implemented.

I had expected it to move functionality from one page to another. Instead, it duplicated the functionality correctly. Once I clarified, it removed the old implementation without issue.

The article didn’t make me a believer. The experiments did.

Work isn’t disappearing tomorrow, and AI isn’t replacing engineers overnight.

But one thing feels clear:

🚀 This is a great time to learn how to leverage AI to your advantage.

Skeptical developers — have you tried an AI coding tool yet? Yes / No — and why?

#ArtificialIntelligence #SoftwareDevelopment #AICoding #FutureOfWork