top of page
3d-ai.png
agentic-horiz-new.png

Gio Scalzo

AI & Engineering Lead
greg-fawson-foto.1024x1024 Kopie.jpg

Giordano Scalzo is a London-based AI and engineering lead at NewDay and CTO of a set of consumer AI apps, including Pocket Sommelier (AI wine pairing) and the AI-driven wellness apps Neurona and Selena. A published author of six books, he spends much of his time exploring what happens when AI meets real engineering work. He speaks regularly at communities such as NSLondon and Londroid and, more recently, at the LeadDev New York (LDX3) conference, ships AI products, and is especially interested in the gap between impressive demos and messy production reality.

Just One More Prompt: Why Agentic Coding Hooks Us

At 10:48pm the agent is nearly there. The feature works, almost. You could stop. Or ask for one more tiny thing: a better empty state, a shortcut, nicer animation. One more prompt. Then another. By 1:30am the app does more than you planned, and you're too wired to sleep.

Agentic coding can form an addictive loop: low friction, fast feedback, variable rewards, and that gambling-like feeling that maybe the next try pays off. Neurodivergent developers may feel it sharply. The uncomfortable part is that anyone can get caught.

I want to name that pattern without turning it into panic. We'll look at what the loop feels like, when flow becomes compulsion, and why "just one small feature" is so hard to resist. Then we'll get practical.

You'll leave with:
- Warning signs to watch for
- A model: tool loop + person + team culture
- Stop rules for late-night feature sessions
- Team habits that protect everyone without singling anyone out

bottom of page