About This Project
OnTilt is a project inspired by research in psychology, HCI, and digital wellbeing, exploring the behavioral patterns that emerge when developers work intensively with AI coding tools like Claude Code, Cursor, and GitHub Copilot.
What we study
The interaction loop of agentic coding tools may share structural similarities with slot machines: low cost per try, variable quality outcomes, instant retry, and near-miss effects that keep you going. We call this the "slot machine hypothesis" — not because AI tools are gambling, but because the reinforcement patterns are similar enough to warrant investigation.
What we are NOT
- Not anti-AI. We use AI tools extensively ourselves.
- Not a clinical service. Our self-check is a screener, not a diagnosis.
- Not moralizing. This is about work hygiene, like sleep hygiene — not abstinence.
Who's behind this
This project is maintained by c3z — a software engineer with three decades in the industry. Uses AI tools daily and noticed patterns worth studying.
Research basis
Our working model draws on 32 peer-reviewed sources spanning operant conditioning, behavioral addiction, dark flow research, HCI design patterns, and workplace wellbeing. Full bibliography available on the Research page.