The two engineers who built the AI brain behind one of the most popular coding platforms just told the world that AI-generated code is becoming a ticking time bomb. And they would know, because they helped light the fuse.

Mario Zechner and Armin Ronacher, creators of Pi, the core AI coding agent powering the widely used OpenClaw platform, sat down with the Wall Street Journal to deliver a warning that should make anyone shipping AI-written software a little uncomfortable. The practice of generating code through plain English prompts, what the industry calls “vibe coding,” is flooding the internet with buggy, insecure, and fundamentally broken software.

They have a name for this growing wave of digital garbage: “vibe slop.”

What exactly is vibe slop, and why should you care

Vibe slop, as Zechner and Ronacher define it, is the inevitable output of this process: code that technically works but lacks the structural integrity that keeps software secure, maintainable, and reliable over time. It skips the boring-but-essential stuff that separates production-grade software from a science fair project.

“We can play this game for a couple more months, or maybe even years, but eventually, it will catch up to us,” Zechner said.

Zechner, who previously developed the libGDX game framework, pointed to a fundamental accountability gap. When AI writes the code, human developers don’t feel the “pain” of managing bug-ridden systems. The feedback loop that traditionally forced engineers to write better code, the misery of debugging their own mistakes at 2 AM, simply doesn’t exist anymore.

The data behind the diagnosis

Ronacher’s concerns aren’t based on vibes alone, which is ironic given the terminology. He spoke with over 30 teams and found a noticeable decline in code quality directly attributable to the prevalence of vibe coding practices. Serious projects, not weekend hobby apps, are shipping what he categorizes as vibe slop.

OpenClaw itself has had a rough start to 2026 on the security front. The platform saw over 135,000 exposed instances flagged in February, followed by several chainable CVEs that required patches through April. While these security issues preceded the Wall Street Journal interview and aren’t explicitly blamed on vibe coding, they paint a picture of an ecosystem where speed has been prioritized over safety for too long.