# AI-Generated Code Cleanup Spawns New Industry Lack of discipline in AI-assisted software development creates new business opportunities. - **Source:** [www.theregister.com - Articles](https://brifing.si/source/2c34ea50-d341-4a10-925d-12c012a9ad10) - **Published:** 2026-08-22 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/89c069e9-0ba6-4ca1-8a77-e41ca9b692ca ## AI coding indiscipline creates new sector With the rise of generative AI, 'vibe-coded' applications created by those without software experience are causing major issues. Konstantin Klyagin, founder of Lisbon-based Redwerk, notes a new industry has emerged to clean up such apps. Clients seek help fixing quickly AI-generated applications that are hard to maintain and contain security vulnerabilities. ## Superficially working but problematic applications Klyagin emphasizes that AI-generated apps may look good externally but have serious logic flaws. In one New York client's app, they found price inconsistencies in payment pages and security flaws allowing payments without profile creation. Forms incompatible with screen readers and inadequate test coverage were other common issues. ## Code duplication and infrastructure problems One of the most prevalent issues in AI-generated code is duplication. Klyagin cites examples like price mismatches between front pages and onboarding flows. He notes these problems stem from lack of proper software architecture knowledge. Technical teams handle these issues more easily than non-technical founders. ## Evolution of traditional code review processes Pre-AI, clients typically requested code reviews or pre-acquisition audits. But with tools like Claude Code and Codex, QA and bug fixing needs have multiplied. Klyagin says his company manages this increased workload with AI's help. While AI accelerates everyone, discipline remains essential in development processes. ## AI boosts productivity but requires discipline Klyagin states AI encourages faster code production but some use this power recklessly. Redwerk focuses on teaching AI coding discipline to founders and enterprises. 'Discipline isn't going anywhere,' Klyagin stresses, emphasizing the continued need to define what's being built and to test it thoroughly. ## Lack of production readiness AI-generated code often isn't production ready. Klyagin clarifies cleanup involves more than reducing lines - it makes apps truly customer-ready. He notes code volume isn't a success metric, and less code doesn't automatically mean better software. The focus should be on maintainability and real-world functionality. ## Validation and security vulnerabilities Critical issues in AI-generated apps include validation gaps and security flaws. Klyagin explains these apps focus on ideal scenarios but fail to handle arbitrary user behavior. Serious vulnerabilities like permission errors allow users to skip directly to payment pages, creating substantial risks. ## Software development discipline in the new era As AI-assisted coding proliferates, development discipline grows more crucial. Klyagin emphasizes the need to steer AI models toward proper architecture and infrastructure. Non-technical founders particularly struggle with this discipline. Companies like Redwerk fill this gap by promoting robust development practices in the AI age. ## Sources covering this story - https://www.theregister.com/ai-and-ml/2026/08/22/ai-slop-is-good-for-business-if-you-know-what-youre-doing/5291382 — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._