Renowned hacker George Hotz has raised concerns about the impact of AI programming agents on software quality. After six months of using AI tools in real-world projects, Hotz concluded that while these tools can initially boost productivity, they often lead to code that is superficially polished but riddled with underlying issues. He warns that less skilled engineers may struggle to identify these problems, potentially leading to a decline in overall code quality masked by increased output volumes. Hotz's warning contrasts with the views of AI researcher Andrej Karpathy, who recently joined Anthropic and believes next-generation AI models are transforming software development. Despite differing opinions, the debate highlights the potential risks of rapidly adopting AI coding tools across the software industry, with Hotz cautioning that widespread implementation could compromise software quality over the next two years.