Tech leaders describe obsessive AI agent use as 'psychosis'; research shows 90% of firms report zero measurable productivity gains

Y Combinator CEO Garry Tan and OpenAI cofounder Andrej Karpathy both publicly used the term 'psychosis' to describe their relationship with AI agents. Tan claimed to be shipping 37,000 lines of code per day across five projects while running YC full-time; a developer's analysis found the resulting website made 169 server requests, shipped 28 test files to production, and loaded 78 JavaScript controllers for non-existent homepage features.

A wave of 'AI orchestration' platforms—including Paperclip (30,000 GitHub stars), Autoflowly, AgentShelf, Alacritous ($3,000/month), and RuFlow—provide dashboards, org charts, and budget controls that create the feeling of commanding agents (such as 'CEO' or 'VP of Marketing'), but lack mechanisms to define requirements or measure whether agents produce business outcomes.

An NBER study of nearly 6,000 CEOs and CFOs across the US, UK, Germany, and Australia found roughly 90% of firms reported zero measurable impact on productivity or employment from AI over the past three years. Separately, every 25% increase in AI adoption correlates with a 1.5% decrease in delivery speed and a 7.2% drop in system stability; teams using AI heavily complete 21% more tasks but experience 154% larger pull requests and 9% higher bug rates.