Startup founders are describing their experience managing AI agents with terms like "seductive," "intoxicating," and "all-consuming," highlighting a phenomenon where these tools are dramatically increasing productivity but also raising concerns about mental well-being. This intense engagement, sometimes likened to addiction or "cyber psychosis," is driven by the ability of AI agents to automate significant portions of work, from drafting code and research to summarization and task coordination. Microsoft's 2026 Work Trend Index found that 58% of AI users are producing work they couldn't a year ago, with 80% among heavy AI users, and 49% of workplace AI conversations supporting cognitive tasks.
The allure for founders is substantial; solo founders leveraging AI-augmented operations report roughly three times the revenue and twice the odds of profitability compared to those who don't. Companies like Naïve are capitalizing on this by offering infrastructure to automate most of the work in setting up and running a business, having raised $28.5 million in a Series A round. Naïve's platform, which has attracted over 30,000 developer customers, handles tasks like setting up payments, email accounts, cloud infrastructure, and even provides templates for AI SEO and full-stack SaaS apps.
Despite the productivity gains, the constant oversight and interaction required by AI agents are causing mental fatigue, dubbed "brain fry" by some researchers. OpenAI co-founder Andrej Karpathy described being in a "state of AI psychosis," shifting his coding ratio from 80/20 hand-written to 0/100 AI-delegated code and spending 16 hours a day issuing commands. Founders report sleeplessness and the need for medication to cope, comparing the agent interaction to slot machines due to the dopamine hits from successful prompts and the frustration of failures. This indicates a double-edged sword where extreme productivity comes at a potential cost to mental health, prompting calls for careful usage and better understanding of these tools' psychological impact.