Adam.dev shares his experience of developing an unhealthy obsession with AI-assisted coding, which led to burnout and disrupted sleep, prompting him to take a five-week break to recover and regain balance. He emphasizes the importance of using AI tools thoughtfully to enhance craftsmanship rather than productivity alone, highlighting the need for awareness and self-care amid the fast-paced, competitive AI development culture.
The video features a candid conversation with Adam.dev, a seasoned developer known for his work on Open Code and his journey through what he terms “AI psychosis.” Adam recounts how he initially approached AI tools cautiously, building code mostly by hand despite the emergence of models like Claude Code. Over time, particularly with the release of more advanced models like GPT-5.2, he began to lean heavily into AI-assisted coding. This shift led to an intense, obsessive engagement with AI, characterized by running multiple AI instances simultaneously, constant refactoring, and a compulsive need to prompt and review AI outputs, which ultimately culminated in burnout and disrupted sleep patterns.
Adam describes this phase as akin to a dopamine-driven slot machine effect, where the intermittent successes of AI-generated code hooked him into a cycle of continuous use despite diminishing returns. The competitive landscape of AI coding tools and the pressure to keep up with rapid advancements exacerbated his obsession. Social media, particularly platforms like Twitter (now X), contributed to his mental strain by exposing him to constant updates and comparisons with competitors. This relentless pace and the culture of overwork in tech led Adam to take an unprecedented five-week break from programming and AI tools to recover.
During his hiatus, Adam disconnected from AI and social media, spending time offline with family, watching shows like Breaking Bad, and engaging in activities like golf and tennis. This period allowed him to regain perspective and mental clarity. Upon returning, he noticed a healthier approach within his team towards AI-assisted development, focusing more on outcomes rather than sheer output volume. Adam emphasizes the importance of slowing down, reviewing AI-generated code carefully, and using AI as a tool to enhance craftsmanship rather than as a means to churn out code rapidly.
The discussion also touches on the broader phenomenon of AI psychosis within the developer community, highlighting signs such as sleep disruption, obsessive multitasking with AI agents, and the difficulty of recognizing the problem without external intervention. The panelists explore the cultural and psychological factors that fuel this behavior, comparing it to other addictive patterns like gambling or gaming. They suggest that recognizing unhealthy habits, maintaining boundaries, and having supportive social networks are crucial for those deeply immersed in AI development environments.
In closing, Adam reflects on the ongoing nature of his journey with AI, acknowledging that while he has found a more balanced approach for now, the rapid evolution of AI technology means the challenge is far from over. He advocates for a measured, outcome-focused use of AI tools and expresses hope that others can learn from his experience to avoid similar pitfalls. The conversation underscores the need for awareness and self-care in the fast-paced, hype-driven world of AI-assisted programming.