Synopsis
Developers are experiencing burnout from intense AI coding sessions. Tools like Claude Code and OpenAI's Codex are generating vast amounts of code, overwhelming even top engineers. This 'brain fry' leads to mental fatigue and slower decision-making. Some researchers have quit leading AI firms citing exhaustion. Organizations are exploring strategies to mitigate this growing issue.But that is no longer the case. After coming home, he now fires up his computer; multiple projects that he works on - from AI interview platform for radiologists to building AI agentic workflow for his startup - come alive in the dead of night. Sivasailam runs five Claude Code instances, one Google Antigravity and one OpenAI’s Codex. “There is so much code being generated that you can easily get lost in building and verifying them,” he told ET.
Sivasailam is part of a growing number of builders, who are spending more time than ever in intense coding sessions that results in burnout for some. He says this fatigue is quite widespread in his network. He sees some of the top engineers in his team are burning themselves out. This cognitive overload has overwhelmed even the best of engineers working in frontier labs. Multiple people working in OpenAI and xAI have quit in recent months after feeling totally drained.
On February 26, OpenAI researcher Hieu Pham posted on X that he was leaving the firm due to burnout. “I cannot believe I would say this one day, but I am burnt out. All the mental health deteriorating that I used to scoff at is real, miserable, scary, and dangerous,” he wrote on X. He said that he would take a break and move his family to home country Vietnam, where he wanted to try something new and allow himself time to heal. A couple of weeks later, another researcher Haotian Liu announced he was quitting Elon Musk’s xAI after two years of intense work. He helped build the video generation model grok imagine. “...shipping it as a great product used by millions, all within 6 months, at age 28: I feel proud. But now it’s time for me to move on. I’m burnt out..,” Liu wrote on X.
The brain fry
A recent The Harvard Business Review report, authored by CXOs at Boston Consulting Group, has termed this as ‘brain fry’ referring to the mental fatigue from excessive use of AI tools beyond one’s cognitive ability, with 14% of the workers reporting mental fog, headaches and slower decision-making. Multiple founders and industry watchers ET spoke to shared that for power users of the technology, the work had only intensified instead of easing. As technology has improved productivity, more complex workloads have been thrown at developers. Unlike in the past, the number and kind of decisions a developer needs to make has changed, explains Prasanna Krishnamoorthy, managing partner, Upekkha, an AI accelerator. “Earlier, these developers were making coding decisions. But now with AI doing most of the actual coding, the kind of decisions they make are changing, and that is also creating burnout,” he said.
Before the advent of AI coding tools, developers were making decisions about what language and datasets to use. Now that has given way to higher level decisions about the kind of architecture, design and products to be built, which were usually taken by senior developers, and often within a short time adding to the stress. “Given the pace of developments, all these decisions need to be taken quickly,” he added. That is not the only problem. For many developers, who are exploring AI and building side projects , these tools are often addictive.
Ashwin (name changed to protect identity), an AI researcher at one of the frontier AI companies, usually builds a couple of hobby projects on the side. But in the past year, he was able to build at least 4-5 side projects due to the sheer capabilities of these models. “It is like an addiction, because you are seeing how the code works and the more it does, the more you want to continue, leading to burnout,” he added. Since you are not looking for an outcome but building it as a toy project, this becomes addictive like gambling, says Krishnamoorthy.
What can be done?
Ashwin has now started taking breaks between coding marathons. “Earlier I would go for weeks working on one side project after another, till 2 am. But now, I take a break for a week or two between intense coding sessions.” 5C’s Sivasailam says he sits down with his team to talk about this topic. “The first thing to do is reset the mental framework on what can be done using AI,” he says. One of the challenges is that for many senior developers it is harder to accept the shift that the job boils down to managing AI agents instead of writing code. The study by Harvard Business Review points out that where one is using AI tools is also important. It noted that using AI to reduce repetitive tasks see 15% lower burnout rates compared to those who do not use AI. “At the organizational level, directionally, practices like providing clear AI strategy and offering training seemed to help,” the report noted.
