Developers and AI: Widespread Use, Low Trust, Burnout

⚡

Key Takeaways

1Less than 10% of developers trust AI for key decisions

2Nearly half describe themselves as complacent, with burnout being the main reason

3The adoption of AI tools is widespread, but experiences vary greatly

💡Why it matters — The rise of AI is transforming the developer profession, amidst loss of meaning, burnout, and new opportunities, while raising issues of trust and control.

According to a global survey conducted among 30,903 developers, the adoption of coding assistants is widespread, but the trust placed in these tools for critical decisions remains very limited. Well-being indicators are deteriorating, with burnout at the forefront. Contrasting testimonies illustrate a profession undergoing rapid transformation, caught between a decline in meaning and new possibilities.

Trust Remains Low and Skepticism Organizes

Less than 10% of the developers surveyed would be willing to entrust important decisions to AI. Nearly 80% consider source attribution essential for evaluating the generated code or the technical responses provided by these tools. Prashanth Chandrasekar, CEO of Stack Overflow, believes that this skepticism serves as a crucial safeguard as agents automate more tasks. According to him, developers' trust is built through transparency, source attribution, and a solid technical context.

Widespread Adoption and Intensive Use of Tools

About two-thirds of developers use AI coding assistants or agents, and 62% resort to general-purpose AI chat tools. Among those who use them daily, 31% spend four hours or more per day on them. While these tools accelerate work, Stack Overflow's survey notes that they also leave many professionals anxious and exhausted.

Low Morale: Complacency, Discontent, and Burnout

Among the 30,903 developers surveyed in 169 countries, nearly half describe themselves as complacent in their current jobs, one-third as unhappy, and just over 22% as happy. Among the complacent, burnout or technological fatigue is the main reason cited by 21% of respondents. The rise of AI-assisted coding contributes to an identity crisis and very high anxiety for some. Annie Wright, a psychotherapist and executive coach, observes that her clients no longer just describe themselves as exhausted; they now question their identity if their work were to be automated, a significant shift from three years ago.

Opposing Experiences: Loss of Gratification or New Capabilities

Marc Backes, a senior software engineer at Directus, feels that AI has made his work less rewarding: he spends less time designing and more time delegating to agents, in an industry where writing or even reading code by hand is now seen as outdated. He fears being left behind if he doesn't stay up to date and finds the pace of AI difficult to keep up with. In contrast, Simon Willison, founder of the open-source project Datasette, believes that AI coding tools make his work more satisfying, allowing him to undertake projects that would have previously required a team. He explains that being his own manager protects him from certain pressures, even though he finds the pace of innovations exhausting, and anticipates increased competition due to lower development costs. He remains optimistic and views the effective use of agents as a deep skill he possesses.

A Profession in Transition Due to Assisted Coding

Software engineering jobs are evolving rapidly, with an increasing role of AI-assisted coding in daily tasks. Some professionals see their role reduced to execution, which heightens tensions regarding the meaning and content of their work.