Their jobs were among the first to be changed by AI. Now some are trying to avoid it

In 2024, software engineer Laura Housh started integrating AI into her coding and other professional tasks. Just as she began to form a mixed opinion about AI, she now finds herself embracing and resisting the technology simultaneously. Housh is keen to highlight AI's limitations, stating that it cannot match the contextual understanding of human coders.

Survey data from Stack Overflow, a leading platform for programmers, supports her claim. Six in ten tech workers who participated in the recent survey reported attempts to avoid using AI. Housh isn't alone in her skepticism – the survey, which included responses from 30,000 software developers across 169 countries, revealed that software engineers were among the first industries to adopt AI.

According to a recent Indeed Hiring Lab report, "software developer" was identified as the job most likely to be transformed by AI. However, the survey also revealed that 65 percent of responders use coding assistants or agents, while a substantial 90 percent of tech professionals globally use AI in their jobs, a 14 percent increase from the previous year.

Despite this widespread use, job dissatisfaction among respondents has risen, with nearly one in three feeling unsatisfied in their current roles, compared to 28.4 percent last year. A significant portion of these developers describe themselves as complacent, with only 22.3 percent reporting genuine happiness. Burnout, tech fatigue, economic uncertainty, and unclear career progression pathways were cited as the primary reasons for their discontent.

While AI advancements may exacerbate these concerns, experts like Stack Overflow CEO Prashanth Chandrasekar argue that the real challenge lies in the rapid pace of change. Senior programmers may find their established skills less relevant, while entry-level workers now need to operate as if they possess three years of experience, given the accelerated AI integration.

AI's limitations become glaringly apparent when developers realize its lack of broad general knowledge and specific job context. Housh and her colleague Audrey Eschright, a 20-year veteran in the field, agree that AI cannot replicate the nuanced understanding and problem-solving skills of seasoned developers. Eschright emphasizes that writing code is integral to problem-solving, and using AI to generate code can leave developers out of the loop, stripping away the personal touch that comes with manual coding.