Adaptability, not any single tool, is the skill engineers need now
IEEE Spectrum examines what it means for engineers to be 'adaptable' as AI reshapes daily work, arguing the advice is repeated constantly but rarely explained. Samantha Brunhaver, an associate professor of engineering at Arizona State University who received a National Science Foundation award in 2020 to study workplace adaptability, has spent the project interviewing managers, early career employees, and undergraduates. She found employers all say adaptability matters but define it differently: in software engineering it might mean adjusting to daily tool turnover, while aerospace or biomedical engineers track changing procedures and regulations. Brunhaver defines the skill generally as recognizing that change or uncertainty is occurring and responding effectively to it, and breaks the process into three steps: perceive a need to adapt, evaluate the options, and act.
The disruption is measurable. A June 2026 report from PwC on AI's effects found jobs in technology, media, and telecom are experiencing the fastest pace of skill turnover of any sector. The World Economic Forum's most recent Future of Jobs Report, published in 2025, found employers across all sectors expect 39 percent of workers' core skills to change by 2030.
Jenna Butler, a Microsoft research scientist who studies developer well-being and productivity, describes the current period as a 'chaos period' typical of paradigm shifts, one she calls 'the uncomfortable middle': code review volume is rising sharply, and, in her words, 'if you ask 20 developers, you get 23 different ways of working with it.' She expects software engineering to lean more on prompting models and managing agents than on writing code line by line, predicting people drawn to the field for problem-solving will find that shift enjoyable while those who prize the craft of writing code will not.
Andy Hunt, who coauthored The Pragmatic Programmer in 1999 and revisited it for a 20th-anniversary edition, argues the core of the job is more stable than it looks: problem solving and communication endure regardless of which tools are current. He compares identifying with a single language, such as calling oneself a Java programmer, to a carpenter calling themselves a hammer user, and says hiring processes that filter resumes by specific languages or years of experience work against the more expansive mindset the job actually needs.
Both researchers put part of the responsibility on institutions. Brunhaver says educators should be explicit about what adaptability means and give students varied real-world experience, such as internships and team projects, before they graduate. Butler urges organizations to set aside dedicated learning time, even just an hour a week, without expecting output during it, warning that developers given a new tool but no real guidance, while still pressured to ship more, tend to fall back on what they already know and burn out.
Key facts
- The World Economic Forum's 2025 Future of Jobs Report found employers across all sectors expect 39 percent of workers' core skills to change by 2030.
- A June 2026 PwC report found technology, media, and telecom jobs have the fastest pace of skill turnover of any sector amid AI's effects.
- ASU's Samantha Brunhaver, running an NSF-funded study since 2020, defines adaptability as recognizing change is occurring and responding effectively, broken into perceive, evaluate, and act.
- Microsoft's Jenna Butler says code review volume is rising and developers are diverging into wildly different ways of working with AI: 'ask 20 developers, you get 23 different ways.'
- Pragmatic Programmer coauthor Andy Hunt argues the job's core, problem solving and communication, is stable even as the tools around it change constantly.
Why it matters
AI has made 'be adaptable' the standard career advice for engineers, but the article's central point is that almost nobody explains what the skill actually requires or how to build it. That gap matters because the disruption behind the advice is now quantified: PwC's June 2026 report puts technology, media, and telecom ahead of every other sector on skill turnover, and the WEF projects 39 percent of core skills changing across all industries by 2030. Treating adaptability as a vague virtue rather than a trainable process leaves engineers and the people managing them without a way to act on it.
Who it affects
Working engineers navigating daily tool churn, students and educators deciding what a curriculum should teach when specific languages or frameworks may not last, and managers responsible for their teams' skill development. Software engineering is singled out as one of the fields most exposed, since Butler describes code review workload rising and expects more of the job to shift toward prompting models and managing agents rather than writing code directly.
How to use it
Brunhaver's three-step model gives a concrete process: notice that change or uncertainty is happening, weigh the available responses, then act, rather than waiting for the situation to resolve on its own. She also stresses metacognition, reflecting on how you are adapting, as what makes the skill effective, and ties agency (believing you can get through a situation) to whether people actually adapt. Hunt's practical suggestion is to build systems thinking rather than identify with one tool or language, since the job's stable core is problem solving and communication, not any specific technology. For managers, Butler's concrete ask is to set aside real learning time, even one hour a week, with no expectation of output during it.
How solid is it
The piece rests on named, on-record sources with clear credentials: Brunhaver's claims come from an NSF-funded research project running since 2020 that interviewed managers, early-career employees, and students, though the article does not say how many; Butler speaks from her own research role at Microsoft on developer well-being; Hunt draws on his experience writing and later revisiting The Pragmatic Programmer. The two headline statistics, the WEF's 39 percent figure and PwC's sector-turnover finding, are attributed to named, dated reports (2025 and June 2026 respectively) rather than presented as the outlet's own analysis, though the article does not link or quote the underlying PwC figures directly.
Risks and caveats
The article does not give a specific number behind its claim of a 'significant increase' in code review for software engineers, nor does it name which AI tools or models engineers are described as adopting, so those points should be read as general trends rather than measured figures. It also does not specify how many people Brunhaver has interviewed for her ongoing study. Butler is candid that the guidance is easier stated than followed: without enough instruction, developers under pressure to be more productive tend to fall back on familiar tools instead of adapting, and she says she expects the next several years to be genuinely difficult for the field.
“We tell engineers that they need to be adaptable when they graduate, but we don't actually explain what that means, demonstrate what that looks like, or help make sure that they're developing it.”
— Samantha Brunhaver, associate professor of engineering, Arizona State University