A previously made prediction by Bill Gates has resurfaced recently: he suggested that developers could be among the survivors of the AI revolution. That observation is timely, since artificial intelligence is already capable of performing an increasing number of software development tasks autonomously.
How the developer role is changing
József Boda, co‑founder of the Codecool digital skills training centre, argues that future demand will concentrate on professionals who not only use AI tools effectively, but also understand business processes and can turn technological possibilities into tangible business value. Traditional measures of developer competence — knowledge of programming languages, algorithmic thinking and technical problem solving — may no longer suffice on their own.
"We can leave coding to AI, but we still need someone to decide what is worth building, which business problem we are solving, and how a new feature or application fits into existing systems. The reliability, security and business usefulness of the delivered solution must also be ensured, and for that human expertise remains indispensable," says József Boda.
According to Boda, the developer’s role increasingly resembles that of a conductor: rather than only writing code, the developer coordinates and directs AI agents that implement individual development steps. Beyond technical knowledge, this requires business awareness, complex problem‑solving skills and confident use of AI tools.
New skill demands, outdated hiring practices
Codecool’s labour market experience shows that companies are placing growing emphasis on roles that link technological implementation directly to business problem solving. One example is the AI‑transformation specialist, also known as a forward deployed engineer, who helps design and roll out AI‑based automated solutions by learning the company’s operations and business processes.
At the same time, recruitment practices often still test candidates with the same traditional programming tasks used years ago. "We often see companies expect a high level of AI‑tool usage from developers, yet interviews still assess the same conventional programming exercises as five years ago. It’s like asking a finance professional to perform complex calculations on paper at an entrance exam, while their job has long been done with digital systems," Boda points out.
To assess the changed competencies, companies should adopt new selection methods: during interviews they should evaluate how a candidate approaches a complex problem, which AI tools they choose to solve it, and how they verify the result.
Introducing tools is not the same as AI transformation
Boda stresses that introducing AI tools alone does not automatically increase a company’s productivity. Real AI transformation requires rethinking workflows, developing employees’ skills and establishing clear rules for technology usage.
Applying AI also raises new data security and corporate governance issues: organisations must define what business or confidential information can be shared with AI systems, how reliable the systems’ outputs are, and who is responsible for decisions based on those outputs.
In Boda’s view, the key question in the coming years will be less about whether AI can take over certain programming tasks, and more about how well companies and workers can convert the expanding capabilities of the technology into genuine business advantage.
