Old-school programmers are reportedly more attuned to 'vibe coding', according to a recent study
In a recent survey by Fastly, it has been revealed that senior developers are increasingly using AI-assisted coding to deliver software more efficiently. This shift comes as no surprise, as senior engineers are often tasked with more than just writing code, juggling responsibilities such as testing, architecture, and mentoring.
The impact of AI on job enjoyment is noteworthy, with over 30% of respondents finding it significant. However, the larger consequences of widespread automation remain uncertain.
As AI plays an increasingly prominent role in software development, senior developers are spending extra effort reviewing for machine-created flaws to ensure the quality of their work. Junior developers, on the other hand, are expected to understand the ramifications of their code more in the future.
The survey also found that more senior engineers understand the second and third effects of their code in how it relates to users and their community. In contrast, fewer than half of younger coders felt that AI coding sped up their work.
Interestingly, a third of senior developers produce more than half of their finished work through AI code generation. This statistic contrasts with the less frequent use of AI tools by junior programmers. Only 13% of junior programmers report a similar level of use of AI code generation as senior developers.
AI coding tools used by senior developers include GitHub Copilot, developed by GitHub (now owned by Microsoft), OpenAI's tools like ChatGPT and enterprise AI solutions, and other platforms such as Amazon SageMaker, DataRobot, Apple CreateML, and Akkio.
Austin Spires, senior director of developer engagement at Fastly, suggests that senior engineers use AI to prototype quickly to recreate the "fun dopamine hit" that drew many developers to programming in the first place.
However, not all developers are on board with this shift. Many younger programmers prefer to hand-craft solutions themselves, and nearly one in ten developers have no idea how much power their code consumes. This raises concerns about the energy costs of AI-assisted coding, with 80% of older coders considering the energy costs of their work, compared to barely half of younger developers.
Spires finds it heartening that less experienced programmers still value traditional development. He also notes that there's not a lot of incentive for AI coding tools to disclose their carbon footprint.
As software development continues to shift towards AI tools for code generation, senior engineers are setting the tone, blending efficiency with caution as they navigate an evolving coding culture.
