For a long time, many people advised young individuals entering the job market to “learn to code.” However, now it appears that some programmers might also need guidance on this front. Namanyay Goel, a seasoned developer, expresses concern about the significant reliance of newer developers on advanced AI tools.
In a recent blog post titled “New Junior Developers Can’t Actually Code,” Goel shares his observations. He notes that many junior developers constantly use AI tools like Copilot, Claude, and GPT to help them write code faster than ever before. While this might sound impressive, Goel highlights a critical issue—many of these junior developers lack fundamental knowledge about why their code functions as it does. When asked about the reasoning behind their coding decisions or how they would handle unique or tricky scenarios, they often seem puzzled and unprepared.
Goel reflects on how the understanding gained from working through problems on their own is largely absent in today’s junior developers. He compared this issue to the past, when algebra teachers complained about calculators making students reliant on them. Goel’s main concern isn’t against AI itself; rather, he believes it acts as an easy crutch instead of encouraging developers to struggle with coding challenges and learn from them.
In earlier times, developers had platforms like StackOverflow where they could ask questions and engage in discussions with experienced professionals. While this platform remains popular, many new coders are now turning to AI for quick answers. Goel points out that junior developers can simply copy and paste error messages they encounter into chat applications and receive instant responses. Although this might speed up the coding process, Goel argues that it doesn’t promote critical thinking or problem-solving skills.
Using StackOverflow required developers to read through various expert perspectives to gain a comprehensive understanding of issues. This process, although slower, ultimately led to better knowledge about not only effective solutions but also the reasoning behind them.
Research supports Goel’s concerns. A study by Microsoft and Carnegie Mellon indicates that increased reliance on AI to find answers can weaken people’s critical thinking abilities, much like a muscle that becomes weak from lack of exercise. While the study has limitations—such as relying on self-reported participant data for measuring effort—it’s not a surprising conclusion that offloading cognitive tasks to technology can hinder mental development.
Moreover, many AI models aren’t always reliable in producing accurate code, often leading to errors. Although these tools can help speed up work, evidence suggests they might introduce more mistakes into coding projects.
Despite these challenges, Goel emphasizes that we can’t turn back the clock on AI developments. He believes that the key moving forward isn’t whether we use AI, but rather how we utilize it. Currently, he warns, there’s a risk that we are sacrificing in-depth understanding and the struggle for quick solutions. This could lead to significant consequences in the future when developers face complex challenges that demand strong foundational knowledge and critical thinking skills.