Explained: How AI-powered assistants are altering the lives of coders

AI-powered assistants are enabling coders to save time and be more productive

Rohit Sharma (name changed), a 26-year-old software developer from Pune, was thrilled to find his artificial intelligence (AI) pair for JavaScript programming. No longer burdened with lengthy function calls, complex syntax, or nested structures, the coder is able to save at least 35 per cent of his time at work.

With significant breakthroughs in natural language processing and deep learning technology, AI assistants are now able to suggest code snippets and entire programming functions to coders in real time, based on the context of the code. These assistants try to read the mind of the coder and suggest the next part of the programme.

“The code completion feature of AI tools has been a tremendous help. It suggests completions for partially written code lines. It’s like having an assistant that understands my intentions and assists me in completing the code swiftly. This not only saves time but also reduces the likelihood of typos or syntax errors,” says Adarsh Thalikandhitil, a developer working with Tech Mahindra in Bengaluru.

Popular AI companions include GitHub Copilot (based on OpenAI’s Codex), Visual Studio IntelliCode by Microsoft, Amazon CodeWhisperer, and Tabnine.

Thalikandhitil, who has been in software development for the past three years, discovered an AI assistant around six months ago.

“One of the most remarkable ways in which AI tools have saved me time is through code generation. With GitHub Copilot, I can quickly generate code snippets based on my context, eliminating the need to write everything from scratch. This has been especially valuable for repetitive or boilerplate code that I often encounter in my projects,” he says.

GitHub Copilot uses OpenAI Codex, trained on billions of lines of public code, to suggest code and even entire functions in a developer’s integrated development editor. While such tools offer significant assistance, it remains crucial for developers to have an understanding of coding language syntax.

Soham Shirole, a Pune-based freelance Android and JavaScript developer, says that the AI assistant has helped him explore new technologies, frameworks, and languages and broaden his understanding of different programming paradigms by saving time on repetitive tasks. “I can take online courses, participate in coding challenges, or engage in open-source projects to further develop my expertise.”

Complementary relationship

The advent of AI has transformed the lives of coders in many ways. “AI coding tools have improved code quality, increased productivity, and reduced production-level incidents. We can anticipate the continued acceleration of AI-driven programming tools. These advancements will likely enable more individuals without programming backgrounds to work with technology, while programmers themselves focus on tackling increasingly complex software challenges,” says Hari Krishnan Nair, co-founder of Great Learning, a Gurugram-based ed-tech company.

Experts have suggested that AI cannot replace coders but complements them. “The collaboration between software developers and AI has resulted in a decrease of 45-50 per cent in task completion time. Rather than replacing human coders, AI has become a powerful tool that they can leverage to enhance their work,” says Sanjay Shetty, director of professional search and selection at Randstad India, a human resource services firm.

“While AI systems excel in automating certain aspects of coding, they still need historical data to function effectively. Hence, they lack the ability to think creatively, make intuitive decisions, and understand human behaviour comprehensively. These are areas where human coders excel. The human touch remains essential in developing new software applications that require empathy, intuition, and a deep understanding of user needs,” adds Shetty.

New opportunities

AI is also creating new opportunities for coders. “The demand for coders with expertise in machine learning, deep learning, and AI algorithms has surged. While AI has automated certain tasks, it has also created new opportunities for coders,” says Mayank Kumar, co-founder and managing director of upGrad, an edtech platform. “The need for coders to build and maintain AI systems, develop AI-driven applications, and ensure the ethical and responsible use of AI has grown substantially. This has resulted in an overall increase in coder demand, especially for those who can navigate the intersection of AI and their specific domains.”

Despite reducing the time-consuming tasks of coders, AI tools have not disrupted their pay-scales so far, according to multiple staffing firms.

The monthly average salary for entry-level software programming/developer roles in the current year is Rs 45,000-60,000. Those with experience of three or four years can draw up to Rs 1.1-1.5 lakh a month, according to NLB Services, a staffing solutions firm.

“The average salaries of software programmers/developers are influenced by factors such as experience, skills, education, location, and industry demand. Over the past 12 months, the salaries offered for software development roles have remained consistent without significant changes,” says Krishna Vij, business head, IT staffing, at TeamLease Digital.

AI-enabled coding assistants, he adds, could address the talent shortage and enhance software development. “Skilled programmers who can leverage AI tools, exhibit creativity, and tackle complex problems will continue to be in demand, potentially leading to maintained or increased salaries.”

Data from the employment website foundit (a global online provider of employment solutions formerly known as Monster) shows that average salaries for software programming and developer roles have not seen much difference over the last year.

“There is much scope for newer roles with the growth of AI. The demand for AI talent has grown 11 per cent in the last six months, with the IT industry harbouring the biggest demand for such roles,” says Sekhar Garisa, chief executive officer of foundit.

To be sure, AI is fast evolving and its true ramifications have yet to be seen. “We are still in the very early stages of the massive shift in how software is built and run, and while tools like AI-powered coding assistants are already making developers more productive and able to build applications much more quickly, we expect many more changes in how developers work,” says a spokesperson for MongoDB University, an online education platform for coders.