In the last two years, there has been a surge in AI coding assistants. OpenAI's competitors have introduced popular generative AI developer tools like GitHub Copilot and Anthropic's Claude. The increase in such AI coding assistants have changed the way developers work and the future of coding depends on its affordability and productivity. Also, the rapid rise in the use of low-code, no-code tools is transforming the landscape.
Many developer tools such as the Tower Git GUI, charge between $69 and $99 per user per year. This is a relatively small investment compared to developer salaries. Other tools are often free or open-source, making them accessible without direct costs.
In the US, software developers earn between $40,000 and $150,000 annually, depending on experience and specialisation, with an average salary of around $89,190 per year. In India, salaries range from INR 1.2 Lakhs to INR 13 Lakhs annually, with full-time developers typically costing companies between $10,000 and $25,000 per year, depending on experience.
Moreover, investing in these tools can lead to significant productivity gains, quickly offsetting their costs. For example, if a tool saves even a few minutes of a developer's time or helps prevent errors, it can easily justify its price in comparison to a developer's salary. The economic case for investing in developer tools is strong; the costs are low relative to the potential productivity gains and improvements in software quality they provide.
Subhojit Mondal, founder of SACHI, a talent consulting firm, notes that these tools will enable more people to create software solutions without traditional coding skills, fundamentally altering the job market.
While observing a clear segmentation among job seekers over the past three years, categorising them into-below average, average and exceptional, Mondal notes that exceptional candidates are able to secure significant salary hikes. This again comes in the backdrop of COVID-19 crisis and rise in funding among startups disrupting the job market. However, he feels that the early-career candidates are struggling to break into the market.
Sumit Mansinghani, 29, a senior data scientist at PepsiCo, echoed similar views, and said we're nearing a tipping point. He observes that while salaries have indeed risen, they are approaching saturation point, and the ongoing economic pressures will temper future hikes.
"The dramatic increases of 100% to 200% are unsustainable and will likely revert to more manageable levels of 30% to 40% as market conditions stabilise," he remarked.
As AI continues to streamline coding processes, one can expect software development to become cheaper and the job market for coders to evolve, creating a more accessible environment for talent, although at lower price points.
Furthermore, Using AI for coding could help bridge the experience gap between junior and senior developers, as advancing typically requires years of experience and personal projects.
Tech giants rely on developers to improve the appeal of their platforms, while non-tech company leaders seek more coders to drive digital transformation in hopes of boosting productivity and attracting consumers.
"Coding is also full of feedback loops and tests that check if software works properly," noted Nathan Benaich, of Air Street Capital.
The rise of AI coding tools has hastened, with around 250 startups creating them, according to PitchBook. Whereas, Leading tech firms are at the forefront, According to a GitHub survey, over 97% of 2,000 respondents from the US, Brazil, Germany, and India reported using AI coding tools at work. GitHub, owned by Microsoft, launched Copilot in 2022, which generates code on demand and has 2 million paying users.
Additionally, a study analysing data from Microsoft, Accenture, and an anonymous Fortune 100 electronics manufacturer found that generative AI code suggestion tools can boost software developer productivity by 26%.
AI's usefulness in coding remains somewhat limited. A survey by Evans Data found that 35% of coders said AI tools save them just 10% to 20% of their time, primarily by generating basic code. However, these tools are not flawless. A study by GitClear revealed a decline in code quality over the past year, potentially due to the use of AI models. Additionally, a survey by Synk reported that over half of organisations had identified security issues with AI-generated code.
Although there is no standard definition, the number of GitHub users from poorer countries exceeded those from wealthier nations. Online forums, such as Stack Overflow, contain vast archives of questions and answers shared by coders.
Whenever a new technology emerges, the immediate concern is job losses. Historically, while some jobs disappear, new ones emerge. In the world of generative AI, the same debate persists, though the impact may be different this time.
Many industry experts interviewed by AIM believe that generative AI will increase productivity and make an organisation's workforce more efficient. However, in a 2023 interview with The Atlantic, OpenAI CEO Sam Altman, said, "A lot of people working on AI pretend that it's only going to be good; it's only going to be a supplement; and no one is ever going to be replaced," he continued, "Jobs are definitely going to go away, full stop."
Surprisingly, Altman isn't the only CEO of a major generative AI company predicting job losses due to AI. Back then, Emad Mostaque, CEO of Stability AI, said on a call with UBS analysts that AI could result in most of India's outsourced coders losing their jobs.
ChatGPT, powered by OpenAI's GPT models, has impressive coding capabilities, having passed a Level-3 coding interview at Google. Trained on amounts of human-written code, LLMs can generate code much faster, potentially displacing entry-level coding jobs.
Natarajan Radhakrishnan, president and CIO, Hinduja Global Solutions, believes that AI will not take away coding jobs entirely, however, "The lower end of coding may indeed face replacement, not just in India, but globally," he told AIM.
With the rise of generative AI, lower-level coding tasks have diminished in value, as LLMs can easily generate code, rendering such tasks unnecessary, according to Anurag Sahay, managing director -- AI and Data Science, at Nagarro. "Coders and engineers will have to bring two important value propositions to the table - an ability for complex system engineering and an ability to leverage LLMs quickly to create digital assets," he said.
