iStock; Rebecca Zisser/BI
Software engineers know that the days of simply being able to code are behind them.
After all, artificial intelligence taking over jobs is not new — numerous tech giants over the past two years announced that they were axing jobs to focus more on artificial intelligence development.
Last year, Google slashed 12,000 jobs — 6% of its then workforce — as the company began shifting its priorities toward AI.
In May 2023, IBM announced that it would pause hiring for roles that could be replaced by AI. The affected roles, which amounted to 26,000 jobs, included those in human resources and other non-consumer-facing departments.
But AI has had a direct impact on software engineers, too.
Although AI has helped to increase their productivity by helping to debug and generate code, software engineers that Business Insider spoke to said that they expect the industry to shrink simply because it doesn't need as many workers as they do today.
This, they said, means they have to work smarter if they want to keep their jobs.
Arpita Kaushik, a data scientist at Rocket Mortgage, told BI that AI tools like ChatGPT and Github Copilot do a lot of the "heavy lifting" for workers like her.
This meant that companies could get by with leaner teams without compromising on output, she said.
"I think there will be a change in the way a software engineer works, and instead of five software engineers you might need four in future," she said.
Kaushik's point was echoed by Aditya Swami, who heads product development efforts at Hatcher+, a venture capital firm in Singapore. In fact, Swami said that shrinking headcounts will be even more apparent in bigger companies.
"That's going to happen," he said. "As an industry, it's going to shrink, and only people who really understand what they're doing are going to survive."
Aditya Swami
And the problem isn't just about there being fewer jobs available. Kaushik said the rise of AI means that the bar for hiring is much higher now for interns and entry-level software engineers.
"Those people have to stretch their limits; they have to do a little bit of senior-level roles even at the starting of their careers," she added. "Because the basic things they could do, now they could do with AI."
However, software engineers who think they can rely on AI to solve their problems may be disappointed to learn that one's technical abilities still matter.
Swami from Hatcher+ said that he still expects his engineers to have strong coding skills, be able to understand code and explain what they are doing.
He said that people who studied software engineering before the rise of AI tools have a stronger foundation because they have "written code from scratch."
"The biggest problem right now is that the AI large language model hallucinates. It's not correct 100%, only 50 to 60% of the time," he said. "You still need to be very careful."
"When it comes to an interview, you should be able to explain line by line what the code is doing. That's still very important, that hasn't changed," he said.
Adrian Goh, cofounder of NodeFlair, a job board for tech professionals in Asia, told BI that one's technical foundations have only become more important in the age of AI.
"Companies value engineers with strong fundamentals who are dedicated to continuous learning and self-upgrading. These qualities ensure adaptability to evolving roles and technologies," Goh said.
Academics BI spoke to also said the rise of AI has not diminished the value of a computer science degree and the technical skills that come with earning it
"Computer science is not just about AI," said Kan Min Yen, an associate professor of computer science at the National University of Singapore.
"Proper development and use of AI still require fundamental knowledge of software engineering, data management, and security, all tenets of a holistic computing education," he added.
While some may think that tools like GitHub Copilot will render them obsolete, Kan said such concerns may be overblown.
Software professionals, he said, could still add value by understanding the limitations of AI tools and how to troubleshoot them.
"You would go to a mechanic to fix a car and to someone who knows the math and training methods for large language modules to fix and diagnose problems with LLMs, such as ChatGPT," Kan explained.
Bo An, who heads the AI division at Singapore's Nanyang Technological University (NTU), said that "the core value of a CS degree remains robust," and advancements in AI do not erode the value of a CS degree.
An added that foundational knowledge in "algorithms, data structures, software engineering, and other critical areas" is essential for advancing AI technologies.
That said, software engineers looking to score a job in software engineering must strike a tricky balance between integrating AI into their workflows and honing their fundamentals.
David Malan, a computer science professor at Harvard, told BI that AI tools could be a plus for aspiring software engineers who know how to leverage them.
"In the near term, AI is posed to amplify the impact of individual software engineers," Malan said. "Consider just how many more features they can implement, how many more bugs they can fix if they have a virtual assistant by their side."
Malan also advised job seekers to continue building up their portfolio of projects if they haven't already.
Working on your own projects, he said, could come in handy during interviews when employers want to see how candidates approach problems.
"Continuing one's education beyond school, whether formally or informally, forever expanding your own knowledge and skills, can certainly help too," Malan said. "There will always be something new to learn."
Tejas Rajagopal, a recent computer science graduate from NTU, said he's been trying to expand his skillset to strengthen his chances of landing a job.
"You can't just go to classes, watch lectures, do tutorials, and expect to be useful out there," Rajagopal said. "Because if you haven't really done any projects, built anything up by yourself, your education probably won't be adequate."
