For a long time, software engineers have been generally considered to be well-paid and easy-to-find professions. However, as AI gradually matures, this impression is gradually being broken. The New York Times pointed out in a recent article that programming has always been considered a guaranteed job in the past, but after the AI wave hit, this situation began to change.
In Taiwan, people who have studied in university are not able to find a job, and they want to change tracks in the middle of their career to a job with better pay. Some people with similar situations or thoughts will choose to take accelerated program courses to become "engineers." There are many short-term courses like this in the United States. However, some students are beginning to find that even after completing the courses and becoming engineers, it is still difficult to find a job.
After being fired from construction companies one after another, 36-year-old Florencio Rendon decided to find a more stable job, so he took a programming training course in the hope of becoming an engineer. However, after successfully graduating from the four-month course, he found that it was still difficult to find a job with his programming skills.
This is not an isolated case. Dan Pickett, the founder of Launch Academy, a computer technology cram school, decided to suspend courses indefinitely starting in May this year because the rate of students successfully finding jobs dropped sharply from 90% to 60%.
There are thousands of people like Rendon, and the challenges they face seem to reflect one thing: Programming is no longer a highly-paid job. According to statistics from Layoff.fyi, technology companies have laid off more than 550,000 people in the past three years. The technology industry is in an era of major downsizing. Coupled with the popularity of AI technology, the demand for programming jobs is no longer strong.
According to data from CompTIA, the Computing Technology Industry Association, software development-related job openings have been cut in half compared to five years ago, a significant reduction of 56%. For newcomers with no experience, the problem is even greater, with job openings reduced by as much as 67%. Learning to code no longer seems to be the road to success. "This is the worst job-hunting environment I have seen at the grassroots level in technology in 25 years," said Venky Ganesan, a partner at Menlo Ventures, a venture capital firm.
Is AI taking away software engineers' jobs?
The rise of AI technology is considered to be one of the driving forces behind the harsh job hunting environment in the technology industry. Although the outside world generally predicted in the past that transactional jobs would be the first to be replaced by AI, the current situation shows that engineers seem to be also threatened.
In 2022, Google's AI team claimed that they tested the AI model AlphaCode in a programming competition and found that the effect was comparable to that of novice engineers who had been trained for several months to a year. By this year, Google had stated at the financial report conference that more than 1 in 4 new code is now generated by AI.
AWS CEO Gallman told employees this year that most people may not be responsible for writing programs in the next 24 months, asking employees to continue to improve themselves.
Amazon AWS CEO Matt Garman also told employees this year that in the next 24 months, most developers may not be responsible for writing programs, and these basic tasks will be replaced by AI, and he hopes employees can continue to improve themselves and develop new capabilities. "This means that we must better understand customer needs and know what we want to do. Rather than actually writing programs on the seat, this will become more and more part of our job."
However, AI has begun to enter programming work. In addition to taking away job vacancies, it also represents the necessary ability to master AI technology to become an engineer. The job is no longer just about writing programs.
In fact, AI can indeed improve the work efficiency of engineers. A study co-authored by scholars from Princeton University, MIT, Microsoft Research, etc. pointed out that the use of AI tools can improve the work efficiency of developers by 26.08%, and the improvement is even greater for junior engineers.
A Stack Overflow survey found that more than 60% of developers have used AI tools in their workflow.
In an annual survey by the programming Q&A website Stack Overflow, more than 60% of developers claimed that they have begun to use AI tools in their workflow, and about 14% of developers claimed that they plan to use them in the future.
According to the Stack Overflow survey, in the next year, most developers believe that artificial intelligence tools will be more integrated in the way they record code (81%), testing code (80%), and writing code (76%).
Software engineers are still in demand, but the threshold has become higher
However, programming work will not be completely replaced by AI in a short time. MIT computer scientist Armando Solar-Lezama points out that AI still lacks many basic skills. Even advanced models like GPT-4 still truly understand the problems they solve and sometimes make outrageous mistakes.
Matt., assistant professor of technology management at the University of California, Santa Barbara. Matt Beane also mentioned that the code generated by AI is often mixed with errors that are difficult for newcomers to find. At present, experienced engineers are still better able to control this technology. And when it comes to learning to use AI technology, having a mathematics or engineering background is an advantage in itself.
The New York Times mentioned that the biggest change that AI technology has brought to the field of programming may not be that it replaces engineers, but that it sets a higher threshold for becoming a software engineer. You must not only be able to write programs, but also know how to use AI in your work. middle. A recent survey by Microsoft and LinkedIn found that 66% of business executives claimed that they would not hire employees without AI skills.
Source: New York Times, New York Post, Stack Overflow
