Discussions about AI replacing software engineers have been heating up lately. In his latest blog post, OpenAI CEO Sam Altman shared his thoughts on the future of AI-driven software engineering:
Let’s imagine the case of a software engineering agent, which is an agent that we expect to be particularly important. Imagine that this agent will eventually be capable of doing most things a software engineer at a top company with a few years of experience could do, for tasks up to a couple of days long. It will not have the biggest new ideas, it will require lots of human supervision and direction, and it will be great at some things but surprisingly bad at others.
In other words, AI agents could handle tasks typically performed by mid-level engineers with a few years of experience.
Mark Zuckerberg echoed this sentiment in a recent interview, revealing that Meta is developing AI capable of replacing mid-level engineers:
Probably in 2025, we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of mid-level engineer that you have at your company that can write code.
As a software engineer early in my career, these statements from industry leaders make me think hard about the future of our profession. To better understand the impact of AI on the job market, I also looked into a research paper by Anthropic that explores AI’s influence on the economy.
AI’s Impact on the Job Market
Anthropic analyzed 4 million anonymous AI conversations with Claude and mapped them to a U.S. Department of Labor job database to assess AI’s broader economic effects. The findings were eye-opening.
AI is most frequently used in coding, writing enhancement, and analytical tasks—areas where AI excels. Their research found that:
- 4% of occupations have 75% of their tasks automatable by AI.
- 36% of occupations have at least 25% of tasks that AI can handle.
This suggests that while AI will significantly impact many jobs, some roles remain difficult to automate.
The research also found that AI is particularly strong in cognitive skills like reading, writing, and critical thinking. However, it is far less effective in areas requiring physical interaction, management, and negotiation.
One chart from the report stood out: it showed the percentage of workers in different professions compared to how often AI was used in related discussions. For example:
- Software engineers make up only 3.4% of the U.S. workforce, yet account for 37.2% of AI-related discussions.
- Art and media professionals make up 1.4% of the workforce but represent 10.3% of AI-related discussions.
- Education professionals also have a high presence in AI-related conversations.
Another key metric was Job Zones, a U.S. Department of Labor classification that ranks job preparation levels from 1 to 5:
- Zone 1: Jobs requiring no formal education (e.g., dishwashers).
- Zone 5: Highly skilled jobs requiring extensive education (e.g., lawyers, biologists, doctors).
The study found that AI discussions disproportionately affected Zone 4 jobs, which require a college degree but are still highly automatable. In contrast, Zone 1 and Zone 5 jobs had the least AI impact—suggesting that both low-skill manual labor and high-expertise professions are harder to replace.
The Future of Software Engineers
After reading this report, one thing became clear: AI is already surpassing humans in knowledge and cognitive ability, making white-collar office jobs particularly vulnerable. Many Zone 4 roles, which require education but involve repetitive knowledge work, are prime candidates for AI automation. However, roles requiring deep expertise and experience will remain safe—for now.
Another concern is AI’s impact on human learning and skill development. Since AI is so good at reading, writing, and problem-solving, people may become overly dependent on it, leading to skill degradation over time.
The recent layoffs at my company reinforced another harsh reality: layoffs often come down to luck, not just effort. Looking at the bigger picture, the golden era for software engineers is fading. The days when solving Leetcode problems could land you a high-paying job at a top tech company are likely behind us.
While AI will likely increase the demand for software overall, the demand for professional software engineers might decrease. Just look at social media platforms like Xiaohongshu, where non-programmers are using AI to build full applications. AI is already replacing mid-level engineers, similar to how TurboTax reduced the need for entry-level accountants—simple tax filings are now automated, while accountants focus on more complex financial services.
What’s Next? Adapting to the AI Era
Looking at history, different professions have dominated each era:
- 1960s: Manufacturing workers—factories flourished after WWII, offering high wages.
- 1970s: Oil industry workers—oil price spikes created lucrative jobs.
- 1980s: Financial analysts and stockbrokers—Wall Street boomed, driving massive salaries.
- 1990s: IT professionals—the rise of the internet made programmers highly sought after.
- 2000s: Real estate agents and construction workers—housing market surges led to huge profits.
If history teaches us anything, industries change, and job markets shift. The best thing I can do is adapt and stay ahead of AI, rather than being replaced by it.
So what are the key takeaways?
- Learn AI – Mastering AI tools will enhance my efficiency and career prospects.
- Develop a broader skillset – The best job security is being exceptional at something AI can’t easily replace.
- Prepare for financial uncertainty – Layoffs are unpredictable, so having a financial safety net is crucial.
AI isn’t going away, and the role of software engineers is evolving. The key question is: Will you let AI replace you, or will you learn how to work with it and stay ahead?