OpenAI CEO Sam Altman has unveiled a bold vision for the future of AI agents, describing them as virtual co-workers that could eventually take over many of the tasks currently handled by software engineers. In a recent blog post, Altman explained how these AI systems are being designed to complete complex engineering tasks, albeit with human oversight.
Altman provided an example of a software engineering AI agent that could perform tasks similar to those assigned to an early-career software engineer at a top company. However, he clarified that while AI could execute tasks over a few days, these models are not yet capable of generating breakthrough ideas or operating without supervision.
“In the long term, we could have thousands or even millions of AI agents deployed across different industries, redefining knowledge-based work,” Altman said. He compared the rise of AI to the impact of transistors—which once revolutionised computing and are now embedded in nearly every aspect of modern life.
Altman’s comments come as the field of Artificial General Intelligence (AGI) moves closer to reality. He envisions a world where AI-driven automation leads to increased productivity, allowing individuals to accomplish more in the next decade than even the most influential figures of today.
The push toward AI-driven workforces isn’t just limited to OpenAI. Other major tech giants, including Meta and Google, have also signalled a shift toward AI replacing mid-level software engineering roles.
Meta CEO Mark Zuckerberg recently revealed that the company plans to replace mid-level engineers with AI-powered tools in 2025. “We will get to a point where all the code in our apps and the AI it generates will also be written by AI engineers instead of human engineers,” Zuckerberg stated.
Meanwhile, Google CEO Sundar Pichai confirmed last year that 25% of all new code at Google is now AI-generated, with human engineers primarily reviewing AI-generated code instead of writing it from scratch.
Tech companies have assured that these developments will not immediately replace human jobs, but the impact of automation on employment remains uncertain. In recent years, mass layoffs in the tech industry have been justified under cost-cutting and restructuring efforts. With AI demonstrating its ability to perform highly skilled tasks, concerns about job security in software development and other fields are mounting.
Altman acknowledged that the balance of power between capital and labor is shifting. He suggested that early interventions may be necessary to prevent inequality as AI begins to reshape the job market. One potential solution, he proposed, is providing individuals with increased access to AI tools, though he admitted this idea might seem unconventional.