As AI coding agents become more capable, many developers fear their jobs could disappear. Andrew Ng disagrees. The Google Brain co-founder believes the future belongs to developers who know how to work with AI, not compete against it.
Google Brain co-founder Andrew Ng says software developers should not fear artificial intelligence replacing them. Instead, he believes AI will reshape software development by changing how developers build products and solve problems.
Ng shared his views in an open letter explaining “loop engineering,” a new approach that has gained attention across the AI industry. The concept shifts developers away from writing one prompt at a time. Instead, they design systems that allow AI agents to write code, test it, fix errors and repeat the process with minimal human intervention.
According to Ng, loop engineering revolves around three connected feedback loops. The first is the agentic coding loop. In this stage, a developer gives an AI agent a product specification and, in many cases, performance tests. The agent writes code, evaluates its own work, fixes bugs and keeps improving the software until it meets the required standard. The second is the developer feedback loop.
Here, the human developer reviews the AI’s work, adjusts priorities and refines the product. Ng argues that developers now spend less time finding bugs and more time making decisions about features, user experience and product direction.
The third is the external feedback loop. This stage brings in feedback from users, testers or production data. Developers use those insights to improve the product before sending updated instructions back to the AI agent. Ng believes humans still hold an important advantage.
See Also: The US Is Now Using Anthropic’s AI to Hunt Hidden Security Flaws in Government Code
He argues that developers understand users, business goals and product context far better than AI systems do. That judgment helps teams build products people actually want instead of simply generating functional code. “Many people describe this human contribution as ‘taste,’ but I prefer to think of it as humans having a context advantage,” Ng wrote.
The discussion follows growing claims that traditional coding is becoming obsolete. Recent comments from Claude Code creator Boris Cherny and other AI leaders sparked debate by suggesting developers should focus less on prompt engineering and more on designing autonomous AI workflows. Ng agrees that workflows matter, but he rejects the idea that developers themselves are becoming obsolete.
The software industry is already moving in that direction. Companies increasingly deploy AI coding agents to write code, run tests and automate repetitive tasks. Developers now supervise those systems instead of manually completing every step themselves.
Ng believes this shift will make developers more productive rather than unnecessary. He expects engineers to spend more time solving problems, understanding customer needs and shaping products while AI handles much of the repetitive coding work.
His message offers a different perspective from many predictions about AI replacing programmers. The biggest opportunity may not belong to the developers who write the fastest code. It may belong to those who learn how to build the smartest systems around AI.
As software development evolves, success will depend less on writing every line of code and more on designing the loops that keep intelligent systems improving long after the first prompt is written.
About the Author
Marcel Chidozie is a tech analyst and writer covering foreign news, fintech, and emerging technologies at TechRegard. Based in Nigeria, He's passionate about translating complex tech developments into compelling, accessible stories for diverse audiences. His work focuses on how technology shapes innovation across Africa and globally.
