Last Updated on May 2, 2025 by factkeeps
Artificial intelligence isn’t just a supporting tool anymore—it’s now actively writing significant portions of software. Microsoft CEO Satya Nadella recently revealed that as much as 30% of Microsoft’s internal code is currently generated by AI, a figure that is only expected to grow in the near future. The comment, made during a conversation with Meta CEO Mark Zuckerberg at the company’s LlamaCon AI developer event, is a striking indicator of how the nature of coding jobs may be evolving.
“I’d say maybe 20%, 30% of the code that is inside of our repos today and some of our projects are probably all written by software,” Nadella said, referring to Microsoft’s use of AI in code generation.
While Nadella didn’t explicitly discuss job losses or layoffs, the subtext was clear: the work traditionally done by human developers is increasingly being handled by AI systems. The transition from human-led to AI-assisted or AI-led coding isn’t theoretical—it’s happening now inside one of the largest tech companies in the world.
The Shift in Software Development
Historically, software development has been viewed as a deeply creative and highly skilled profession. Engineers spend years learning programming languages, debugging systems, and building products line by line. But with AI tools like GitHub Copilot (developed by Microsoft in collaboration with OpenAI) and large language models like GPT and Gemini, much of that coding work can now be semi-automated—or even fully generated from natural language prompts.
Nadella’s remark reflects how internal development workflows are being restructured. Instead of coding everything from scratch, developers may increasingly become supervisors, prompting AI tools, reviewing output, and integrating it with broader systems.
How Much Code Can AI Really Write?
Microsoft isn’t the only company leaning heavily on AI in software development. Google CEO Sundar Pichai has stated that over 25% of new code at Google is now written by AI. At Meta, while Zuckerberg said he didn’t have exact numbers, he predicted that in the next year, “maybe half the development is going to be done by AI.”
The trend points to a significant realignment of developer responsibilities. Rather than eliminating jobs outright, AI could reshape them—turning software engineers into AI orchestrators who refine machine-generated output, rather than write code from scratch.
Job Security in the Age of Code-Writing AI
The implications for coders, especially early-career developers and contractors, are enormous. As AI’s capabilities grow, companies are beginning to reassess how many human developers they truly need. Shopify CEO Tobi Lütke, for example, has told staff that no new hires will be approved unless it’s proven AI cannot do the job. Duolingo CEO Luis von Ahn made a similar announcement, indicating the company will phase out contract roles where AI can take over.
Even McKinsey and Goldman Sachs have issued reports predicting that 30% to 50% of jobs could be automated by 2030–2045. While not all of these are in tech, software development is high on the list of roles that AI can readily augment or replicate.
What Comes Next?
It’s not all bad news for developers. As AI handles more repetitive and boilerplate tasks, human programmers may be freed to focus on higher-order thinking, product design, architecture, and systems-level problem-solving. New roles may emerge around prompt engineering, AI supervision, and human-in-the-loop workflows.
Still, the transition won’t be painless. Coders who fail to adapt to this AI-infused future risk being sidelined in favor of more AI-literate peers. The industry is entering a period where knowing how to code with AI is becoming as important as knowing how to code at all.
Nadella’s 30% figure isn’t just a milestone—it’s a warning bell for the global coding workforce. AI isn’t coming for the coding profession. It’s already here, reshaping it in real time.
With AI inputs.
