The rapid rise of artificial intelligence is beginning to reshape how software is built and it may even redefine the foundations of computer science itself. According to Aravind Srinivas, advances in AI could shift the discipline away from routine coding and back toward its deeper roots in mathematics, physics, and systems thinking.
Srinivas, the chief executive of Perplexity AI, recently endorsed a viral discussion on X suggesting that large language models (LLMs) are automating many of the repetitive programming tasks traditionally performed by software engineers. In a brief response to the post, he wrote “Well said,” signaling agreement with the broader argument that AI is transforming the nature of software engineering.
The original post argued that as AI tools increasingly handle boilerplate coding, the center of gravity in computer science may move away from writing syntax-heavy code toward deeper theoretical reasoning and system-level design. Instead of spending most of their time writing code manually, future engineers may need stronger foundations in mathematical logic, physics, and computational theory to guide and architect complex AI-driven systems.
This view reflects a growing conversation across the technology industry about how AI tools from code-generation assistants to autonomous development systems are changing the skills required to build software.
The Changing Nature of Software Engineering
Artificial intelligence models are already capable of generating code, debugging programs, and navigating large codebases. Developers increasingly rely on these systems to automate repetitive tasks, accelerate development cycles, and improve productivity.
As a result, the role of a software engineer may gradually evolve from primarily writing code to designing architectures, defining system behavior, and supervising AI-generated outputs.
Several industry leaders have predicted that this shift could arrive faster than expected. Dario Amodei, the chief executive of Anthropic, has suggested that advanced AI models may soon be capable of performing the majority of software engineering tasks.
In public discussions earlier this year, Amodei said the industry might be “six to twelve months away” from AI systems handling most of the end-to-end work currently done by developers. Inside AI companies themselves, some engineers already rely heavily on AI-generated code, editing and refining outputs rather than writing programs from scratch.
This growing reliance on automation is fueling speculation that coding itself could become a smaller part of the profession.
Diverging Views Among Tech Leaders
Not everyone agrees on what the future of software engineering will look like.
Amjad Masad, who leads the developer platform Replit, believes traditional programming roles may eventually fade as AI takes over much of the technical work. In his view, the industry may shift toward broader problem-solvers who guide AI systems rather than write detailed code themselves.
Masad has argued that companies could increasingly rely on “generalist product people” capable of identifying problems, designing solutions, and directing AI tools to implement them.
Others see AI as a powerful augmentation tool rather than a replacement for human developers.
Jensen Huang, the head of Nvidia, has repeatedly emphasized that artificial intelligence will transform jobs rather than eliminate them entirely. According to Huang, workers who learn to collaborate effectively with AI systems will have a significant advantage in the evolving job market.
His widely quoted warning reflects this perspective, people are unlikely to lose jobs directly to AI, but they may lose them to individuals who know how to use AI more effectively.
Impact on Education and Computer Science Training
If AI continues to automate large portions of coding work, it could also change how computer science is taught in universities.
For decades, programming education has focused heavily on learning languages, syntax, and coding techniques. However, if AI systems can generate much of that code automatically, educators may place greater emphasis on foundational subjects such as algorithms, mathematics, logic, and system design.
This potential shift aligns with the early origins of computer science, which emerged from mathematical theory, physics research, and electrical engineering. Before software engineering became a large-scale profession, the field was dominated by theoretical work on computation and information processing.
Under an AI-driven model of development, those foundational disciplines could regain prominence as engineers increasingly focus on conceptual design rather than manual coding.
Industry Context: AI and Tech Workforce Changes
The debate over AI’s impact on software engineering is unfolding alongside broader changes in the technology industry.
In 2026, several major technology companies including Meta, Amazon, and Block announced workforce reductions while simultaneously ramping up investments in artificial intelligence.
Many firms are integrating AI tools into development workflows to automate coding, analyze data, and streamline internal operations. These capabilities allow companies to maintain productivity with smaller engineering teams, intensifying concerns that entry-level programming roles could shrink over time.
At the same time, demand for specialists in machine learning, AI infrastructure, and advanced computing systems continues to grow.
The result is a shifting landscape for software professionals. Routine coding tasks may increasingly be handled by machines, while human engineers focus on higher-level problem solving, system architecture, and the development of new AI technologies.
What Happens Next
Whether AI ultimately reduces the number of software engineering jobs or simply transforms them remains an open question.
What appears increasingly clear, however, is that the definition of a “software engineer” may change dramatically in the coming years. As AI systems take on more coding responsibilities, the most valuable skills in the field could shift toward analytical thinking, mathematical reasoning, and the ability to design complex computational systems.
For industry leaders like Srinivas, the transformation may represent not the end of computer science as a discipline but a return to its intellectual foundations.