Large language models are changing how software is developed and how the next generation of programmers is taught. This CCC and CRA-I whitepaper synthesizes roundtable discussions with experts in LLM development, software engineering, computing education, and industry deployment.
The report argues that LLMs make code reading, evaluation, debugging, testing, and decomposition more important, not less. It highlights research and education challenges around helping developers vet generated code, preparing students to use LLMs effectively, providing stable and affordable access to models, and building durable curricula in a rapidly changing technical landscape.
@misc{guha:future-programming-llms,
title = {The Future of Programming in the Age of Large Language Models},
author = {Arjun Guha and Ben Zorn},
year = {2025},
}