Using LLMs in Software Design: An Empirical Study of GitHub and A Practitioner Survey Journal: TOSEM Volume: 0 0 0 0 CCS: Software and its engineering Software development techniques

Yifei Wang, Ruiyin Li, Peng Liang, Yangxiao Cai, Zengyang Li, Mojtaba Shahin, Arif Ali Khan, Qiong Feng · 2026

Read Using LLMs in Software Design: An Empirical Study of GitHub and A Practitioner Survey Journal: TOSEM Volume: 0 0 0 0 CCS: Software and its engineering Software development techniques by Yifei Wang, Ruiyin Li, Peng Liang, Yangxiao Cai, Zengyang Li, Mojtaba Shahin, Arif Ali Khan, Qiong Feng (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-01392.txt.

Description

Yifei Wang and seven co-authors report how software developers actually use large language models while designing software, a stage long seen as pure human judgment. Hard facts on this use have been scarce, so the team mixes two sources of proof.

First, they mine 291 developer chats with ChatGPT that were shared on GitHub. From these they sort nine kinds of design task, from system architecture to data models and design patterns. They find most use is knowledge lookup and design-related code drafting at the detail level. Second, they survey 65 working developers. Those developers name seven gains, such as better tech choices and early flaw spotting, and six limits, such as long outputs, broken code, and weak grasp of project context. Together the two views ground a factual picture of current practice.

Catalog record

CREATORYifei Wang, Ruiyin Li, Peng Liang, Yangxiao Cai, Zengyang Li, Mojtaba Shahin, Arif Ali Khan, Qiong Feng
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE85 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.01392

Browse the full library · About the Tower of Babel