# Building Software by Rolling the Dice: A Qualitative Study of Vibe Coding

arXiv: https://arxiv.org/abs/2512.22418 · HTML: https://mlsanity.com/paper/2512.22418/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Dec 2025.

## Built on by (most cited first, 5 of 5)

- [Cheap Code, Costly Judgment: A Case Study on Governable Agentic Software Engineering](https://mlsanity.com/paper/2607.01087/) (2026)
- [Is Vibe Coding the Future? An Empirical Assessment of LLM Generated Codes for Construction Safety](https://mlsanity.com/paper/2604.12311/) (2026)
- [Where Did the Variability Go? From Vibe Coding to Product Lines by Regeneration](https://mlsanity.com/paper/2606.19042/) (2026)
- [VibeJam: A User Study Platform for Web Development with Agents](https://mlsanity.com/paper/2608.29889/) (2026)
- [Vibe Analysis: Exploring LLM Adoption by Data Visualization Practitioners](https://mlsanity.com/paper/2609.31922/) (2026)

## Abstract (arXiv)

Large language models (LLMs) are reshaping software engineering by enabling "vibe coding," in which developers build software primarily through prompts rather than writing code. Although widely publicized as a productivity breakthrough, little is known about how practitioners actually define and engage in these practices. To shed light on this emerging phenomenon, we conducted a grounded theory study of 20 vibe-coding videos, including 7 live-streamed coding sessions (about 16 hours, 254 prompts) and 13 opinion videos (about 5 hours), supported by additional analysis of activity durations and prompt intents. Our findings reveal a spectrum of behaviors: some vibe coders rely almost entirely on AI without inspecting code, while others examine and adapt generated outputs. Across approaches, all must contend with the stochastic nature of generation, with debugging and refinement often described as "rolling the dice." Further, divergent mental models, shaped by vibe coders' expertise and reliance on AI, influence prompting strategies, evaluation practices, and levels of trust. These findings open new directions for research on the future of software engineering and point to practical opportunities for tool design and education.