VIBE CODING TRONG PHÁT TRIỂN PHẦN MỀM: MỘT TỔNG QUAN HỆ THỐNG VỀ TÁC ĐỘNG ĐẾN TỐC ĐỘ, CHẤT LƯỢNG VÀ NỢ KỸ THUẬT

Authors

  • Quách Thị Hạnh, Nguyễn Tố Uyên
  • Vũ Tất Điệp

DOI:

https://doi.org/10.59266/houjs.2025.726

Keywords:

vibe coding, mô hình ngôn ngữ lớn, GitHub Copilot, năng suất lập trình, chất lượng phần mềm

Abstract

Sự phát triển nhanh chóng của mô hình ngôn ngữ lớn (LLM) đã thúc đẩy sự xuất hiện của “vibe coding” - phương pháp lập trình dựa trên mô tả yêu cầu bằng ngôn ngữ tự nhiên để AI sinh mã. Bài báo này thực hiện tổng quan hệ thống (Systematic Literature Review) nhằm tổng hợp các bằng chứng thực nghiệm mới nhất về ảnh hưởng của vibe coding tới tốc độ và chất lượng phát triển phần mềm. Quy trình sàng lọc và phân tích tuân thủ hướng dẫn PRISMA 2020, với dữ liệu từ 17 nghiên cứu công bố trong giai đoạn 2022 - 2025. Kết quả cho thấy vibe coding giúp cải thiện tốc độ phát triển trung bình từ 19% đến 23%. Tuy nhiên, độ chính xác của mã sinh ra chỉ đạt 48%, và tỉ lệ lỗi hoặc lỗ hổng bảo mật trong lần sinh đầu tiên ở mức 31%, phản ánh những rủi ro đáng kể về chất lượng. Dù vậy, khả năng sửa lỗi của hệ thống có thể đạt tới 89% khi áp dụng chiến lược tương tác nhiều vòng hoặc prompt giàu ngữ cảnh. Nghiên cứu làm rõ mối quan hệ giữa tốc độ, chất lượng và bảo mật trong vibe coding, đồng thời nhấn mạnh nhu cầu xây dựng mô hình cộng tác con người - AI hiệu quả hơn nhằm giảm thiểu nợ kỹ thuật và đảm bảo độ tin cậy của phần mềm.

References

[1]. Andrej Karpathy [@karpathy]. (2025, Tháng Hai 2). There’s a new kind of coding I call “vibe coding”, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It’s possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper [Tweet]. Twitter. https://x.com/karpathy/ status/1886192184808149383

[2]. Asare, O., Nagappan, M., & Asokan, N. (2024). Is GitHub’s Copilot as Bad as Humans at Introducing Vulnerabilities in Code? (No. arXiv:2204.04741). arXiv. https://doi.org/10.48550/arXiv.2204.04741

[3]. Bakal, G., Dasdan, A., Katz, Y., Kaufman, M., & Levin, G. (2025). Experience with GitHub Copilot for Developer Productivity at Zoominfo (No. arXiv:2501.13282). arXiv. https:// doi.org/10.48550/arXiv.2501.13282

[4]. Barke, S., James, M. B., & Polikarpova, N. (2022). Grounded Copilot: How Programmers Interact with Code-Generating Models (No. arXiv:2206.15000). arXiv. https://doi. org/10.48550/arXiv.2206.15000

[5]. Bird, C., Ford, D., Zimmermann, T., Forsgren, N., Kalliamvakou, E., Lowdermilk, T., & Gazit, I. (2023). Taking Flight with Copilot. Communications of the ACM, 66(6), 56–62. https://doi.org/10.1145/3589996

[6]. Finnie-Ansley, J., Denny, P., Becker, B. A., Luxton-Reilly, A., & Prather, J. (2022). The Robots Are Coming: Exploring the Implications of OpenAI Codex on Introductory Programming. Proceedings of the 24th Australasian Computing Education Conference, 10–19. https://doi. org/10.1145/3511861.3511863

[7]. Fu, Y., Liang, P., Tahir, A., Li, Z., Shahin, M., Yu, J., & Chen, J. (2025). Security Weaknesses of Copilot-Generated Code in GitHub Projects: An Empirical Study (No. arXiv:2310.02059). arXiv. https:// doi.org/10.48550/arXiv.2310.02059

[8]. Hong, Q. N., Pluye, P., Fàbregues, S., Bartlett, G., Boardman, F., Cargo, M., Dagenais, P., Gagnon, M.-P., Griffiths, F., Nicolau, B., O’Cathain, A., Rousseau, M.-C., & Vedel, I. (2019). Improving the content validity of the mixed methods appraisal tool: A modified e-Delphi study. Journal of Clinical Epidemiology, 111, 49- 59.e1. https://doi.org/10.1016/j. jclinepi.2019.03.008

[9]. Li, Y., Choi, D., Chung, J., Kushman, N., Schrittwieser, J., Leblond, R., Eccles, T., Keeling, J., Gimeno, F., Lago, A. D., Hubert, T., Choy, P., d’Autume, C. de M., Babuschkin, I., Chen, X., Huang, P.-S., Welbl, J., Gowal, S., Cherepanov, A., … Vinyals, O. (2022). Competition-Level Code Generation with AlphaCode. Science, 378(6624), 1092–1097. https://doi. org/10.1126/science.abq1158

[10]. Liang, J. T., Yang, C., & Myers, B. A. (2024). A Large-Scale Survey on the Usability of AI Programming Assistants: Successes and Challenges. Proceedings of the IEEE/ACM 46th International Conference on Software Engineering, 1–13. https://doi. org/10.1145/3597503.3608128

[11]. Liu, Z., Tang, Y., Luo, X., Zhou, Y., & Zhang, L. F. (2024). No Need to Lift a Finger Anymore? Assessing the Quality of Code Generation by ChatGPT (No. arXiv:2308.04838). arXiv. https://doi. org/10.48550/arXiv.2308.04838

[12]. Oertel, J., Klünder, J., & Hebig, R. (2025). Don’t settle for the first! How many GitHub Copilot solutions should you check? Information and Software Technology, 183, 107737. https://doi. org/10.1016/j.infsof.2025.107737

[13]. Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71

[14]. Peng, S., Kalliamvakou, E., Cihon, P., & Demirer, M. (2023). The Impact of AI on Developer Productivity: Evidence from GitHub Copilot (No. arXiv:2302.06590). arXiv. https://doi. org/10.48550/arXiv.2302.06590

[15]. Roose, K. (2025, Tháng Hai 27). Not a Coder? With A.I., Just Having an Idea Can Be Enough. The New York Times. https://www.nytimes.com/2025/02/27/technology/personaltech/vibecoding-ai-software-programming.html

[16]. Solohubov, I., Moroz, A., Tiahunova, M. Y., Kyrychek, H. H., & Skrupsky, S. (không ngày). Accelerating software development with AI: exploring the impact of ChatGPT and GitHub Copilot.

[17]. Tabachnyk, M., & Nikolov, S. (2022, Tháng Bảy 26). ML-Enhanced Code Completion Improves Developer Productivity. https://research.google/blog/ml-enhanced-code-completion- improves-developer-productivity/

[18]. Tyndall, J. (2010). AACODS Checklist: Appraisal tool for grey literature (tr 2) [Checklist]. Flinders University.

https://fac.flinders.edu.au/dspace/api/ core/bitstreams/e94a96eb-0334-4300- 8880-c836d4d9a676/content

[19]. Vaithilingam, P., Zhang, T., & Glassman, E. L. (2022). Expectation vs. Experience: Evaluating the Usability of Code Generation Tools Powered by Large Language Models. CHI Conference on Human Factors in Computing Systems Extended Abstracts, 1–7. https://doi. org/10.1145/3491101.3519665

[20]. Yetiştiren, B., Özsoy, I., Ayerdem, M., & Tüzün, E. (2023). Evaluating the Code Quality of AI-Assisted Code Generation Tools: An Empirical Study on GitHub Copilot, Amazon CodeWhisperer, and ChatGPT (No. arXiv:2304.10778). arXiv. https://doi. org/10.48550/arXiv.2304.10778

[21]. Zhang, B., Liang, P., Zhou, X., Ahmad, A., & Waseem, M. (2023). Practices and Challenges of Using GitHub Copilot: An Empirical Study. 124–129. https://doi.org/10.18293/SEKE2023-077

[22]. Ziegler, A., Kalliamvakou, E., Li, X. A., Rice, A., Rifkin, D., Simister, S., Sittampalam, G., & Aftandilian, E. (2024). Measuring GitHub Copilot’s Impact on Productivity. Communications of the ACM, 67(3), 54–63. https://doi.org/10.1145/3633453

Loading...