Challenges to Using Large Language Models in Code Generation and Repair

Liliana Pasquale, Antonino Sabetta, Marcelo d'Amorim, Péter Hegedüs, Mehdi Tarrit Mirakhorli, Hamed Okhravi, Mathias Payer, Awais Rashid, Joanna C. S. Santos, Jonathan M. Spring, Lin Tan, Katja Tuma

Abstract: Large language models (LLMs) hold great promise in solving many challenges arising from software complexity, including the possibility of automating code generation and repair. Although we cannot deny the groundbreaking nature of LLM-based code repair, we must be realistic in positioning current results.