Developers are increasingly relying on large language models (LLMs) for everyday computing tasks such as fixing bugs, explaining code and automating text-processing tasks like filtering logs.