The field of scientific computing is witnessing a significant shift towards autonomous systems, driven by the increasing capabilities of large language models (LLMs). Recent developments have focused on leveraging LLMs to automate various aspects of scientific computing, including code generation, data analysis, and experimental design. These advancements have the potential to accelerate scientific discovery, improve productivity, and democratize access to complex computational tools. Notably, innovative systems have been proposed to automate tasks such as OpenQASM programming, distribution grid analysis, and data-driven scientific discovery in plant science. The integration of LLMs with domain knowledge, data analysis, and machine learning has enabled the development of autonomous agents that can iteratively refine solutions, identify meaningful features, and converge on interpretable models without human intervention. The use of collaborative frameworks, such as multi-agent systems and rewriting-resolution-review-revision logical chains, has also improved the reliability and accuracy of autonomous code generation. Overall, the field is moving towards a future where autonomous systems can assist scientists in complex tasks, freeing them to focus on higher-level creative work and accelerating the pace of scientific progress. Noteworthy papers include: CelloAI, which leverages LLMs for HPC software development in high energy physics, and QAgent, which automates OpenQASM programming with a multi-agent system.