Quantum100 ⊗ AI Workshop

The rise of AI coding assistants and agents is transforming how scientific software is written, debugged, and maintained. In this talk, I explore how large language models have reshaped research workflows, share a personal perspective on integrating these technologies, and discuss strategies for maximizing their effectiveness. I examine what current benchmarks measure, why vibe coding is an unsustainable pitfall, and why providing sufficient context is crucial. Finally, I offer an outlook on what the next few years may hold for AI-assisted scientific software and the evolving role of physicists in this new landscape.