物理学变天，「AI主导」论文首次登顶刊，人类科学家沦为验证者？

Core Insights - The article discusses a groundbreaking research paper by physicist Stephen Hsu from Michigan State University, which was inspired by AI, specifically GPT-5, to rethink the foundations of quantum mechanics [1][4][18] - This research marks a significant shift in the role of AI in scientific inquiry, suggesting that AI can provide core theoretical breakthroughs rather than just assist in editing or formatting [5][18] Research Overview - Hsu's paper, published in the journal Physics Letters B, explores the evolution of quantum mechanics and questions whether it is strictly linear, a fundamental aspect of standard quantum mechanics governed by the Schrödinger equation [8][9] - The research aims to examine the compatibility of nonlinear quantum evolution with relativity, using the Tomonaga-Schwinger (TS) formalism as suggested by GPT-5 [10][11] Methodology - The TS formalism allows for a more flexible approach to quantum mechanics by enabling the use of arbitrary "slices" of spacetime, which is crucial for maintaining relativistic covariance [13] - Hsu's findings indicate that introducing nonlinear or state-dependent modifications to quantum mechanics poses significant challenges, often leading to violations of relativistic principles [16][19] AI's Role in Research - The collaboration between Hsu and GPT-5 exemplifies a new paradigm in theoretical physics, where large language models (LLMs) actively contribute to generating ideas and deriving equations [18][19] - Hsu describes the workflow as a "Generate-Verify" protocol, where one model generates hypotheses and another verifies their consistency [18][19] Future Implications - The article envisions a future where human-AI collaboration becomes standard in formal sciences, potentially accelerating discoveries and enhancing understanding of fundamental laws of nature [23]