The paper analyzed big data from three major preprint repositories and found that the use of large language models (LLMs) accelerates manuscript production.LLMs reduce barriers for non-native English speakers.3 These models support diversification of discovery of prior literature.[3] The production of scholarly work is increasing rapidly, but traditional signals of quality, such as the complexity of language, are becoming unreliable.3 The authors emphasize that advances in AI will require changes in assumptions about research quality, scholarly communication, and intellectual work.3 Science policymakers need to adapt institutions to the rapidly changing process of scholarly production.[3