The history of artificial intelligence for scientific discovery is not a two year story about chatbots learning to write papers. It is a sixty year story about science repeatedly handing its bottlenecks to machines—first inference, then search, then measurement, then the full workflow—only to discover that each delegation solves one problem and exposes a harder one underneath. This paper traces that history from DENDRAL (1965) through the construction of open scholarly infrastructure (arXiv, Google Scholar, ORCID), the oracle breakthroughs of AlphaFold, and the current era of LLM driven autonomous research agents. Three interlocking threads are followed including AI as research instrument, AI for research infrastructure, and the reshaping of scholarly profiles and incentives by machine readable metrics. The central tension throughout is between automation and augmentation between building systems that replace human researchers and tools that amplify human creativity and judgement. The paper presents that the most consequential development is not any single tool but the emergence of an interconnected ecosystem where AI agents, preprint platforms, open source codebases, and citation infrastructure form a feedback loop that is fundamentally restructuring who can do science, how fast discoveries propagate, and what counts as a valid scientific contribution.