AI Project‑Coordinators Run Research

AI systems are beginning to act as persistent project managers that decompose research questions, delegate to specialized subagents (literature review, coding, proof search), maintain shared workspaces and versioned artifacts, and iterate across long workflows. That combination turns transient chat models into durable research infrastructure that can carry a multi‑step scientific or engineering project to completion with little human micromanagement. — If AI can autonomously coordinate and execute research workflows, funders, universities, and regulators will need to rethink scientific credit, reproducibility, oversight, and the labor market for researchers.