The field of scientific research is witnessing a significant shift towards autonomous discovery and synthesis, with a focus on developing AI systems that can automate the research workflow. Recent developments have led to the creation of AI agents that can perform tasks such as generating ideas, checking the literature, developing research plans, and writing papers. These agents are capable of combining ideas from different disciplines, leading to novel contributions to the scientific literature. The use of structured world models and modular architectures has enabled these agents to coherently pursue objectives and generate accurate and trustworthy results. Noteworthy papers include the Denario project, which presents a deep knowledge AI agent for scientific discovery, and Kosmos, an AI scientist that automates data-driven discovery. The Jr. AI Scientist system also demonstrates the potential of autonomous scientific exploration, while highlighting the importance of understanding the risks and limitations of AI-driven research.