Why LLMs Aren't Scientists Yet: Lessons from Four Autonomous Research Attempts
IntermediateDhruv Trehan, Paras ChopraJan 6arXiv
The authors built a simple six-agent system to see if today’s AI models could plan, run, and write a research paper mostly on their own.
#autonomous research pipeline#implementation drift#training data bias