The Chaotic Experiment That Shows AI Isn’t Ready to Take Your Job
A Carnegie Mellon University experiment, "TheAgentCompany," tested AI models in a simulated office setting, revealing significant limitations. Top AI performer, Anthropic's Claude 3.5 Sonnet, completed only 24% of tasks, showcasing…