Over the past couple of years, AI has made huge progress in coding, writing, and multimodal tasks. But there are still areas where progress feels slower than many predicted. For me, long-term planning and reliably handling complex, multi-step tasks still seem inconsistent. What's one capability you expected AI to master by now, but it still struggles with? I'm interested in hearing real-world experiences from people who use AI regularly, whether for research, development, or everyday work....