Self-driving cars have been one of AI's most high-profile and most humbling challenges. The technology has progressed enormously - autonomous vehicles can navigate complex urban environments, handle highway driving, and respond to many traffic situations without human intervention. Companies like Waymo operate commercial robotaxi services in several US cities. But the promise of fully autonomous vehicles being commonplace has been repeatedly pushed back. The core difficulty is handling edge cases - the unusual, unexpected situations that human drivers navigate instinctively but that can confuse AI systems. A plastic bag blowing across the road, an unusual roadwork layout, or a pedestrian behaving unpredictably can create situations that fall outside the AI's training. The safety standard is also extraordinarily high: to be accepted, autonomous vehicles arguably need to be significantly safer than human drivers, not just comparable. Beyond passenger cars, autonomous technology is making faster progress in more controlled environments - mining trucks, port vehicles, highway freight, and agricultural machinery - where the conditions are more predictable and the paths more constrained. Full urban autonomy remains the hardest problem and the biggest prize.