AI alignment remains a major unsolved challenge. It requires major technical and scientific breakthroughs and will potentially be a major enabler of progress across the field. All else being equal, we would prefer to use safer and more aligned models so alignment will likely become a key feature of advanced AI systems.