CZ: The incentive mechanism for AI data annotation can refer to the bell curve and give rewards based on the position of the curve

CZ published an article on the X platform today suggesting the use of blockchain to achieve AI data annotation, advocating decentralized tool development, which can make use of low-cost global labor and instant payment through cryptocurrency to break geographical restrictions. In response, George Kikvadz, founder of Bitfury, asked what incentive mechanisms can be designed to reward consistently high-quality labeling users while preventing spam/low-quality submissions? CZ responded: "The general method I learned is: label 100 or 500 people, and the results will be a bell curve. Then give rewards based on their position on the bell curve. People near the end get nothing. Robot prevention is another aspect."