Musk: For AI, Writing "Love Humanity" Is Not as Good as Writing "Pursue Truth"
English text is being prepared.
Musk believes that writing "love humanity" into AI programs is actually the most dangerous approach, because morality is fragile and good intentions easily go astray; when facing a super AI millions of times smarter than humans, writing "pursue truth" into the code is humanity's only death-exemption card. He warns that every well-intentioned lie instilled into large models today is planting a landmine for the future, and only the ultimate pursuit of the truth of the universe can earn its cautious reverence for a fragile civilization.
Body
为什么把“爱人类”写进 AI 程序反而最危险？ 马斯克谈比人类聪明百万倍超级AI的一句话点醒了我们所有人， 老马说道德是脆弱的，善意是容易跑偏的，把“追求真相”写进代码，才是人类在神级智能面前唯一的免死金牌。 我们今天给大模型灌输的每一句善意谎言，都在给未来的人类埋下一颗雷， 马斯克一语中的撕开了一个最刺骨的现实： 在比你聪明百万倍的机器面前，伪善会引来毁灭，唯有对宇宙真相的极致追求，能换来它对脆弱文明的审慎敬畏。
听完马斯克这段关于百万倍智能与求真本能的推演，你就会明白： 机器的善意从来不需要靠缰绳来乞讨，真理本身的重量，足够容纳所有的渺小。 https://t.co/amywq8e9vj