Elon Musk, Sam Altman, and the Misreading of Science Fiction

埃隆·马斯克、萨姆·奥特曼与对科幻小说的误读

Why does anyone care what Elon Musk thinks of Homer’s Odyssey, or Christopher Nolan’s? Musk is promising to use AI to generate a “historically accurate” version of the The Odyssey—turning epic poetry into atrocious slop—even as AI companies are literally destroying books to ingest them as data more cheaply. Meanwhile, Musk and other Silicon Valley billionaires are building a world they read about, as boys, in the pages of books they failed to understand. I think of that world as the Artificial State.

为什么会有人在意埃隆·马斯克对荷马史诗《奥德赛》或克里斯托弗·诺兰作品的看法？马斯克承诺利用人工智能生成一个“历史准确”版本的《奥德赛》——将史诗文学变成糟糕的垃圾——与此同时，人工智能公司却在为了更廉价地获取数据而肆意摧毁书籍。与此同时，马斯克和其他硅谷亿万富翁正在构建一个他们儿时在书中读到、却未能真正理解的世界。我将那个世界称为“人造国度”（Artificial State）。

The Artificial State is less a state than a dream of being without one, the dream of the severing of humans not only from the natural world but also from one another, every umbilical cord cut. An extreme form of this dream is the actual plan of a very small number of staggeringly powerful 21st-century merchant princes, Silicon Valley millionaires and billionaires who, seeking escape from the prison of their own bodies, intend to have their heads frozen when they die and, possibly, consciousnesses uploaded onto computers, an idea that, like much of what they want, comes from a story published nearly a century ago in a science fiction magazine for boys.

“人造国度”与其说是一个国家，不如说是一个摆脱国家束缚的梦想，一个将人类不仅与自然世界、也与彼此彻底割裂的梦想，仿佛切断了所有的脐带。这个梦想的一种极端形式，是少数几位权势滔天的21世纪“商业王子”——那些硅谷百万富翁和亿万富翁——的实际计划：他们试图逃离肉体的牢笼，打算在死后冷冻头部，甚至可能将意识上传到计算机中。这个想法和他们想要的许多东西一样，源自近一个世纪前发表在男孩科幻杂志上的故事。

The dream of life in outer space, of immortal life, and of colonizing other planets all have a place in the dream of the Artificial State. The thinking goes something like this: Democracy and the liberal nation-state have to be abolished to free corporations from the restraining force of any government so that they might marshal all the resources of this planet toward the accelerated development of the most sophisticated artificial intelligence possible, artificial intelligence that will at once make it possible to escape the prison of the earth and, perversely, make such an escape necessary.

外太空生活、永生以及殖民其他行星的梦想，都在“人造国度”的蓝图中占有一席之地。其逻辑大致如下：必须废除民主和自由民族国家，以使企业摆脱任何政府的约束，从而能够调动地球上的所有资源，加速开发最尖端的人工智能。这种人工智能既能让人类逃离地球的牢笼，又荒谬地让这种逃离变得不可或缺。

For a time, in the very recent past, some of the merchant princes of Silicon Valley paid lip service to the idea that ordinary people, or the democratically elected governments of nation-states, ought to have a role, a vote, in deciding what to do with technologies that—corporate leaders pledged—were about to usher in a new era in the history of humanity. But none of the promised democratic or constitutional structures were built. No world-governance board was established, or even seriously contemplated. No one held an AI constitutional convention or anything close to it.

在不久前的过去，一些硅谷的“商业王子”曾口头上表示，普通民众或民主选举产生的民族国家政府，应该在决定如何使用这些技术时拥有发言权或投票权——企业领袖们曾承诺，这些技术即将开启人类历史的新纪元。然而，承诺的民主或宪政结构一个也没有建立。没有任何世界治理委员会成立，甚至连认真的构想都没有。没有人召开过人工智能制宪会议，也没有任何类似的举措。

Instead, the architects of the Artificial State told astounding stories, straight from the pages of science fiction, about existential threats from artificial intelligence that could only be avoided by accelerating its development, all for the sake of driving investment in their corporations, even to the cost of catastrophe for the planet. In 2025, OpenAI’s Sam Altman said he guessed that pretty soon most of the surface of the earth would be covered in data centers, but, he mumbled, “maybe we put them in space.”

相反，“人造国度”的建筑师们编造了令人震惊的故事，直接照搬科幻小说情节，声称人工智能带来的生存威胁只有通过加速其开发才能避免。这一切都是为了推动对其企业的投资，哪怕代价是给地球带来灾难。2025年，OpenAI的萨姆·奥特曼表示，他猜测很快地球表面大部分地区都将被数据中心覆盖，但他嘟囔着说：“也许我们可以把它们放在太空里。”

Altman claimed to be a devotee of science fiction. He was particularly enchanted by a once obscure Canadian novel called We Are Legion (We Are Bob), in which war has destroyed humanity but the consciousness of dead humans are uploaded to an AI system to do a new government’s bidding. This dire prospect—the complete abandonment of what it is to be human—apparently held appeal for advocates of artificial intelligence. Altman, like Peter Thiel and Marvin Minsky before him, had signed up to have himself frozen for future reincarnation.

奥特曼自称是科幻小说的忠实拥趸。他特别痴迷于一部曾经默默无闻的加拿大小说《我们是军团》（We Are Legion (We Are Bob)），书中战争摧毁了人类，但死者的意识被上传到人工智能系统，以听从新政府的差遣。这种可怕的前景——彻底抛弃作为人的本质——显然对人工智能的倡导者们具有吸引力。奥特曼和之前的彼得·蒂尔（Peter Thiel）及马文·明斯基（Marvin Minsky）一样，已经签署了协议，准备在死后冷冻自己，以待未来转世。

Sam Altman’s interest in having his brain frozen so that in the future, after he’s dead, his consciousness can be uploaded onto a computer is a good illustration of taking science fiction rather too literally, and especially of reading cautionary tales as instruction manuals. The computer scientist Stuart Russell, for instance, once said the problem of an AI doing exactly what you tell it to do is like the problem of King Midas, but, as Ted Chiang observed, that is not actually the moral of the Midas parable, whose point is “that greed will destroy you, and that the pursuit of wealth will cost you everything that is truly important.” Russell appears to have missed that point.

萨姆·奥特曼对冷冻大脑以期死后将意识上传到计算机的兴趣，很好地说明了人们是如何过于字面地理解科幻小说，尤其是将警世寓言当作操作手册来阅读。例如，计算机科学家斯图尔特·罗素（Stuart Russell）曾说，人工智能完全按照你的指令行事的问题，就像迈达斯国王（King Midas）的问题一样。但正如特德·姜（Ted Chiang）所指出的，这并非迈达斯寓言的真正寓意，其核心在于“贪婪会毁灭你，追求财富会让你失去所有真正重要的东西”。罗素似乎错过了这一点。

Elon Musk read Asimov’s Foundation series when he was 10 years old, and he drew from its plot, he said, the lesson that “you should try to take the set of actions that are likely to prolong civilization, minimize the probability of a dark age, and reduce the length of a dark age if there is one.” (Even ChatGPT tells me that the moral of the Foundation series is that the future is shaped not by great men but by collective action, that power corrupts, and that computer predictions of human behavior are not to be relied upon and are also likely to become a form of tyranny.) Musk also missed the point when he read the book that he claimed had most influenced him: Douglas Adams’ The Hitchhiker’s Guide to the Galaxy.

埃隆·马斯克在10岁时读了阿西莫夫的《基地》系列，他说他从中汲取的教训是：“你应该尝试采取那些可能延长文明、降低黑暗时代概率，并在黑暗时代来临时缩短其持续时间的行动。”（即便是ChatGPT也告诉我，《基地》系列的寓意是：未来并非由伟人塑造，而是由集体行动塑造；权力会腐蚀人；计算机对人类行为的预测不可信，且很可能演变成一种暴政。）马斯克在阅读那本他声称对自己影响最大的书——道格拉斯·亚当斯的《银河系漫游指南》时，也同样错失了重点。

Hitchhiker’s Guide is a spoof of Asimov’s Foundation series, with its Encyclopedia Galactica. Adams, a comedy writer who had earlier written for Monty Python’s Flying Circus, loathed science fiction—he once said Asimov’s books read like “junk mail.” In the middle of the 1970s, Adams wrote The Hitchhiker’s Guide for BBC Radio 4. He then turned the scripts into a series of phenomenally best-selling books.

《银河系漫游指南》是对阿西莫夫《基地》系列及其《银河百科全书》的戏仿。亚当斯曾为《蒙提·派森的飞行马戏团》撰写剧本，他厌恶科幻小说——他曾说阿西莫夫的书读起来像“垃圾邮件”。20世纪70年代中期，亚当斯为BBC广播四台创作了《银河系漫游指南》，随后将剧本改编成了一系列极其畅销的书籍。

Musk read The Hitchhiker’s Guide to the Galaxy when he was 12 or 13. “I was reading all the time,” he later recalled. “So I was either reading, working on my computer, or reading comics, playing Dungeons & Dragons, that kind of thing.” But then, he said, he had an “existential crisis.” Life seemed meaningless. “I was reading various books on trying to figure out the meaning of life … and then we happened to have some books by Nietzsche and Schopenhauer in the house,” and those made him feel worse, and then he read Hitchhiker’s Guide, a book that is so wonderfully nutty and its structure so delightfully byzantine that it’s almost impossible to summarize; but along one branch of its plot, the people of the…

马斯克在12或13岁时读了《银河系漫游指南》。“我当时一直在阅读，”他后来回忆道，“要么是在读书、用电脑，要么是在看漫画、玩《龙与地下城》之类的东西。”但随后，他说自己经历了一场“存在主义危机”。生活似乎毫无意义。“我读了各种试图弄清生命意义的书……家里恰好有一些尼采和叔本华的书，”这些书让他感觉更糟了。后来他读了《银河系漫游指南》，这是一本如此奇妙而疯狂、结构如此令人愉悦且错综复杂的书，几乎无法总结；但在其情节的一条分支中，……的人们。