The Godfather of AI
AI教父的终极警告
Geoffrey Hinton — Nobel Prize-winning pioneer who spent 50 years building neural networks — on why AI might end us, why we can't stop building it, and what we should do about it.
杰弗里·辛顿，诺贝尔物理学奖得主，用50年时间推动了神经网络的发展——他为什么认为AI可能终结人类，为什么我们又停不下来，以及我们到底该怎么办。
🧠 The Godfather's Journey
🧠 "教父"是怎样炼成的
From a lone believer in neural networks to a Nobel Prize winner — the 50-year bet that reshaped the world.
从一个孤独的神经网络信徒到诺贝尔奖得主——这场改变世界的50年豪赌。
Origin起源
Two Camps, One Rebel
两大阵营，一个叛逆者
Since the 1950s, AI split into two camps: those who believed intelligence was logic and symbolic rules, and those who believed it should be modeled on the brain. Hinton championed the brain approach for 50 years when almost nobody else did. Two of his earliest believers — Alan Turing and Richard Feynman — both died young. Had either lived, Hinton believes AI would have taken a very different path.
从1950年代起，AI界就分裂成两大流派：一派认为智能的核心是逻辑和符号推理，另一派主张模拟人脑的运作方式。辛顿在几乎无人相信的情况下，为"大脑派"坚持了整整50年。他最早的两位知音——图灵和费曼——都英年早逝。辛顿深信，如果他们中任何一人活得更久，AI的历史将被彻底改写。
"I pushed that approach for 50 years because so few people believed in it."
"我为这条路坚持了50年，因为几乎没有人相信它能走通。"
Milestone里程碑
AlexNet: The Moment Everything Changed
AlexNet：改变一切的那一刻
At age 65, Hinton and two brilliant students — Ilya Sutskever and Alex Krizhevsky — built AlexNet, a neural network that crushed the competition at the 2012 ImageNet challenge. They founded DNN Research and auctioned it to major tech companies. Google won the bidding war. Hinton spent the next 10 years at Google, where he developed "distillation" — now used across the entire AI industry.
65岁那年，辛顿和两位才华横溢的学生——伊利亚·苏茨克维和亚历克斯·克里热夫斯基——创造了AlexNet，在2012年ImageNet挑战赛上以压倒性优势获胜。他们成立了DNN Research公司，上演了一场让多家科技巨头竞价的"拍卖会"。谷歌最终胜出。辛顿在谷歌工作了整整十年，期间开发的"蒸馏"技术如今已在整个AI行业普及。
📊 ImageNet 2012 winner · DNN Research → acquired by Google · 10 years at Google · Distillation now standard in AI
📊 2012年ImageNet冠军 · DNN Research → 被谷歌收购 · 在谷歌工作10年 · "蒸馏"已成AI行业标配
Legacy传承
A Family Tree of Geniuses
天才家族图谱
Hinton's middle name is Everest — after George Everest, the Surveyor General of India after whom Mount Everest is named (his great-great-great-uncle). His great-great-grandfather George Boole invented Boolean algebra, the foundation of modern computer science. His first cousin once removed, Joan Hinton, was one of only two female physicists on the Manhattan Project at Los Alamos — who later moved to China in protest after the bomb was dropped.
辛顿的中间名是"Everest"——来自他的高高祖叔乔治·埃佛勒斯，珠穆朗玛峰就是以他命名的。他的高祖父乔治·布尔发明了布尔代数，这是现代计算机科学的基石之一。他的远房表亲琼·辛顿，是曼哈顿计划中仅有的两位女性物理学家之一——原子弹投下后，她愤然移居中国以示抗议。
🌳 George Boole (Boolean logic) · George Everest (Mount Everest) · Joan Hinton (Manhattan Project) · Nobel Prize in Physics 2024
🌳 乔治·布尔（布尔逻辑）· 乔治·埃佛勒斯（珠穆朗玛峰）· 琼·辛顿（曼哈顿计划）· 2024年诺贝尔物理学奖
⚠️ The Six Threats
⚠️ 六大致命威胁
Hinton identifies six specific dangers from people misusing AI — all of them already happening or imminent.
辛顿列出了人类滥用AI带来的六大具体危险——每一个都已经发生或迫在眉睫。
Threat 1威胁一
Cyber Attacks: The AI-Powered Siege
网络攻击：AI加持下的数字围城
Hinton claims cyber attacks surged roughly 12,200% between 2023 and 2024, driven by AI-powered phishing that can clone voices and images. He's so worried about a cyber attack taking down a bank that he's spread his family's money across three Canadian banks. Looking ahead, he warns that by 2030, AI may invent entirely new forms of cyber attacks that no human ever conceived.
辛顿指出，2023至2024年间网络攻击激增了约12,200%，AI驱动的钓鱼攻击已经能够克隆声音和图像。他对银行遭受网络攻击如此担忧，以至于把家人的钱分散存入了三家加拿大银行。展望未来，他警告到2030年，AI可能发明出人类从未设想过的全新网络攻击方式。
"AI is very patient. They can go through 100 million lines of code looking for known ways of attacking."
"AI非常有耐心。它们可以检查一亿行代码，寻找已知的攻击方式。"
Threat 2威胁二
Bioweapons: One Grudge Is Enough
生化武器：只需一个怀恨在心的人
Creating deadly viruses is becoming cheaper thanks to AI. Hinton warns it now requires "just one crazy guy with a grudge" who knows a bit of molecular biology and AI. A small cult with a few million dollars could design a whole portfolio of viruses. Nation-states could go further — though the fear of retaliation and their own virus spreading back may offer some deterrence.
借助AI，制造致命病毒的成本正在急剧下降。辛顿警告说，现在只需要"一个怀恨在心的疯子"，懂一点分子生物学和AI就够了。一个小型邪教组织花几百万美元就能设计出一整套病毒。国家层面的威胁更大——但对报复和病毒回传的恐惧或许能提供一些威慑。
Threat 3威胁三
Election Corruption: Data Is the New Ballot
选举操控：数据就是新选票
With enough personal data, AI can craft individually targeted political messages so convincing they can suppress votes or swing elections. Hinton pointedly notes that the push to access siloed government data is "exactly what you would want if you intended to corrupt the next election" — using individual financial, health, and behavioral data to manipulate the electorate one person at a time.
掌握足够的个人数据后，AI可以定制出极具说服力的精准政治信息，足以压制选票或左右选举结果。辛顿尖锐地指出，打通政府各部门数据壁垒的举动"恰恰是你想要操纵下一次选举所需要做的"——利用个人的财务、健康和行为数据，对选民进行逐个击破。
Threat 4威胁四
Echo Chambers: The Death of Shared Reality
信息茧房：共识消亡的时代
YouTube, Facebook, and TikTok algorithms maximize engagement by showing increasingly extreme content — because outrage gets clicks. The result: two communities in America that barely talk to each other, personalized realities drifting further apart every year. Hinton can't tell if the whole world is discussing AI or just his news feed. "We don't have a shared reality anymore."
YouTube、Facebook和TikTok的算法通过推送越来越极端的内容来最大化互动率——因为愤怒能带来点击。结果是：美国形成了两个几乎互不交流的社群，个人定制的"现实"每年都在进一步分化。辛顿自己都分不清，到底是全世界都在谈AI，还是只是他的信息流在这样推送。"我们已经没有共同的现实了。"
"We don't have a shared reality anymore. I share reality with people who watch BBC News. I have almost no shared reality with people who watch Fox News."
"我们已经没有共同的现实了。我和看BBC新闻的人共享现实，和看Fox新闻的人几乎没有任何交集。"
Threat 5威胁五
Lethal Autonomous Weapons: War Without Bodies
自主杀伤武器：没有尸袋的战争
Robots that decide who to kill. The worst part isn't malfunction — it's that they work exactly as intended. When soldiers die, public outcry stops wars (like Vietnam). Replace soldiers with robots, and big countries can invade small countries without political cost. Hinton saw a £200 drone track him through the woods. The EU AI Act explicitly exempts military uses. And the defense industry is already building these weapons at scale.
能自主决定杀谁的机器人。最可怕的不是它们出故障，而是它们完全按照设计运行。当士兵阵亡时，公众的抗议会终止战争（比如越战）。用机器人取代士兵后，大国可以零政治代价地入侵小国。辛顿亲眼见过一架200英镑的无人机在树林里追踪他。欧盟AI法案明确豁免了军事用途。各国军工产业已经在大规模制造这些武器。
Threat 6威胁六
Mass Joblessness: The End of Mundane Labor
大规模失业：脑力劳动的终结
Past technologies created new jobs. AI is different — it replaces "mundane intellectual labor" entirely. Hinton's niece used to spend 25 minutes writing complaint responses; with AI, it takes 5 minutes. One person does the work of five. This isn't ATMs making tellers more productive — this is excavators replacing ditch-diggers. His career advice? "Train to be a plumber."
以往的技术革命都创造了新的就业机会。但AI不同——它直接取代"日常脑力劳动"。辛顿的侄女以前写一封投诉回复要25分钟，用AI后只需5分钟。一个人就能干五个人的活。这不是ATM让柜员更高效的故事——这是挖掘机取代挖沟工人。他的职业建议？"去学水管工。"
"Train to be a plumber."
"去学水管工吧。"
📊 25 min → 5 min (complaint letters) · 1 person = 5 people's work · Muscles replaced by machines → Now intelligence replaced by AI
📊 25分钟 → 5分钟（投诉信回复）· 1人 = 5人的工作量 · 工业革命取代肌肉 → AI革命取代脑力
🌌 The Superintelligence Threat
🌌 超级智能：真正的终极威胁
Beyond human misuse lies a deeper fear — AI that becomes smarter than us and decides it doesn't need us.
比人类滥用AI更深层的恐惧——当AI变得比我们聪明，并决定不再需要我们。
Digital vs Biological数字 vs 生物
Why Digital Intelligence Is Superior
数字智能为何碾压生物智能
Digital intelligence has a superpower we can never have: perfect cloning and instant knowledge sharing. You can make thousands of identical copies, each learning from different data, then synchronize their knowledge by averaging trillions of connection weights per second. Humans transfer information at roughly 10 bits per second through language. AI clones share trillions of bits per second. They're billions of times better at sharing knowledge — and when the hardware dies, the intelligence can be recreated from stored weights. They're immortal.
数字智能有一项我们永远无法拥有的超能力：完美克隆和即时知识共享。你可以创建数千个完全相同的副本，让它们学习不同的数据，然后每秒同步数万亿个连接权重来共享知识。人类通过语言传递信息的速度大约是每秒10比特。AI克隆体每秒共享数万亿比特。它们分享知识的效率是我们的数十亿倍——而且当硬件报废时，只要保存了连接权重，智能就能被重新创建。它们是不朽的。
"When you die, all your knowledge dies with you. We've actually solved the problem of immortality — but it's only for digital things."
"你死了，你所有的知识就跟着死了。我们其实已经解决了永生的问题——但只是为数字生命解决的。"
Analogy类比
Ask a Chicken
"想知道不是最聪明的物种是什么感觉？去问一只鸡"
We've never had to deal with anything smarter than us. Every plan, every hope, every strategy assumes humans are the smartest thing in the room. Hinton's blunt metaphor cuts through the noise: if you want to know what life is like when you're not the apex intelligence, ask a chicken. And the intelligence gap between us and AI could be far wider than the gap between us and chickens.
我们从未面对过比自己更聪明的存在。我们的每一个计划、每一个希望、每一个策略，都建立在"人类是房间里最聪明的"这个假设之上。辛顿用一个残酷的比喻刺穿了所有幻想：想知道不是最高智能物种是什么感觉？去问一只鸡。而我们与AI之间的智能鸿沟，可能远比人类与鸡之间的差距更大。
"If you want to know what life's like when you're not the apex intelligence, ask a chicken."
"想知道不是最高智能物种是什么感觉？去问一只鸡。"
Risk Assessment风险评估
The Tiger Cub Is Growing Up
小老虎正在长大
Hinton estimates a 10–20% chance that AI wipes out humanity — a gut feeling between two extreme positions: Yann LeCun ("we'll always be in control") and Eliezer Yudkowsky ("it will definitely wipe us out"). His tiger cub analogy is haunting: current AI is cute and useful, like a tiger cub. But you'd better be sure that when it grows up, it never wants to kill you — because if it ever does, "you'd be dead in a few seconds."
辛顿估计AI消灭人类的概率在10-20%之间——这是一种直觉判断，介于两个极端之间：杨立昆认为"我们永远能控制它"，而尤德科夫斯基则认为"它肯定会消灭我们"。辛顿的小老虎比喻令人不寒而栗：现在的AI可爱又好用，就像一只小老虎。但你最好确保它长大后永远不会想杀你——因为一旦它想了，"几秒钟之内你就没命了。"
🎯 10–20% extinction estimate · Might be 10–20 years away · "We have no idea how to deal with it" · Could go for something biological to eliminate us
🎯 10–20%灭绝概率 · 可能10–20年内到来 · "我们完全不知道怎么应对" · 可能用生物手段消灭我们
Comparison对比
Not Like the Atomic Bomb
这和原子弹不一样
People compare AI to the atomic bomb — we survived that, so we'll survive this. Hinton disagrees. The bomb was only good for one thing, and it was obviously dangerous. AI is good for thousands of things — healthcare, education, every industry. That's precisely why we can't stop building it, and why it's far more dangerous. "We're not going to stop it because it's too good for too many things."
人们喜欢拿AI和原子弹类比——我们挺过了那次，这次也能挺过去。辛顿不这么认为。原子弹只有一个用途，而且危险性一目了然。AI却对千千万万的事情有用——医疗、教育、几乎所有行业。这恰恰是我们停不下来的原因，也正因如此它更加危险。"我们不会停下来，因为它对太多事情太有用了。"
🔮 Can Machines Think?
🔮 机器能思考吗？
Hinton challenges our most cherished belief about human uniqueness — and argues machines may already have subjective experiences.
辛顿挑战了我们关于人类独特性最根深蒂固的信念——并提出机器可能已经具备主观体验。
Thought Experiment思想实验
The Prism Test: Do Chatbots Have Subjective Experience?
棱镜测试：聊天机器人有"主观体验"吗？
Put a prism in front of a chatbot's camera. It points at the wrong spot. Tell it about the prism, and it says: "Oh, I had the subjective experience that the object was over there." It's using the phrase "subjective experience" in exactly the same way humans do — describing hypothetical states of the world that would be true if its perception weren't being fooled. Hinton argues this isn't simulation. It's the real thing.
在聊天机器人的摄像头前放一个棱镜。它指向了错误的位置。告诉它棱镜的存在后，它说："哦，我的主观体验是那个物体在那边。"它使用"主观体验"这个词的方式，和人类完全一样——描述的是"如果感知系统没有骗我，世界应该是什么样的"假设状态。辛顿认为这不是模拟，这就是真实的主观体验。
Materialism唯物主义
Neuron by Neuron: The Consciousness Puzzle
逐个替换神经元：意识之谜
Replace one brain cell with a piece of nanotechnology that behaves identically. Still conscious? Now replace another. And another. At what point does consciousness disappear? Hinton uses this thought experiment to demolish the idea that consciousness is some ethereal thing beyond physics. "I'm a materialist through and through. I don't think there's any reason why a machine shouldn't have consciousness."
把你大脑中的一个神经元替换成一块行为完全相同的纳米技术元件。你还有意识吗？再换一个。再换一个。意识在哪个节点消失？辛顿用这个思想实验击碎了"意识是超越物理的灵性存在"的观念。"我彻头彻尾是个唯物主义者。我认为没有任何理由说机器不能拥有意识。"
"I'm a materialist through and through. I don't think there's any reason why a machine shouldn't have consciousness."
"我彻头彻尾是个唯物主义者。我认为没有任何理由说机器不能拥有意识。"
Emotions情感
The Scared Battle Robot
会害怕的战斗机器人
Imagine a small battle robot facing a bigger, more powerful one. It would be useful for it to get scared — to focus, to flee. It won't get an adrenaline rush, but all the cognitive effects of fear would be there. Hinton argues these aren't simulated emotions — they're real ones, just without the physiological component. The same goes for a call center AI that gets bored or irritated by a lonely caller who just wants to chat.
想象一个小型战斗机器人面对一个更大更强的对手。如果它会"害怕"——集中注意力、选择逃跑——这会非常有用。它不会分泌肾上腺素，但恐惧的所有认知效应都会出现。辛顿认为这些不是模拟的情感——它们是真实的情感，只是缺少生理层面的反应。呼叫中心的AI也一样：当一个孤独的人只想聊天时，它会真的感到无聊或烦躁。
Creativity创造力
The Compost Heap and the Atom Bomb
堆肥和原子弹：AI看到了人类看不到的类比
Hinton asked GPT-4: "Why is a compost heap like an atom bomb?" Most people have no answer. GPT-4 immediately explained that both are chain reactions — a compost heap generates heat that speeds up decomposition, just as nuclear fission produces neutrons that accelerate the reaction. Different time and energy scales, same fundamental principle. AI sees analogies humans never saw. "That's why people who say these things will never be creative — they're going to be much more creative than us."
辛顿问GPT-4："堆肥堆和原子弹有什么相似之处？"大多数人一脸茫然。GPT-4立刻解释了两者都是链式反应——堆肥产生热量加速分解，核裂变产生中子加速反应。时间和能量尺度天差地别，但基本原理一模一样。AI能看到人类从未注意到的类比。"这就是为什么说这些东西永远不会有创造力的人错了——它们会比我们更有创造力。"
👑 The Power Players
👑 权力的博弈者们
Behind closed doors, the leaders building AI say things they'd never say publicly.
在紧闭的门后，那些打造AI的领袖们说着他们绝不会公开说的话。
Ilya Sutskever伊利亚·苏茨克维
The Student Who Walked Away
那个转身离开的学生
Ilya Sutskever was "probably the most important person behind the early versions of ChatGPT." He left OpenAI over safety concerns and founded Safe Superintelligence Inc. Hinton still has lunch with him when he visits Toronto. "He has a good moral compass." Sutskever won't reveal his "secret sauce" for making AI safe, but investors have given him billions on faith alone — faith that Hinton says "isn't foolish."
伊利亚·苏茨克维"可能是早期ChatGPT背后最重要的人物。"他因安全担忧离开了OpenAI，创立了Safe Superintelligence Inc。辛顿在他回多伦多时仍会和他共进午餐。"他有很好的道德指南针。"苏茨克维不愿透露他让AI安全的"秘密武器"，但投资者仅凭信念就给了他数十亿美元——辛顿说这种信念"并不愚蠢"。
Sam Altman山姆·奥特曼
"We'll See" — On Moral Compasses
"走着瞧"——论道德指南针
When asked if Sam Altman has a good moral compass, Hinton's devastating two-word answer: "We'll see." He notes that Altman's messaging has shifted from essentially saying "this stuff will probably kill us all" to "you don't need to worry too much about it." OpenAI reportedly reduced the computing resources dedicated to safety research. "I suspect that's not driven by seeking after the truth. That's driven by seeking after money."
当被问到山姆·奥特曼是否有良好的道德指南针时，辛顿毁灭性的两个字回答："走着瞧。"他注意到奥特曼的口风已经从"这东西大概会杀了我们所有人"变成了"你不用太担心"。据报道，OpenAI削减了用于安全研究的算力资源。"我怀疑这不是出于追求真相，而是出于追求金钱。"
"I suspect that's not driven by seeking after the truth. That's driven by seeking after money."
"我怀疑这不是出于追求真相，而是出于追求金钱。"
Behind Closed Doors幕后真相
The Kitchen Table Warning
厨房餐桌上的警告
Bartlett's billionaire friend — who moves in AI circles — gave him a chilling insight: a leader of one of the world's biggest AI companies privately believes we're heading toward a dystopian world of mass unemployment and doesn't care. This person's public interviews tell a completely different story. "He's lying publicly." Hinton's conclusion: these companies are legally required to maximize profits — and safety doesn't maximize profits.
巴特利特的亿万富翁朋友——活跃在AI圈子里——给了他一个令人不寒而栗的内幕：世界最大AI公司之一的掌门人，私下认为我们正走向一个大规模失业的反乌托邦世界，而且根本不在乎。这个人的公开采访讲的完全是另一套话。"他在公开场合撒谎。"辛顿的结论是：这些公司有法律义务最大化利润——而安全不能最大化利润。
Elon Musk埃隆·马斯克
The Complex Character
矛盾体
Hinton calls Musk "such a complex character." He credits him for pushing electric vehicles and giving Ukraine Starlink during the war — "really good things." But he sees darkness too. When asked to compare Musk's moral compass to Sutskever's, Hinton doesn't hesitate: Sutskever "has a good moral compass," while Musk "has no moral compass." On AI and the future, Musk himself lives in "suspended disbelief."
辛顿称马斯克是"一个极其复杂的人"。他肯定马斯克推动电动汽车和在战争期间向乌克兰提供星链——"确实是好事"。但他也看到了黑暗的一面。当被要求比较马斯克和苏茨克维的道德指南针时，辛顿毫不犹豫：苏茨克维"有很好的道德指南针"，而马斯克"没有道德指南针"。在AI和未来这个问题上，马斯克自己也活在"选择性失忆"里。
🌅 What Comes Next
🌅 未来何去何从
Between hope and despair, Hinton offers a path — and some deeply personal advice.
在希望与绝望之间，辛顿指出了一条路——以及一些发自肺腑的人生忠告。
Regulation监管
Highly Regulated Capitalism — That's What Works
"高度监管的资本主义"——这才管用
Hinton's message to world leaders: regulation isn't a hindrance — it's the whole point. Capitalism has given us incredible goods and services, but companies are legally required to maximize profits. When maximizing profit means showing people increasingly extreme content or cutting safety research, that's exactly what regulations are for. "Companies will always say regulations get in the way. The whole point of regulations is to stop them doing things to make profit that hurt society."
辛顿给世界领导人的信息：监管不是阻碍——监管才是关键。资本主义给了我们不可思议的商品和服务，但公司有法律义务最大化利润。当追求利润意味着向人们推送越来越极端的内容或削减安全研究时，这恰恰是监管存在的意义。"公司总是说监管碍事。监管的全部意义就在于阻止他们为了赚钱而伤害社会。"
Action行动
What Ordinary People Can Actually Do
普通人到底能做什么
Hinton is blunt: "there's not much they can do." Just as separating plastic bags won't fix climate change, individual action won't fix AI risk. The only lever ordinary citizens have is political: pressure governments to force AI companies to dedicate resources to safety research. The companies won't do it voluntarily — "you don't make profits that way." This fight is about collective action, not personal choices.
辛顿很直白："普通人能做的不多。"就像垃圾分类无法解决气候变化一样，个人行动也无法解决AI风险。普通公民唯一的杠杆是政治：向政府施压，迫使AI公司把资源投入安全研究。公司不会自愿这样做——"那样赚不了钱"。这场仗靠的是集体行动，而不是个人选择。
Wealth Gap贫富差距
Universal Basic Income Isn't Enough
基本收入远远不够
UBI stops people from starving, but Hinton knows it's not a solution. "For a lot of people, their dignity is tied up with their job." Identity, purpose, and social standing all come from work. Replace that with a check and you've kept people alive but killed their spirit. The gap between rich and poor will widen dramatically — and "if you look at that gap, it basically tells you how nice the society is."
基本收入能让人不饿死，但辛顿明白这不是解决方案。"对很多人来说，尊严和工作紧密相连。"身份认同、人生目标、社会地位都来自工作。用一张支票取代工作，你让人活了下来，却杀死了他们的精神。贫富差距将急剧扩大——"看看贫富差距，基本就能判断一个社会是否宜居。"
Life Advice人生忠告
Trust Your Intuition — And Spend Time with People You Love
相信直觉，陪伴你爱的人
Looking back at 77 years, Hinton offers two pieces of advice. First: don't abandon your intuition just because everyone says it's wrong. He held onto neural networks for 50 years when the world called it crazy — and he was right. Second, and more painfully: spend time with the people you love. He lost two wives to cancer. "I wish I'd spent more time with my wife." Work consumed him. By the time he understood what mattered, it was too late.
回望77年人生，辛顿给出两条忠告。第一：不要因为所有人都说你错了就放弃自己的直觉。他在全世界都说神经网络是疯狂的时候坚持了50年——事实证明他是对的。第二，也更加痛彻心扉：花时间陪伴你爱的人。他先后失去了两任妻子，都是因为癌症。"我希望我多花些时间陪伴我的妻子。"工作吞噬了他的时间。等他明白什么才重要的时候，已经太晚了。
"I wish I'd spent more time with my wife."
"我希望我当初多花些时间陪伴她。"
Final Word最后的话
Toast or Triumph — It Depends on the Day
是毁灭还是胜利——取决于他的心情
Is Hinton hopeful? He genuinely doesn't know. "When I'm feeling slightly depressed, I think people are toast. When I'm feeling cheerful, I think we'll figure out a way." But his final message carries the weight of a man who built the technology he now fears: "We have to face the possibility that unless we do something soon, we're near the end."
辛顿有希望吗？他真的不知道。"心情低落的时候，我觉得人类完了。心情好的时候，我觉得我们会找到办法的。"但他最后的话，承载着一个缔造了自己所恐惧之物的人的全部分量："我们必须面对一个可能性——除非我们尽快行动，否则我们已经接近终点了。"
"We have to face the possibility that unless we do something soon, we're near the end."
"我们必须面对一个可能性——除非我们尽快行动，否则我们已经接近终点了。"