AI 革命：驾驭好与坏

你好 Sam，最近怎么样？

Hello Sam, what's happening?

没什么。谢谢你邀请我来，很感激。谢谢你的招待。

Not much. Thanks for coming in here, appreciate it. Thanks for having me.

那么你在 AI 方面做了什么？我的意思是，每个人都对它着迷，对目前的能力感到震惊，并想知道未来的潜力，以及这是否是一件好事。

So what have you done with AI? I mean, everyone is fascinated by it, absolutely blown away at the current capability and wondering what the potential for the future is, and whether or not that's a good thing.

我认为这会是一件伟大的事情，但不会全是好事。这就是所有复杂性所在。这不是一个简单的故事：我们做这件事，一切都会很棒。它总体上会很棒，但它将是一场技术革命，一场社会革命。而这些总是伴随着变化。即使总体上很美好，我们也会在这个过程中失去一些东西：某些工作、某些生活方式、某些生活部分会改变或消失。无论好处有多大——我相信它会非常好——我们都需要应对很多事情。这对任何人来说都是一个复杂的问题，而且围绕着它有着深刻且非常可以理解的情感。

I think it's going to be a great thing, but I think it's not going to be all a great thing. That's where all the complexity comes in for people. It's not this clean story of we're going to do this and it's all going to be great. It's going to be net great, but it's going to be like a technological revolution, a societal revolution. And those always come with change. Even if it's net wonderful, there are things we're going to lose along the way: some kinds of jobs, some parts of our way of life, some parts of the way we live are going to change or go away. No matter how tremendous the upside is—and I believe it will be tremendously good—there's a lot of stuff we got to navigate through to make sure. That's a complicated thing for anyone to wrap their heads around, and there are deep and super understandable emotions around that.

这是一个非常诚实的回答，并非一切都会好。但在这一点上，它似乎是不可避免的。

That's a very honest answer, that it's not all going to be good. But it seems inevitable at this point.

是的，这绝对是不可避免的。我对世界的看法：当你还是个学生时，你了解到这个技术革命和那个技术革命。我现在回顾过去和展望未来的观点是，这是一场漫长的技术革命。首先我们必须发展农业，这样我们才有资源和时间去研究如何制造机器。然后我们有了工业革命，这也让我们学到了很多其他科学发现。它让我们实现了计算机革命，而现在，随着我们扩展到这些庞大的系统，它让我们实现了 AI 革命。但这实际上只是人类发现科学和技术并与之共同进化的一个漫长故事。我认为这是有史以来最激动人心的故事。我认为这是我们走向富足世界的方式。尽管我们确实需要应对这些事情，并且会有这些负面影响，但如果你思考这对世界和人们生活质量意味着什么，如果我们能到达一个智能成本大幅下降、随之而来的富足大幅上升的世界，我认为我们会对能源做同样的事情。我认为这是我们所想要的一切的两个关键输入。所以如果我们能拥有丰富且廉价的能源和智能，那将极大地改善人们的生活。

Yeah, it's definitely inevitable. My view of the world: when you're a kid in school, you learn about this technological revolution and then that one. My view now, looking backward and forward, is that this is one long technological revolution. First we had to figure out agriculture so that we had the resources and time to figure out how to build machines. Then we got the Industrial Revolution, and that made us learn about a lot of other scientific discovery too. It let us do the computer revolution, and that's now letting us, as we scale up to these massive systems, do the AI revolution. But it really is just one long story of humans discovering science and technology and co-evolving with it. I think it's the most exciting story of all time. I think it's how we get to this world of abundance. Although we do have these things to navigate and there will be these downsides, if you think about what it means for the world and for people's quality of life, if we can get to a world where the cost of intelligence dramatically falls and the abundance that comes with that goes way up, I think we'll do the same thing with energy. I think those are the two key inputs to everything else we want. So if we can have abundant and cheap energy and intelligence, that will transform people's lives largely for the better.

就像如果我们现在能回到 500 年前，看看某人的生活，我们会说有一些很棒的东西，但他们没有这个，没有那个。你能相信他们没有现代医学吗？这就是人们将来回顾我们时的感觉。但在 50 年后，当你想到那些目前依赖将被 AI 取代的工作的人——无论是卡车司机、自动化工人、工厂流水线工人——可以采取什么策略来减轻这些工作被 AI 淘汰的负面影响？

In the same way that if we could go back now 500 years and look at someone's life, we'd say there are some great things but they didn't have this, they didn't have that. Can you believe they didn't have modern medicine? That's what people are going to look back at us like. But in 50 years, when you think about the people that currently really rely on jobs that AI will replace—whether it's truck drivers, automation workers, people that work in factory assembly lines—what strategies can be put in place to mitigate the negative downsides of those jobs being eliminated by AI?

我会谈一些一般性的想法，但我发现做出非常具体的预测很困难，因为技术的发展方式甚至与我自己的直觉大相径庭。我们能不能先停一下，稍微回顾一下？

I'll talk about some general thoughts, but I find making very specific predictions difficult because the way the technology goes has been so different than even my own intuitions. Can we maybe stop there and back up a little?

你最初的想法是什么？

What were your initial thoughts?

如果你 10 年前问我，我会说：首先，AI 会取代蓝领劳动——基本上它会开卡车、做工厂工作、操作重型机械。然后在那之后，它会做一些认知劳动，但不会是真正困难的事情——它不会证明新的数学定理、发现新的科学、编写代码。然后最终，也许最后，也许永远不会，因为人类创造力是这种神奇特殊的东西，最后它才会取代创意工作。这就是我过去会说的。现在，在我看来，在一段时间内，AI 更擅长完成任务而不是完成工作。它可以很好地完成这些小部分，但有时会出错；它无法保持很长的连贯性。所以人们反而能够更高效地完成现有工作，但今天你仍然需要人类。而且 B，它正朝着完全相反的方向发展：它可能先做创意工作，然后是编码之类的事情，再然后是其他类型的认知劳动，而我们离人形机器人最远。

If you had asked me 10 years ago, I would have said: first, AI is going to come for blue-collar labor—basically it's going to drive trucks and do factory work, handle heavy machinery. Then maybe after that, it'll do some kinds of cognitive labor, but not the really hard stuff—it won't be proving new mathematical theorems, discovering new science, writing code. And then eventually, maybe last of all, maybe never because human creativity is this magic special thing, last of all it'll come for the creative jobs. That's what I would have said. Now, it looks to me like, for a while, AI is much better at doing tasks than doing jobs. It can do these little pieces super well, but sometimes it goes off the rails; it can't keep very long coherence. So people are instead just able to do their existing jobs way more productively, but you really still need the human there today. And B, it's going exactly the other direction: it could do the creative work first, stuff like coding second, other kinds of cognitive labor third, and we're the furthest away from humanoid robots.

所以回到最初的问题：如果我们确实有东西完全取代了工厂工人，完全取代了卡车司机、送货司机等等，这会在我们的社会中造成巨大的真空。

So back to the initial question: if we do have something that completely eliminates factory workers, completely eliminates truck drivers, delivery drivers, things along those lines, that creates this massive vacuum in our society.

我认为我们会做一些好事，但还不够。所以我认为在某个时候，我们会做一些事情，比如全民基本收入（UBI）或其他某种非常长期的失业保险之类的东西。我们会有某种方式给人们重新分配社会中的金钱，作为人们在寻找新工作时的缓冲。我或许应该提一下：我完全不相信不会有很多新工作。我认为人类的创造力、对地位的渴望、想要以不同方式竞争、发明新事物、成为社区的一部分、感到被重视——这些不会消失。人们一直担心这个。实际情况是，我们得到更好的工具，然后我们发明新事物和更令人惊叹的事情去做。而且外面有一个巨大的宇宙——我字面意思是指太空真的很大，但也有很多我们可以做的事情。如果我们真的到达了这个智能充裕的世界，你可以想到一个新想法，它就会被创造出来，但同样，这并不能为今天失业的人提供很好的解决方案。所以说未来会有这些伟大的无限的东西，人们会问：‘我们今天在做什么？’所以，作为一个社会，我们会做像 UBI 和其他再分配方式的事情，但我不认为这能成为人们核心的……

I think there are things that we're going to do that are good to do but not sufficient. So I think at some point we will do something like a UBI or some other kind of very long-term unemployment insurance, something. We'll have some way of giving people, like redistributing money in society as a cushion for people as they figure out the new jobs. And I maybe should touch on that: I'm not a believer at all that there won't be lots of new jobs. I think human creativity, desire for status, wanting different ways to compete, invent new things, feel part of a community, feel valued—that's not going to go away. People have worried about that forever. What happens is we get better tools and we just invent new things and more amazing things to do. And there's a big universe out there—I mean that literally, in that space is really big, but also there's just so much stuff we can all do. If we do get to this world of abundant intelligence, where you can just think of a new idea and it gets created, but again, that doesn't provide great solutions to people who are losing their jobs today. So saying there's going to be this great indefinite stuff in the future, people are like, 'What are we doing today?' So we'll, as a society, do things like UBI and other ways of redistribution, but I don't think that could be at the core of what people...

我认为人们想要的是自主权、自我决定权，能够与社会其他成员一起在构建未来中发挥作用，能够表达自己并创造对自己有意义的东西。此外，很多人做着他们讨厌的工作。作为一个社会，我们总是有点困惑，是想要工作更多还是更少，但无论如何，我们都能做一些有意义的事情，并在推动未来发展中扮演自己的角色。这非常重要。我希望随着那些长途卡车驾驶工作的消失——人们一直错误预测这会发生多快，但它终将发生——我们不仅能通过每月给人们钱来解决经济问题，还能找到一种分享所有权和未来决策权的方式。我经常说，关于 AGI，每个人都意识到我们必须分享它的好处，但我们也必须分享决策权和系统本身的访问权。我更期待这样一个世界：与其给地球上的每个人八十分之一的 AGI 资金——我们也应该这样做——不如说每个人得到系统的一小部分。你可以把它卖给其他人、卖给公司、与他人合并、用于任何你想要的创造性追求，或者用来想办法创办新企业。有了这个，你就获得了关于这一切如何使用的投票权。所以 AGI 变得越好，你那小小的八十分之一所有权就越有价值。

I think what people want is agency, self-determination, the ability to play a role in architecting the future along with the rest of society, the ability to express themselves and create something meaningful to them. Also, a lot of people work jobs they hate. As a society, we're always a bit confused about whether we want to work more or less, but somehow we all get to do something meaningful and play our role in driving the future forward. That's really important. I hope that as those long-haul truck driving jobs go away—people have been wrong about predicting how fast that will happen, but it will happen—we figure out not just a way to solve the economic problem by giving people money every month, but also a way to share ownership and decision-making over the future. I often say about AGI that everyone realizes we'll have to share the benefits, but we also have to share decision-making and access to the system itself. I'd be more excited about a world where, rather than giving everyone on Earth one eight-billionth of the AGI money—which we should also do—we say you get one eight-billionth slice of the system. You can sell it to someone else, to a company, pool it with others, use it for any creative pursuit you want, or to figure out how to start a new business. With that, you get a voting right over how it's all used. So the better the AGI gets, the more your little one eight-billionth ownership is worth to you.

前几天我们在播客上开玩笑，我说我们需要一个 AI 政府，应该有一个 AI 总统，让 AI 做所有决定。一个完全无偏见、绝对理性、掌握整个人类历史积累知识的东西，包括所有心理学和心理学研究的知识。但这伴随着一大堆人们担心的陷阱和问题。

We were joking around the other day on the podcast where I was saying that what we need is an AI government, that we should have an AI president and have AI make all the decisions. Something that's completely unbiased, absolutely rational, with the accumulated knowledge of all human history at its disposal, including all knowledge of psychology and psychological studies. But that comes with a host of pitfalls and issues that people have with it.

我认为我们离一个足够强大和可靠、让我们任何一个人都想要的系统还很远。但我要告诉你，有一天我喜欢它的地方。假设这个东西被造出来了。它可以和地球上的每个人交谈，在非常深的层面上理解他们的确切偏好——他们如何看待这个问题和那个问题，如何权衡取舍，想要什么——然后理解所有这些，并集体优化人类或美国公民的集体偏好。这太棒了，只要它不被收买。我们现在的政府已经被收买了。我们确信我们的政府受到特殊利益的严重影响。如果我们能有一个人工智能政府，没有任何东西能影响它——多么迷人的想法。这是可能的，我认为这可能是获得完全客观、绝对最聪明的决策的唯一途径，几乎可以解决我们社会面临的每一个问题和困境。

I think we're still very far away from a system that is capable enough and reliable enough that any of us would want that. But I'll tell you something I love about that someday. Let's say that thing gets built. The fact that it can go around and talk to every person on Earth, understand their exact preferences at a very deep level—how they think about this issue and that one, how they balance trade-offs, what they want—and then understand all that and collectively optimize for the collective preferences of humanity or citizens of the US. That's awesome, as long as it's not co-opted. Our government currently is co-opted. We know for sure that our government is heavily influenced by special interests. If we could have an artificial intelligence government that has no influence on it, nothing has influence on it—what a fascinating idea. It's possible, and I think it might be the only way to get completely objective, the absolute most intelligent decision for virtually every problem and dilemma we face in society.

你真的愿意交出最终决策权，说‘好吧 AI，你来决定’吗？

Would you truly be comfortable handing over final decision-making and saying, 'All right AI, you got it'?

不，但我对任何人这样做都不舒服。我对《爱国者法案》感到不舒服，我对许多正在做出的决定感到不舒服。有太多明显的证据表明，决策并非为了人民的整体福祉。它们是为了那些捐款的巨型公司、军工复合体、制药复合体的利益。金钱对我们的社会和做出的选择，以及对民众的整体好坏，有着巨大的影响。

No, but I'm not comfortable doing that with anybody. I was uncomfortable with the Patriot Act, I'm uncomfortable with many decisions being made. There's so much obvious evidence that decisions are not being made in the best interest of the overall well-being of the people. They're made in the interest of whatever gigantic corporations that have donated, the military-industrial complex, the pharmaceutical industrial complex. Money has a massive influence on our society and the choices that get made, and the overall good or bad for the population.

我完全同意当前系统非常糟糕，不为人民服务，极其腐败，而且令人难以置信地被金钱操控。我认为 AI 可以在某种程度上做得更好。但这可能只是整天与系统打交道、观察它们所有失败方式的一个因素。我们还有很长的路要走。

I have no disagreement that the current system is super broken, not working for people, super corrupt, and unbelievably run by money. I think there is a way to do a better job than that with AI, in some way. But this might just be a factor of sitting with the systems all day and watching all the ways they fail. We've got a long way to go.

当你想到 AGI 和可能的未来时，你是否曾停下来推断：‘如果这个东西有了意识，并且有能力制造更好的自己，我们离真正面对一个神还有多远？’

When you think of AGI and the possible future, do you ever extrapolate and pause, saying, 'If this thing becomes sentient and has the ability to make better versions of itself, how long before we're literally dealing with a God?'

我对这个问题的看法是，AGI 曾经是一个非常二元的时刻——之前和之后。我认为我完全错了。正确的思考方式是把它看作一个智能的连续体，一条平滑的指数曲线，一直回溯到那种技术革命的平滑曲线。我们可以投入系统的算力量，关于如何使其更高效、更智能、赋予它推理能力、思考如何自我改进的科学想法——这些都会到来。但长期以来我的模型——如果你看看 AGI 思想家的世界，有两个轴很重要，特别是在安全问题上：到达第一个 AGI 里程碑的短时间线或长时间线，然后一旦我们到达那里，起飞速度——从那里到它能够快速自我改进的点，是一个缓慢还是快速的过程？我认为我们正在走向的世界，也是我认为最可控和最安全的世界，是短时间线和慢起飞象限。我认为我们将会有——曾经有很多非常聪明的人认为你刚才说的事情会在一天或三天内发生。鉴于我们现在对技术的理解，这在我看来不太可能。即使这发生在十年或三十年内，从历史的角度来看，仍然是一眨眼的功夫。要把它做对会有一些真正的挑战，我们做出的决定——安全系统和检查——世界……

The way I think about this is that AGI used to be a very binary moment—before and after. I think I was totally wrong about that. The right way to think about it is a continuum of intelligence, a smooth exponential curve back all the way to that sort of smooth curve of technological revolution. The amount of compute power we can put into the system, the scientific ideas about how to make it more efficient and smarter, to give it the ability to do reasoning, to think about how to improve itself—that will all come. But my model for a long time—if you look at the world of AGI thinkers, there are two axes that matter, particularly around safety issues: short timelines vs. long timelines to the first milestone of AGI, and then takeoff speed once we get there—from there to the point where it's capable of rapid self-improvement, is that a slow or fast process? The world I think we're heading into, and also the world that I think is the most controllable and safest, is the short timelines and slow takeoff quadrant. I think we're going to have—there were a lot of very smart people for a while who thought that the thing you just talked about happens in a day or three days. That doesn't seem likely to me given the shape of the technology as we understand it now. Even if that happens in a decade or three decades, it's still the blink of an eye from a historical perspective. There are going to be some real challenges to getting that right, and the decisions we make—the safety systems and the checks—that the world...

从安全角度为这些项目制定全球监管或规则至关重要。你可以想象很多可怕的事情发生，但我对世界在认真对待这件事上取得的进展感到乐观。这让我想起我读到的关于核武器发展时的对话。就公众意识而言，这出现得非常迅速。我不认为有人真正意识到——人们知道人工智能的概念，但没想到它会如此全面、如此迅速地实现。

It's super important to put in place how we think about global regulation or rules of the road from a safety perspective for those projects. You can imagine many things going horribly wrong, but I feel cheerful about the progress the world is making towards taking this seriously. It reminds me of what I've read about the conversations around the development of nuclear weapons. In terms of public consciousness, this has emerged very rapidly. I don't think anyone was really aware—people were aware of the concept of artificial intelligence, but they didn't think it would be implemented so comprehensively so quickly.

ChatGPT 现在是 4.5 吗？4？4 和 4.5 之间会有某种指数级的能力提升。每一步都会更好一些。人类有一种快速适应任何新技术的能力。

ChatGPT is on 4.5 now? 4? 4 and with 4.5 there'll be some sort of exponential increase in its abilities. It'll be somewhat better each step. Humans have this ability to get used to any new technology so quickly.

我认为 ChatGPT 3.5 和 4 的发布不同寻常的地方在于，人们之前并没有真正关注。这是我们部署的部分原因——我们认为让人们和机构有时间逐渐理解、反应并共同设计我们想要的社会非常重要。如果你秘密地在实验室里建造 AGI，然后一下子扔给世界，我认为那是个非常糟糕的主意。我们一直在试图与世界对话。如果你不给人们一些他们能在生活中感受和使用的东西，他们不会太当真。每个人都很忙。从技术所处的位置到公众意识之间存在巨大的落差。现在这个落差已经弥合了。我认为人们理解了。我不认为从 4 到我们完成 4.5 的跳跃会那么疯狂。我认为人们必须经历的疯狂调整已经基本发生了。大多数人已经从认为 AGI 是科幻小说且非常遥远，转变为认为它即将发生。那是一次性的重新框架。现在，每年你都会得到一部新 iPhone。15 年来它们变得好得多，但 iPhone 到 iPhone 你会觉得，好吧，好了一点。但如果你把第一部 iPhone 和 15 放在一起，那差别就大了。GPT-3.5 到 AGI 将是一个巨大的差异，但一路上它只会逐步改进。

The thing I think was unusual about the launch of ChatGPT 3.5 and then 4 was that people hadn't really been paying attention. That's part of the reason we deploy—we think it's very important that people and institutions have time to gradually understand this, react, and co-design the society we want with it. If you just build AGI in secret in a lab and then drop it on the world all at once, I think that's a really bad idea. We had been trying to talk to the world about this for a while. If you don't give people something they can feel and use in their lives, they don't quite take it seriously. Everybody's busy. There was this big overhang from where the technology was to where public consciousness was. Now that's caught up. I think people understand it. I don't expect the jump from 4 to whenever we finish 4.5 to be crazy. I think the crazy adjustment people have had to go through has mostly happened. Most people have gone from thinking AGI was science fiction and very far off to something that is going to happen. That was a one-time reframe. Now, every year you get a new iPhone. Over 15 years they've gotten dramatically better, but iPhone to iPhone you're like, okay, it's a little better. But if you hold up the first iPhone to the 15, that's a big difference. GPT-3.5 to AGI will be a big difference, but along the way it'll just get incrementally better.

你是否考虑过 Neuralink 和其他竞争技术的融合，这些技术试图在人类生物系统和技术之间建立某种连接？你想在脑子里装一个那样的东西吗？

Do you think about the convergence of things like Neuralink and other competing technologies that try to implement some sort of connection between the human biological system and technology? Do you want one of those things in your head?

我不想要，直到每个人都想要。我对此有个玩笑，但想法是，一旦它开始普及，你几乎不得不拥有，因为每个人都会拥有。关于所有融合的事情，一个难题正是你刚才说的：作为一个社会，我们是否会让一些人融合 AGI 而其他人不行？如果我们这样做，而你选择不融合，那对你意味着什么？你会受到保护吗？如何把握好那个时刻？如果我们想象到科幻的未来，有很多科幻书籍写过如何把握好那个时刻——谁先做，不想做的人怎么办，如何确保先做的人真正帮助所有人一起提升，如何确保那些只想过非常人性化生活的人能够做到？这些事情真的很难，而且老实说离我日常的问题太远了，我没有那么多时间去思考。但似乎如果我们只是逻辑地思考，那将在某个时候成为一个巨大的挑战。人们会想要截然不同的东西，但存在一个关于公平以及这对不这样做的人意味着什么的社会问题。非常复杂。

I don't, until everybody does. I have a joke about it, but the idea is that once it gets going, you kind of have to because everybody's going to have it. One of the hard questions about all the merge stuff is exactly what you just said: as a society, are we going to let some people merge with AGI and not others? If we do, and you choose not to, what does that mean for you? Will you be protected? How do you get that moment right? If we imagine all the way out to the sci-fi future, there have been a lot of sci-fi books written about how you get that moment right—who gets to do that first, what about people who don't want to, how do you make sure the people that do it first actually help lift everybody up together, how do you make sure people who want to just live their very human life get to do that? That stuff is really hard and honestly so far off from my problems of the day that I don't get to think about it as much as I'd like to. But it seems like if we just think logically, that's going to be a huge challenge at some point. People are going to want wildly divergent things, but there is a societal question about fairness and what it means for people who don't do it. Super complicated.

在神经接口方面，短期内，在我们弄清楚如何将某人的意识上传到计算机之前——如果这完全可能的话，而且有很多理由认为它不可能——我最感兴趣的是我们能在不钻开某人头骨的情况下做什么。我们能通过外部安装的设备读取多少内心独白？如果我们有一个不完美、低带宽、低精度的神经接口，人们是否仍然可以学会很好地使用它，从而对他们用新计算平台能做的事情产生强大影响？我的猜测是，我们会解决这个问题。我肯定你见过那个头戴设备的演示：有人问一个问题，他们戴上头戴设备，思考这个问题，然后他们真的用谷歌搜索这个问题并通过头脑得到答案。这就是我们一直在探索的方向。

On the neural interface side, in the short term, before we figure out how to upload someone's consciousness into a computer—if that's possible at all, and there are plenty of sides you could take on why it's not—the thing I find myself most interested in is what we can do without drilling a hole in someone's head. How much of the inner monologue can we read out with an externally mounted device? If we have an imperfect, low bandwidth, low accuracy neural interface, can people still learn how to use it really well in a way that's quite powerful for what they can now do with a new computing platform? My guess is we figure that out. I'm sure you've seen that headpiece demonstration where someone asks a question, they have the headpiece on, they think the question, and then they literally Google the question and get the answers through their head. That's the kind of direction we've been exploring.

在我看来这是第一步。那是最终沉浸式 3D 视频游戏的 Pong。你会得到这些初步步骤，它们看起来有点粗糙和缓慢。它本质上比直接问 Siri 还慢。我认为如果有人构建了一个系统，你可以思考词语——不一定只是问题，也可以是你被动的、漫无边际的内心独白，但当然也可以是问题——然后这些被输入到 GPT-5 或 6，在你的视野中显示回答的词语，那将是 Pong。那仍然是一个非常有价值的工具，而且这似乎是不可避免的。在神经接口方面还有艰苦的工作要做，但我相信它会实现。

That seems to me to be step one. That's the Pong of the eventual immersive 3D video games. You're going to get these first steps and they're going to seem sort of crude and slow. It's essentially slower than just asking Siri. I think if someone built a system where you could think words—it doesn't just have to be a question, it could be your passive rambling inner monologue, but certainly could be a question—and that was being fed into GPT-5 or 6, and in your field of vision the words in response were being displayed, that would be the Pong. That's still a very valuable tool to have, and that seems inevitable. There's hard work to get there on the neural interface side, but I believe it will happen.

我担心的是，这项技术的早期采用者将比普通人群拥有巨大的优势。

My concern is that the initial adopters of this will have such a massive advantage over the general population.

这并不让我担心，因为那就像一台更好的电脑。你并没有把你的大脑接入高风险的东西。你知道你不想用的时候怎么办吗？你摘下眼镜。这感觉没问题。这只是外部设备。

That doesn't concern me because that's just like a better computer. You're not jacking your brain into something high-risk. You know what you do when you don't want them? You take off the glasses. That feels fine. This is just the external device.

你认为我们也能通过技术与外部设备进行心灵感应或半心灵感应式的交流吗？

Do you think we'll be able to communicate with an external device as well, telepathically or semi-telepathically through technology?

是的，我也这么认为。

I do, yeah. I think so too.

我真正担心的是，一旦我们迈出使用实际神经接口的一步，当有实际手术时。

My real concern is once we take the step to use an actual neural interface, when there's an actual operation.

他们正在使用某种植入物，然后那个植入物变得越来越先进。不是 iPhone 1 了，现在是 iPhone 15。随着这些东西越来越好，我们正走在通往半机械人的道路上。我们正走在这样的路上：为什么你想做一个生物人？你真的想住在小木屋里，而你可以活在《黑客帝国》里吗？我的意思是，我们似乎正走在这条路上。我们已经在这条路上走了一点点，对吧？比如如果你拿走一个人的手机，让他们在今天的世界里运作，他们相对于其他人就处于劣势。所以那可能是我们能想象的最轻量级的融合版本。但我认为值得一说的是，如果我们回到那个关于一条指数曲线的事情，我们已经在 x 轴上稍微抬升了一点。而且是的，即使你不完全走到神经接口那一步，VR 也会变得非常好，以至于有些人就是不想摘下来，对吧？这对他们来说没问题，只要我们能解决如何思考世界中的权力平衡这个问题。

And they're using some sort of an implant and then that implant becomes more sophisticated. It's not the iPhone 1 now, it's the iPhone 15. As these things get better and better, we're on the road to cyborgs. We're on the road to like, why would you want to be a biological person? Do you really want to live in a log cabin when you can be in The Matrix? I mean, it seems like we're on this path. We're already a little bit down that path, right? Like if you take away someone's phone and they have to go function in the world today, they're at a disadvantage relative to everybody else. So that's like the lightest weight version of a merge we could imagine. But I think it's worth, if we go back to that thing about the one exponential curve, I think it's worth saying we've lifted off the x-axis already down this path the tiniest bit. And yeah, even if you don't go all the way to a neural interface, VR will get so good that some people just don't want to take it off that much, right? And that's fine for them, as long as we can solve this question of how do we think about what a balance of power means in a world.

我认为会有很多人。我肯定是其中之一，觉得实际上人体和人类体验非常棒。树林里的小木屋很 awesome。我不想一直待在那里。我很想去玩那个很棒的视频游戏，但我也很高兴能偶尔去那里。

I think there will be many people. I'm certainly one of them who's like, actually the human body and the human experience is pretty great. That log cabin in the woods is pretty awesome. I don't want to be there all the time. I'd love to go play the great video game, but I'm really happy to get to go there sometimes.

是的，仍然有一些很棒的人类体验。人类体验，比如和朋友一起大笑，你知道，亲吻一个你从未吻过的人，第一次约会。那些才是真正的时刻。只是和朋友一起喝杯酒，大笑。在 VR 里不太一样。

Yeah, there's still human experiences that are just great. Human experience, just laughing with friends, you know, kissing someone that you've never kissed before, that you're on a first date. Those kinds of things are the real moments. Just laughing, having a glass of wine with a friend and just laughing. Not quite the same in VR.

现在还不是。当 VR 变得超级先进，以至于你无法分辨，你知道，就像你被接入大脑，分不清什么是真实什么是虚假。然后每个人都深入研究模拟假说或类似东方宗教之类的东西。我不知道那时会发生什么。

Not now. When VR goes super far, so you can't, you know, it's like you are jacked into your brain and you can't tell what's real and what's not. And then everybody gets super deep on the simulation hypothesis or the like, Eastern religion or whatever. And I don't know what happens at that point.

你有没有纠结过模拟理论？因为真正的问题是当你把它和概率论结合起来，和那些说“如果你只看数字，我们已经处于模拟中的概率远高于我们不在模拟中的概率”的人交谈时。我一直不清楚该怎么做。就像，好吧，这在理智上很有道理。

Do you ever wrestle with simulation theory? Because the real problem is when you combine that with probability theory and you talk to the people that say, well, if you just look at the numbers, the probability that we're already in a simulation is much higher than the probability that we're not. It's never been clear to me what to do about it. It's like, okay, that intellectually makes a lot of sense.

是的，我想可能吧，当然，那似乎很有说服力。但这是我的现实，这是我的生活，我要去活它。而且，你知道，从大学新生宿舍走廊凌晨两点到现在，我在这方面没有取得更多进展。

Yeah, I think probably, sure, that seems convincing. But this is my reality, this is my life, and I'm gonna live it. And I've, you know, from 2 AM in my college freshman dorm hallway till now, I've made no more progress on it than that.

嗯，这似乎是那种，你知道，如果它是一种可能性，如果它是真实的，首先，一旦它发生了，你会怎么做？我的意思是，那就是新的现实。而且在很多方面，我们的新现实对于 15000 年前的狩猎采集者来说，就像那个新现实对我们现在一样陌生。我的意思是，我们已经进入了一些非常奇怪的领域，你知道，我刚刚和我的孩子们聊天。我们问一些关于某件事的问题，我总是说，让我们猜猜那里面有多少比例是这个，然后我们就谷歌一下，然后问 Siri，我们就查到了。就像，看那个。仅此一项，与我小时候相比就是巫术，那时你必须去图书馆，希望那本书是准确的。

Well, it seems like one of those, there's no, you know, if it is a possibility, if it is real, first of all, once it happens, what are you going to do? I mean, that is the new reality. And in many ways, our new reality is as alien to hunter-gatherers from 15,000 years ago as that would be to us now. I mean, we've already entered into some very bizarre territory where, you know, I was just having a conversation with my kids. We're asking questions about something and I always say, let's guess what percentage of that is this, and then we just Google it and then just ask Siri and we pull it up. Like, look at that. That alone is sorcery compared to how it was when I was a kid, and you had to go to the library and hope that the book was accurate.

完全同意。我很恼火。我读到关于像 ChatGPT 和 Google 这样的系统对环境的影响有多可怕，因为每次查询只消耗极少的能量，你知道我们都在毁灭世界。我当时想，在那之前，人们开车去图书馆。让我们谈谈他们为了回答这个问题燃烧了多少碳，而现在需要多少。拜托。但那只是人们在找一些理由说某件事不好。那不是逻辑的视角。

Totally. I was very annoyed. I was reading about how horrible systems like ChatGPT and Google are from an environmental impact because it's using some extremely tiny amount of energy for each query, and you know how we're all destroying the world. And I was like, before that, people drove to the library. Let's talk about how much carbon they burn to answer this question versus what it takes now. Come on. But that's just people looking for some reason why something's bad. That's not a logical perspective.

我们应该关注的是通过这个可能实现的惊人变化，以及所有的问题，我们在资源、环境、清理海洋、气候变化等方面面临的不可逾越的问题。有这么多问题。我们需要这个来解决所有其他问题。这就是为什么我们需要总统 AI。如果 AI 能做出每一个科学发现，但我们仍然有人类总统，你认为我们会好吗？

What we should be looking at is the spectacular changes that are possible through this and all the problems, the insurmountable problems that we have with resources, with the environment, with cleaning up the ocean, climate change. There's so many problems. We need this to solve all of everything else. And that's why we need president AI. If AI could make every scientific discovery but we still had human presidents, do you think we'd be okay?

不会，因为那些混蛋仍然会中饱私囊，他们有离岸账户，腐败以及如何减轻腐败将永远是一个奇怪的问题。这也是当前技术状态的一个迷人之处，我们对腐败的认识更加深刻。我们有更多的独立报道，我们对实际存在的问题更加清晰。这真的很棒。

No, because those creeps would still be pocketing money and they'd have offshore accounts and it would always be a weird thing of corruption and how to mitigate that corruption. Which is also one of the fascinating things about the current state of technology, is that we're so much more aware of corruption. We're so much more, there's so much independent reporting, and we're so much more cogent of the actual problems that are in place. This is really great.

我观察到的一件事，显然很多其他人也观察到，腐败是社会中完成任何事情、取得进展的巨大障碍。你知道，我年轻时世界观更以美国为中心，随着我对世界的研究更多，必须在世界上更多地方工作，令人惊讶的是腐败仍然如此之多。但向技术赋能世界的转变，我认为是对抗腐败的主要力量，因为一切都更难隐藏。而且我确实认为世界上的腐败会因其曝光而持续下降。

One of the things that I've observed, obviously many other people too, is corruption is such an incredible hindrance to getting anything done in a society, to make it forward progress. And you know, my worldview had been more US-centric when I was younger, and as I've just studied the world more and had to work in more places in the world, it's amazing how much corruption there still is. But the shift to a technologically enabled world, I think, is a major force against it, because everything is, it's harder to hide stuff. And I do think corruption in the world will keep trending down because of its exposure.

是的，通过技术。我的意思是，这是有代价的，而且我认为损失是，我非常担心监控国家能走多远。但在一个支付不再是成袋现金而是以某种方式数字化的世界里，有人，即使你使用比特币，也能观察那些流动，我认为这是减少腐败的事情。

Yeah, through technology. I mean, it comes at a cost, and I think the loss that, like, I am very worried about how far the surveillance state could go here. But in a world where payments, for example, are no longer like bags of cash but done somehow digitally, and somebody, even if you're using Bitcoin, can like watch those flows, I think that's a corruption-reducing thing.

我同意。但我非常担心央行数字货币，以及它与社交信用评分挂钩。超级反对。是的，那把我吓坏了。超级反对。而推动这个不是为了社会的整体利益，而是为了控制。

I agree. But I'm very worried about central bank digital currency and that being tied to a social credit score. Super against. Yeah, that scares the hell out of me. Super against. And the push to that is not for the overall good of society, that's for control.

是的，我认为，我的意思是，有很多事情让我对美国政府的近期作为感到失望，但针对加密货币的战争，我认为那是，我们不能放弃这个，我们要控制这个，诸如此类，那让我对这个国家感到非常难过。

Yeah, I think, I mean, there's many things that I'm disappointed that the US government has done recently, but the war on crypto, which I think is a, like, we can't give this up, like we're going to control this, and all like that, that's a thing that makes me quite sad about the country.

这也让我对这个国家感到非常难过。但你也看到像 FTX 这样的事情，哦，没有监管和没有监督，这可能会变得非常糟糕。

It makes me quite sad about the country too. But then you also see with things like FTX, like, oh, this can get really bad without regulation and without someone overseeing it.

是的，我不反监管。我认为它显然有作用。而且我也认为 FTX 是一种滑稽的糟糕情况，我们不应该从中吸取太多教训。

Yeah, I'm not anti-regulation. I think there's clearly a role for it. And I also think FTX was a sort of comically bad situation that we shouldn't learn much from.

是的，但这是一个有趣的故事。它完全有趣，我喜欢那个故事。

Yeah, but it's a fun one. It's totally fun, and I love that story.

我真的很喜欢他们都在吸毒和互相发生关系这件事。这是一个引人入胜的故事，因为它包罗万象。他们用 QuickBooks 处理税务，同时管理着数十亿美元。我不知道为什么我觉得“polycule”这个词这么好笑，但那就是他们对自己关系的称呼——一个多角恋分子组合。所以有九个人，他们是多角恋。这在硅谷流行过一阵子，我觉得很搞笑。你显然希望有足够的监管来防止那种事发生，但我不反对那种事发生；我反对他们损失所有存款人的钱。我的意思是，polycule 有点像，随他们去吧。但你希望有足够的监管，让 FTX 不能损失所有存款人的钱。这里有一个重要的点：人们谈论的所有其他监管并没有保护我们的安全。基本的事情——比如，我们做那个——并不是关于人们讨论的所有加密货币的东西。

I really love the fact that they were all doing drugs and having sex with each other. It's a gripping story because it had everything. They did their taxes with QuickBooks while dealing with billions of dollars. I don't know why I find the word 'polycule' so funny, but that's what they called their relationship—a polyamorous molecule put together. So there were nine of them, and they were polyamorous. That became a meme in Silicon Valley for a while, which I thought was hilarious. You clearly want enough regulation so that can't happen, but I'm not against that happening; I'm against them losing all the depositor money. I mean, polycule is kind of like, go for it. But you want enough regulation so that FTX can't lose all its depositor money. There's an important point here: all this other regulation people talk about didn't keep us safe. The basic thing—like, let's do that—wasn't about all the crypto stuff people were discussing.

对我来说，真正迷人的加密货币是比特币。我认为它最有可能成为一种通用的可行货币。它供应有限，人们用自己的资源挖矿。这非常迷人，我喜欢它已经被实现。我邀请过 Andreas Antonopoulos 上播客，他完全生活在其中——他所有的钱都花在比特币上，用比特币付房租。

The real fascinating crypto is Bitcoin to me. That's the one I think has the most likely possibility of becoming a universal viable currency. It's limited in supply, people mine it with their own resources. That's very fascinating, and I love that it's been implemented. I had Andreas Antonopoulos on the podcast, and he lives it—he spends all his money in Bitcoin, pays his rent in Bitcoin.

几年前我帮助启动了一个叫 Worldcoin 的项目，所以我对这个领域有了更多了解。我对比特币兴奋的原因同样让我对它兴奋。一个不受任何政府控制的全球货币的想法，在科技树上是一个非常合乎逻辑且重要的一步。

I helped start a project called Worldcoin a few years ago, so I've learned more about the space. I'm excited about it for the same reasons I'm excited about Bitcoin. The idea of a global currency outside the control of any government is a super logical and important step on the tech tree.

同意。为什么政府要控制货币？政府应该处理紧迫的环境、社会、基础设施、外交政策和经济问题——那些我们需要被治理才能拥有和平繁荣社会的事情。

Agreed. Why should the government control currency? The government should deal with pressing environmental, social, infrastructure, foreign policy, and economic issues—the things we need to be governed for a peaceful and prosperous society.

你认为 AGI 之后货币会发生什么？我一直在想这个问题。随着货币数字化，瓶颈在于访问权限。如果我们达到一个所有信息都自由共享的地步——没有秘密，没有界限，没有国界，我们能读心，完全访问所有信息——那货币是什么？货币是这种数字化的东西。如果没有任何信息访问的障碍，你怎么能拥有它？

What do you think happens to money and currency after AGI? I've wondered about that. With money going digital, the bottleneck is access. If we get to a point where all information is freely shared everywhere—no secrets, no boundaries, no borders, we're reading minds, complete access to all information—what is money then? Money is this digital thing. How can you possess it if there's no barrier to accessing any information?

另一种思考方式：货币是交易劳动或有限硬资产（如土地和房屋）的方式。如果智力劳动随时可得且超级便宜，那就会非常不同。我认为总会有我们想要稀缺和昂贵的商品和服务。在那个世界里的货币是一个奇怪的想法。它会变得不同——不再是皮袋里的一袋金子。但问题是如何分配这些货币，以及如何避免一个极权政府发放货币的可怕马克思主义社会。我目前最好的想法：如果 AGI 系统成为影响力的高位比特，你需要重新分配的不只是货币，还有访问权限，这样人们可以自己决定如何使用和治理它。如果你有一个想法，你可以做这个；如果我有另一个想法，我可以做那个。我有权对我的部分做任何我想做的事，如果我想出更好的主意，我会得到奖励。

Another way to think about it: money is a way to trade labor or limited hard assets like land and houses. If intellectual labor is readily available and super cheap, that's very different. I think there will always be goods and services we want to be scarce and expensive. Money in that world is a curious idea. It becomes different—it's not a bag of gold in a leather pouch. But then the question is how that money is distributed and how we avoid a horrible Marxist society with a totalitarian government handing it out. My current best idea: if AGI systems become the high-order bits of influence, you need to redistribute not just the money but the access, so people can make their own decisions about how to use and govern it. If you have one idea, you get to do this; if I have another, I get to do that. I have rights to do whatever I want with my part, and if I come up with better ideas, I get rewarded.

反对福利和全民基本收入的强硬派担心的是人性。他们认为如果你取消激励，直接给人免费的钱，他们会变得上瘾和懒惰。但这难道不是一个人类生物学和心理学的问题吗？也许通过人工智能和某种神经接口，我们可以重新设计人类生物系统，消除嫉妒、欲望、羡慕等问题——所有那些与金钱和地位相关的变量。如果这些被消除，我们就成为人类物种的下一个版本。我们早已远离部落式的野蛮洞穴人社会；我们都同意这是一种更好的生活方式。昨晚我在我的喜剧俱乐部谈到这个。我妻子说每个人都来自洞穴人，这是一个发人深省的想法——所有这些仍然存在于我们的 DNA 中。那些奖励系统可以通过直接给人钱来劫持，这样他们就不必工作或有野心。这就是人们的恐惧。

The hardliners against welfare and universal basic income are concerned with human nature. They believe that if you remove incentives and just give people free money, they become addicted and lazy. But isn't that a human biological and psychological problem? Perhaps with AI and some neural interface, we can re-engineer the human biological system and remove problems like jealousy, lust, envy—all those variables that come into play with money and status. If those are eliminated, we become a next version of the human species. We're very far removed from tribal brutal societies of cave people; we all agree this is a better way to live. I was talking about this at my comedy club last night. My wife said everyone came from cave people, which is a sobering thought—that all that is still in our DNA. Those reward systems can be hijacked by just giving people money so they don't have to work or have ambition. That's the fear people have.

给人免费的钱，但如果我们能真正从工程上改造人类生物载体，消除所有那些陷阱，如果我们能在技术上启迪人类——也许‘启迪’这个词不对——但让人类物种进步到那些不再是困境的地步，因为那些问题通过编码就能轻松解决，通过增强人类生物系统就能轻松解决，比如提高多巴胺水平，让愤怒、恐惧和仇恨变得不可能，它们不存在。如果让每个人都嗑 Molly，那会有多少战争？零场战争。我的意思是，如果你能让地球上每个人在同一天都嗑一次 Molly，那会是件了不起的事，但如果让地球上每个人每天都嗑 Molly，那会是真正的损失。但如果他们服用低剂量的 Molly，让你达到那种状态，每个人都用爱和情感问候他人，不再关心竞争，而是关心创新和创造的魅力，那会怎样？天哪，我们可以用剩下的时间只聊这一个话题，太有趣了。

Giving people free money but if we can figure out how to literally engineer near the human biological vehicle and remove all those pitfalls if we can enlighten people technologically maybe enlighten is the wrong word but advance the human species to the point where those are no longer dilemmas because those are easily solvable through coding, they're easily solvable through enhancing the human biological system perhaps raising dopamine levels to the point where anger and fear and hate are impossible they don't exist. And if you just had everyone on Molly, how many wars would there be? There would be zero wars. I mean, I think if you could get everyone on Earth to all dally once on the same day, that'd be a tremendous thing, but if you got everybody on Earth to dally every day, that'd be a real loss. But what if they did a low dose of Molly where you just get to ah, where everybody greets people with love and affection and there's no longer concern about competition, instead the concern is about the fascination of innovation and creation, creativity. Man, we could talk the rest of the time about this one topic, it's so interesting.

我觉得如果我能按下一个按钮消除所有人类的奋斗和冲突，我不会这么做。首先，我认为这是我们故事和经历中非常重要的一部分。而且，从我们自身的生物历史以及我们对 AI 的了解来看，非常简单的目标系统——适应度函数、奖励模型，随便你怎么叫——能带来令人难以置信的成果。你知道，如果生物本能是生存和繁衍，看看它已经把我们社会带到了多远。我们拥有的所有这些，所有技术、建筑等等，都是通过一个极其简单的目标在极其复杂的环境中实现的，导致了人们在一定程度上实现这种生物本能并想互相打动的丰富性和复杂性。所以我认为进化适应度是一个简单而无比强大的概念。现在，你能仔细地编辑掉它的每一个个体表现吗？也许可以。但我也不想生活在一个每个人都整天嗑 Molly 的无人机社会里。那似乎不是正确答案。我希望我们继续奋斗，继续推进前沿，走出去探索。我实际上认为社会在这方面已经有点偏离轨道了。我们——我不知道，我觉得我——我以为我会在更老的时候才觉得自己是抱怨年轻人的老家伙，但我认为我们已经失去了一些东西。我认为我们需要更多的奋斗，也许更多的冒险，更多的探索精神。

I think if I could push a button to remove all human striving and conflict, I wouldn't do it. First of all, I think that's a very important part of our story and experience. And also, I think we can see both from our own biological history and from what we know about AI that very simple goal systems—fitness functions, reward models, whatever you want to call it—lead to incredibly impressive results. You know, if the biological imperative is survive and reproduce, look how far that has somehow gotten us as a society. All this stuff we have, all this technology, this building, whatever else, that got here through an extremely simple goal in a very complex environment, leading to all of the richness and complexity of people fulfilling this biological imperative to some degree and wanting to impress each other. So I think evolutionary fitness is a simple and unbelievably powerful idea. Now, could you carefully edit out every individual manifestation of that? Maybe. But I don't want to live in a society of drones where everybody is just sort of on Molly all the time either. That doesn't seem like the right answer. I want us to continue to strive, I want us to continue to push back the frontier and go out and explore. And I actually think something's already gotten a little off track in society about all of that. And we're—I don't know, I think I'm—I thought I'd be older by the time I felt like the old guy complaining about the youth, but I think we've lost something. And I think that we need more striving, maybe more risk-taking, more explorer spirit.

你说我们失去了一些东西，是什么意思？

What do you mean by you think we've lost something?

我的意思是，从我的视角来看有一个版本。我长期做创业投资者，过去最优秀的初创公司创始人通常在 20 岁出头、20 多岁中期或甚至 20 多岁后期。现在他们年龄偏大很多。我想知道的是：在当今世界，那些非常出色的 25 岁创始人在哪里？很少——说没有是不公平的，但比以前少了。我认为这对社会各个层面都不好。我以科技公司创始人为例，但像那些去创造新事物、推动有争议或不受欢迎想法的人——我们需要他们来推动进步。我们需要那种精神。我们需要人们能够提出想法，犯错，而不会因此被社会排斥，或者被取消等等。我们需要人们能够在职业生涯中冒险，因为他们相信某个重要的科学探索，这个探索可能不会成功，或者听起来非常有争议或糟糕。当然，当我们创办 OpenAI 时，我们说我们认为 AGI 是真实的，可能实现——不太可能，但如果发生就非常重要——而我们领域的所有老科学家都说那些人是不负责任的，你不应该谈论 AGI，他们在卖骗局，他们鲁莽行事，会导致 AGI 寒冬。我们说了我们相信，我们知道这不太可能，但这是一次重要的探索，我们要去追求它。而反对者——这对社会很重要。

I mean, here's a version of it very much from my own lens. I was a startup investor for a long time, and it often was the case that the very best startup founders were in their early or mid-20s or late 20s maybe even. And now they skew much older. And what I want to know is: in the world today, where are the super great 25-year-old founders? And there are few—it's not fair to say there are none, but there are less than there were before. And I think that's bad for society at all levels. I'm using tech company founders as one example, but like people who go off and create something new, who push on a disagreeable or controversial idea—we need that to drive forward. We need that sort of spirit. We need people to be able to put out ideas and be wrong and not be ostracized from society for it, or not have it be something that they get canceled for or whatever. We need people to be able to take a risk in their career because they believe in some important scientific quest that may not work out or may sound really controversial or bad. Certainly when we started OpenAI and we were saying we think this AGI thing is real and could be done—unlikely but so important if it happens—and all of the older scientists in our field were saying those people are irresponsible, you shouldn't talk about AGI, they're selling a scam, they're being reckless and it's going to lead to an AGI winter. We said we believed, we knew it was unlikely but it was an important quest and we were going to go after it. And the haters—that's important to a society.

你认为根源是什么？为什么现在做这类事情的年轻人比一二十年前少了？

What do you think is the origin? Like, what why do you think there are less young people that are doing those kind of things now as opposed to a decade or two ago?

我对这个话题非常感兴趣。我倾向于归咎于教育系统，但我确定它与社会以各种奇怪的方式相互作用。有趣的是，最近我的推特上到处都是关于什么导致了过去几十年美国男性睾酮水平下降的讨论。没有人说‘这是症状，不是原因’。每个人都说‘哦，是微塑料，是避孕药，是这个，是那个’。我认为这根本不是这个话题最重要的部分，但对我来说社会学上很有趣的是，只有关于什么导致了它的讨论，而没有关于它是社会某种变化的结果的讨论。

I am so interested in that topic. I'm tempted to blame the education system, but I'm sure that interacts with society in all of these strange ways. It's funny, there was this thing all over my Twitter feed recently trying to talk about what caused the drop in testosterone in American men over the last few decades. And no one was like 'this is a symptom not a cause.' Everyone was like 'oh it's the microplastics, it's the birth control pills, it's the whatever.' And I think this is not at all the most important piece of this topic, but it was just interesting to me sociologically that there was only talk about it being about what caused it, not about it being an effect of some sort of change in society.

但难道不是原因吗？

But isn't what caused it?

嗯，有生物学原因——当我们谈论邻苯二甲酸盐和微塑料农药、环境因素时，那些是真实的，完全正确。而且我不——再说一次，我在这方面远远超出了我的深度和专业知识。这只是让我觉得有趣的是，唯一的讨论是关于生物因素，而不是社会可以以某种方式产生影响——嗯，社会肯定有影响。

Well, there are biological reasons why—when we talk about the phthalates and microplastic pesticides, environmental factors, those are real, totally. And I don't—again, I'm so far out of my depth and expertise here. This was just interesting to me that the only talk was about biological factors and not that somehow society can have some sort of effect on—well, society most certainly has an effect.

你知道这个问题的答案吗？

Do you know what the answer to this is?

我不知道。我的意思是，我和写过《倒计时》一书的 Shanna Swan 博士有过交流，那本书全是关于石化产品的引入以及与之相关的睾酮下降、流产率上升。这些是普遍存在的内分泌干扰物——当他们给人们做血液检测时，他们发现一些疯狂的数字，比如 90%以上的人体内都有邻苯二甲酸盐。我很感激金属杯。我们尽量减轻影响，但我的意思是，如果你用微波炉加热食物，你就会摄入它，你就在摄入。如果你吃加工食品，你也会摄入。你的饮食中会有一定量的微塑料，据估计高达每周一张信用卡的量——你的身体每周消耗一张信用卡的微塑料。哇。真正令人担忧的是哺乳动物，因为引入——当他们用哺乳动物做研究时——

I don't. I mean, I've had a—with Dr. Shanna Swan who wrote the book Countdown, and that is all about the introduction of petrochemical products and the correlating drop in testosterone, rise in miscarriages. The fact that these are ubiquitous endocrine disruptors—when they do blood tests on people, they find some insane number, it's like 90 plus percent of people have phthalates in your system. And I appreciate the metal cups. We try to mitigate it as much as possible, but I mean, you're getting it if you're microwaving food, you're getting it, you're just getting it. You're getting it if you eat processed food. You're getting a certain amount of microplastics in your diet, and estimates have been that it's as high as a credit card of microplastics per—in your body, you consume a credit card of that a week. Whoa. The real concern is with mammals, because the introductions—when they've done studies with mammals and—

他们把这种东西引入体内。有一个相关现象是，这些动物的会阴缩小了。比如哺乳动物的会阴，雄性比雌性大 50%到 100%。雄性引入这些东西后，会阴开始缩小，阴茎缩小，睾丸缩小，精子数量减少。所以我们知道这些化学物质与身体相互作用有直接的生物学联系。这是真实存在的。还有我们使用的石化产品数量，我们使用的塑料数量。它已经成为我们文化、社会、文明如此不可或缺的一部分，无处不在。我在想，如果你考虑这些领地猿如何进化成这个新的高级物种，最好的方法之一难道不是消除最成问题的东西之一——睾酮吗？我们需要睾酮。我们需要有攻击性的男人和保护者。但为什么需要他们？因为还有其他邪恶的攻击性男人，对吧？所以我们需要保护者来保护我们自己。我们需要善良强壮的人来保护我们免受邪恶强壮的人的伤害。但如果我们正在与科技融合，如果科技是我们生活中不可避免的一部分，如果它无处不在，你正在使用它，你的微波炉、电视、电脑，你使用的一切，随着时间的推移，它将越来越成为你生活的一部分。随着这些塑料被引入人类生物系统，你看到雄性物种的女性化。你看到出生率下降。你看到所有这些相关因素，它们会引导我们变得更和平、更少暴力、更少攻击性、更少自我驱动，世界确实在随着时间的推移变得如此。我显然支持减少暴力。但我不认为……显然睾酮有很多好处，也有一些坏倾向。但我不认为一个世界，如果我们把那个排除在外，只是说一个世界有精神，我们要保卫自己，我们要找到保护自己和我们的部落、我们的社会进入未来的方法，这可以通过很多其他方式实现，我认为这是一个重要的冲动。但更重要的是，我指的是回到年轻创始人哪里去了的问题？为什么我们没有更多这样的人？我认为不仅仅是科技创业行业。你可以说年轻科学家或其他许多类别也是如此。这些可能只是我最了解的。在一个有任何科技的世界里，我仍然认为我们必须，在某种意义上这是我们的命运，保持在这个曲线上，我们仍然需要去弄清楚下一步是什么，下一个山丘之后，再下一个山丘之后。我的感觉是，这里正在发生一些长期的社会变化，我认为这也让我们更不快乐，对吧？它可能让我们更不快乐。但我要说的是，如果人类物种确实与科技融合，促进这一点的一个好方法难道不是让原始猿类女性化，并淡化角色吗？比如，AGI 世界应该……也许我不知道是不是 AGI，我的意思是，也许这只是科技不可避免的后果，特别是因为我们使用的这种科技确实含有大量塑料，再加上涉及食品系统、防腐剂的技术，以及我们用来确保人们不会饿死的所有这些不同东西。我们在这方面取得了巨大进步。这个国家很少有人饿死。这不是主要问题。但暴力是主要问题。但我们对暴力的担忧，以及对睾酮、强壮男人和有权势的人的担忧，仅仅是因为我们需要防范其他同样的事情。这真的是唯一的原因吗？

They've introduced these into their body. There's a correlating one thing that happens is these animals, their taint shrinks. Like the taint of the mammal, when you look at males, it's 50% to 100% larger than the females. With the introduction of these on the males, the taints start shrinking, the penises shrink, the testicles shrink, sperm count shrinks. So we know there's a direct biological connection between these chemicals and how they interact with bodies. So that's a real one. And it's also the amount of petrochemical products we have, the amount of plastics we use. It is such an integral part of our culture, our society, our civilization. It's everywhere. And I've wondered if you think about how these territorial apes evolve into this new advanced species, wouldn't one of the very best ways be to get rid of one of the things that causes the most problems, which is testosterone? We need testosterone. We need aggressive men and protectors. But why do we need them? We need them because there are other aggressive men that are evil, right? So we need protectors from ourselves. We need the good strong people to protect us from the bad strong people. But if we're in the process of integrating with technology, if technology is an inescapable part of our life, if it is everywhere, you're using it, you have the internet of everything that's in your microwave and your television, your computers, everything you use, as time goes on that will be more and more a part of your life. And as these plastics are introduced into the human biological system, you're seeing a feminization of the males of the species. You're seeing a downfall in birth rate. You're seeing all these correlating factors that would sort of lead us to become this more peaceful, less violent, less aggressive, less ego-driven thing, which the world is definitely becoming over time. And I'm all for less violence, obviously. But I don't look... obviously testosterone has many great things to say for it and some bad tendencies too. But I don't think a world, if we leave that out of the equation and just say like a world that has a spirit that we're going to defend ourselves, we're going to find a way to protect ourselves and our tribe and our society into this future, which you can get with lots of other ways, I think that's an important impulse. More than that though, what I meant is about if we go back to the issue of like where are the young founders? Why don't we have more of those? And I don't think it's just the tech startup industry. I think you could say that about young scientists or many other categories. Those are maybe just the ones that I know the best. In a world with any amount of technology, I still think we've got to, it is our destiny in some sense to stay on this curve, and we still need to go figure out what's next, and after the next hill, and after the next hill. And it would be my perception that there is some long-term societal change happening here, and I think it makes us less happy too, right? It may make us less happy. But what I'm saying is, if the human species does integrate with technology, wouldn't a great way to facilitate that be to kind of feminize the primal apes and to sort of downplay the role? Like, should the AGI world... maybe I don't know if it's AGI, I mean maybe it's just an inevitable consequence of technology, because especially the type of technology that we use, which does have so much plastic in it, and then on top of that the technology that's involved in food systems, preservatives, all these different things that we use to make sure that people don't starve to death. We've made incredible strides in that. There are very few people in this country that starve to death. It is not a primary issue. But violence is a primary issue. But our concerns about violence are, and our concerns about testosterone and strong men and powerful people, is only because we need to protect against other same things. Is that really the only reason?

当然，我是说有多少极其暴力的女性在经营帮派？没有，那部分肯定，显然不多。我更多的是指，社会重视强男性气质的唯一原因就是这个吗？是的，我认为是这样。我认为这是一种生物学上的必然，对吧？我认为这种生物学上的必然是因为我们过去必须抵御入侵的部落、捕食者和动物，我们需要一个比大多数人更强壮的人来保护其他人。这就是军队的概念。这就是为什么海豹突击队训练如此艰难。我们希望强者中的强者处于矛尖。但这只是因为外面有坏人。如果通用人工智能和某种改变人类生物结构的设备被实施，以至于这不再是问题，比如如果你是我，我是你，我通过科技知道这一点，暴力就不可能了。看，如果这完全走上科幻道路，我们都融合成一个单一的行星宇宙意识，那么是的，你不需要睾酮。特别是如果我们能通过其他方式繁殖。这就是外星人假说，对吧？为什么他们看起来那么细长，没有性别？他们有大脑袋和小身体。他们不需要体力。他们有某种心灵感应交流方式。他们可能不需要用嘴发出声音，也不需要我们那种征服和传播 DNA 的冲动。人们做的很多事情都是这些奖励系统，这些系统是在我们是领地猿时建立的。对我来说，有一个问题是你到底能摆脱多少。如果你制造了一个 AGI，它决定实际上我们不需要扩张，我们不需要更多领土，我们只是快乐，你和我，所有一切融合在一起，我们在地球上很快乐。我们不需要变得更大，不需要繁殖，不需要成长。我们只是坐在这里运行。这听起来像无聊的生活。我不同意。我不同意这会是逻辑结论。我认为逻辑结论是他们会寻找我们当前存在无法克服的问题和前沿，比如星际通信和跨……当它在另一个星系遇到另一个 AGI 时会发生什么？如果它遇到一个先进一百万年的 AGI 会发生什么？那会是什么样子？

Sure, I mean how many incredibly violent women are out there running gangs? No, that part for sure, clearly not very many. What I meant more is, is that the only reason that society values strong masculinity? Yeah, I think so. I think it's a biological imperative, right? And I think that biological imperative is because we used to have to defend against incoming tribes and predators and animals, and we needed someone who was stronger than most to defend the rest. And that's the concept of the military. That's why Navy SEAL training is so difficult. We want the strongest of the strong to be at the tip of the spear. But that's only because there are people like that out there that are bad. If general artificial general intelligence and the implementation of some sort of a device that changes the biological structure of human beings to the point where that is no longer a concern, like if you are me and I am you and I know this because of technology, violence is impossible. Look, by the time if this goes all the way down the sci-fi path and we're all merged into this one single planetary universal whatever consciousness, then yes, you don't need testosterone. Especially if we can reproduce through other methods. This is the alien hypothesis, right? Why do they look so spindly and without any gender? They have these big heads and tiny bodies. They don't need physical strength. They have some sort of telepathic way of communicating. They probably don't need sounds with their mouths, and they don't need this urge that we have to conquer and to spread our DNA. That's so much of what people do is these reward systems that were established when we were territorial apes. There's a question to me about how much you can ever get rid of that. If you make an AGI and it decides actually we don't need to expand, we don't need more territory, we're just happy, you and me, the whole thing all together, all merges in, we're happy here on Earth. We don't need to get any bigger, we don't need to reproduce, we don't need to grow. We're just going to sit here and run. That sounds like a boring life. I don't agree with that. I don't agree that that would be the logical conclusion. I think the logical conclusion would be they would look for problems and frontiers that are insurmountable to our current existence, like intergalactic communication and trans... what happens when it meets another AGI the other galaxy over? What happens if it meets an AGI that's a million years more advanced? What does that look like?

是的，这也是我经常想知道的。如果我们，我称我们自己为创造电子蝴蝶的生物毛毛虫。我们现在正在做一个茧。

Yeah, that's what I've often wondered. If we are, I call ourselves the biological caterpillars that create the electronic butterfly. We're making a cocoon right now.

而且我们甚至不知道自己在做什么。我认为这也与消费主义有关。因为消费主义做了什么？消费主义促进了更新更好的事物的创造，因为你总是想要最新、最棒的东西。所以你在汽车、电脑、手机以及所有这些东西上有了更先进的技术，包括医学科学。这当然都是真的。我在听你说话时反思的是，我不认为我像你那样乐观，认为我们能够或者应该克服我们的生物基础到你认为的程度。再往前推一层，我认为社会最幸福的时候是当同一个人或不同的人身上同时存在强烈的女性和男性特质时。我不认为很多这些根深蒂固的东西能够轻易被推到一边，同时还能保持一个有效的系统。当然，我们无法真正思考如果有一天机器有了意识会是什么样子。也许我只是思维太狭隘了，但我认为我们身上有一些东西在非常深层的层面上起作用，进化花了大量的搜索空间才达到这里，我不会轻易否定它。

And we don't even know what we're doing. And I think it's also tied into consumerism. Because what does consumerism do? Consumerism facilitates the creation of newer and better things because you always want the newest, latest, greatest. So you have more advanced technology in automobiles, computers, cell phones, and all of these different things, including medical science. That's all for sure true. The thing I was reflecting on as you were saying that is, I don't think I'm as optimistic that we can or even should overcome our biological base to the degree that I think you think we can. And to even go back one further level, I think society is happiest where there are roles for strong femininity and strong masculinity in the same people and in different people. And I don't think a lot of these deep-seated things are going to be able to get pushed aside very easily and still have a system that works. Sure, we can't really think about what if there were consciousness in a machine someday or whatever what that would be like. And maybe I'm just thinking too small-minded, but I think there is something about us that has worked in a super deep way, and it took evolution a lot of search space to get here, but I wouldn't discount it too easily.

但你不觉得穴居人可能也会对生活、久坐不动的生活方式、坐在电脑前、除了文字之外不互相交流得出同样的逻辑结论吗？嗯，我的意思是，这不就是你所说的正确之处吗？你认为我们的动机、真正带给我们真正快乐的东西，以及我们在深层上的连接方式，与穴居人有多大不同？显然很多其他东西变了，我们有更好的工具，但你认为他到底有多不同？

But don't you think that cave people would probably have those same logical conclusions about life and sedentary lifestyle and sitting in front of a computer and not interacting with each other except through text? Well, I mean, isn't that like what you're saying is correct? How different do you think our motivations are today and what really brings us genuine joy, and how we're wired at some deep level, differently than cave people? Clearly lots of other things have changed, we've got much better tools, but how different do you think he really is?

我认为问题就在这里：在基因层面上，基础没有太大区别，这些奖励系统就是我们与之互动的一切——无论是自我、欲望、激情、愤怒、嫉妒，所有这些不同的东西。而你认为有些人会上传并编辑掉这些？是的，没错。我认为我们担心失去作为人的这一方面，这种认为我们应该总是有冲突和斗争的想法，因为冲突和斗争是我们推动进步的方式，这是真的，对吧？与邪恶作斗争是善良变得更伟大、更强大的方式，如果善良赢了的话。但我的担忧是，这一切都基于一个前提，即我们现在的生物系统是正确的和最优的。我认为我们正在处理的抑郁、焦虑、缺乏意义和存在焦虑的高涨状态，很大程度上是因为作为人类动物的生物现实并没有真正与我们创造的这个世界的整合方式相匹配。这确实是肯定的。

I think that's the problem: genetically at the base level there's not much difference, and these reward systems are all we interact with—whether it's ego, lust, passion, fury, anger, jealousy, all these different things. And you think we'll be some people will upload and edit those out? Yes, yeah. I think that our concern with losing this aspect of what it means to be a person, this idea that we should always have conflict and struggle because conflict and struggle is how we facilitate progress, which is true, right? And combating evil is how the good gets greater and stronger if the good wins. But my concern is that that is all predicated on the idea that the biological system that we have right now is correct and optimal. And I think one of the things that we're dealing with with the heightened states of depression and anxiety and the lack of meaning and existential angst that people experience a lot of that is because the biological reality of being a human animal doesn't really integrate that way with this world that we've created. That's for sure, yeah.

我想知道解决方案是否不是在你被赋予的生物容器中寻找意义，而是将那些有问题的方面从系统中工程化出来，创造一个真正开明的存在。比如有一件事：如果你今天问别人，三年内世界没有战争的几率是多少，那是零。没有人认为没有，人类历史上从来没有过没有战争的时候。如果你必须说我们作为一个物种的头号问题是什么，我会说我们的头号问题是战争。我们的头号问题是这种观念：派遣大量互不相识的人去谋杀大量因政府、因沙线而对立的人是可以的。这是疯狂的事情。如何摆脱它？嗯，摆脱它的方法之一是完全工程化掉所有与资源获取相关的人类奖励系统。那么那时还剩下什么？嗯，我们是一个新事物。我认为我们变成了一个新事物。那个事物想要做什么？我认为那个新事物可能想要与其他更先进的新事物互动。我确实相信科学好奇心可以驱动这一点；这可以是一个长期的伟大前沿。

And I wonder if the solution to that is not to find ways to find meaning with the biological vessel you've been given, but rather to engineer those aspects that are problematic out of the system to create a truly enlightened being. Like one of the things: if you ask someone today what are the odds that in 3 years there will be no war in the world, that's zero. Like nobody thinks no, there's never been a time in human history where we haven't had war. If you had to say what is our number one problem as a species, I would say our number one problem is war. Our number one problem is this idea that it's okay to send massive groups of people who don't know each other to go murder massive groups of people that are somehow opposed because of the government, because of lines in the sand. It's an insane thing. How do you get rid of that? Well, one of the ways you get rid of that is to completely engineer out all the human reward systems that pertain to the acquisition of resources. So what's left at that point? Well, we're a new thing. I think we become a new thing. And what does that thing do? I think that new thing would probably want to interact with other new things that are even more advanced than it. I do believe that scientific curiosity can drive quite that; that can be a great frontier for a long time.

是的，我认为它也可以是一个长期的伟大前沿。我只是想知道，我们看到的由于微塑料导致的睾酮下降，这有点悄悄降临到我们身上——我们甚至不知道这是一个问题，直到人们开始研究它有多确定？我不知道。这是一个非常好的问题。Shanna Swan 博士认为它是睾酮下降、所有流产问题和低出生体重的主要驱动因素。所有这些似乎都有一个直接的环境因素。我确定还有其他因素。睾酮下降，我的意思是，已经证明你可以通过抗阻训练和某些你可以做的事情来增加男性的睾酮。比如他们在日本的一项研究中发现的一个大因素是运动前冷水浸泡能大幅提高睾酮。所以冷水浸泡然后运动？是的，我不知道，我可以找到那个。但这是一个迷人的研究领域。但我认为这与韧性和抵抗力有关，以及你的身体必须对抗这个非常极端的外部因素，这导致身体进入这种保存状态，并实施冷休克蛋白和减少炎症，这也增强了身体的内分泌系统。但除此之外，这种你必须变得更有韧性才能在你每天引入生活中的这个外部因素中生存下来的必要性。所以显然有办法让人类更强壮。我们知道我们可以通过力量训练做到这一点，所有这些实际上都会提高睾酮。你的饮食可以提高睾酮，而糟糕的饮食会降低它，并阻碍你的内分泌系统，阻碍你产生生长激素、褪黑激素的能力，所有这些不同的因素。这似乎是我们可以通过决定、选择和努力来修复或至少缓解的事情。但事实上，这些石化产品——比如 Shanna Swan 博士在她的书中有一张图表显示，在 1950 年代，当他们开始在一切中使用石化产品时——微波炉、塑料、保鲜膜，所有这些不同的东西——实施和下降之间存在直接相关性。这一切似乎都吻合。这似乎是一个主要因素。这对雌激素相关激素有同等影响吗？这是个好问题。

Yeah, I think it can be a great frontier for a long time as well. I just wonder if what we're seeing with the drop in testosterone because of microplastics, which sort of just snuck up on us—we didn't even know that it was an issue until people started studying how certain is that at this point that that's what's happening? I don't know. It's a very good question. Dr. Shanna Swan believes that it's the primary driving factor of the sort of dropping testosterone and all miscarriage issues and low birth weights. All those things seem to have a direct environmental factor. I'm sure there are other factors too. The drop in testosterone, I mean, it's been shown that you can increase males' testosterone through resistance training and through certain things you can do. Like one of the big ones they found through a study in Japan is cold water immersion before exercise radically increases testosterone. So cold water immersion and then exercise? Yeah, I don't know, I can find that. But it's a fascinating field to study. But I think it has something to do with resilience and resistance and the fact that your body has to combat this external factor that's very extreme that causes the body to go into this state of preservation and the implementation of cold shock proteins and the reduction of inflammation, which also enhances the body's endocrine system. But then on top of that, this imperative that you have to become more resilient to survive this external factor that you've introduced into your life every single day. So there are ways obviously that you can make a human being more robust. We know that we can do that through strength training, and all that stuff actually does raise testosterone. Your diet can raise testosterone, and a poor diet will lower it and will hinder your endocrine system, hinder your ability to produce growth hormone, melatonin, all these different factors. That seems to be something that we can fix, or at least mitigate, with decisions, choices, and effort. But the fact that these petrochemical products—like there's a graph that Dr. Shanna Swan has in her book that shows during the 1950s when they started using petrochemical products in everything—microwave, plastic, Saran wrap, all this different stuff—there's a direct correlation between the implementation and the dip. It all seems to line up. That seems to be a primary factor. Does that have an equivalent impact on estrogen-related hormones? That's a good question.

实际上，我知道他们讨论的一些化学物质确实会增加男性雌激素。我不确定，但我知道它会增加流产风险。所以我认为这总体上对人类是有害的。在短时间内对整个社会的内分泌系统造成干扰，这看起来是件坏事，而且很难完全理解。污染物和环境毒素，再加上杀虫剂、除草剂以及其他这些东西，还有微塑料——有很多因素导致我们的系统无法正常运作。

Actually, I know some of these chemicals they're talking about actually increase estrogen in men. I don't know, but I do know that it increases miscarriages. So I just think it's overall disruptive to humans. Definitely a society-wide disruption of the endocrine system in a short period of time seems like a bad thing and difficult to wrap our heads around. Pollutants and environmental toxins, on top of pesticides, herbicides, and all these other things, and microplastics—there are a lot of factors that are leading our systems to not work well.

但我真的想知道，我们是不是只是紧紧抓住这个猴子的身体不放？我们是否决定‘我喜欢我的猴子身体’？我也喜欢，听着，我爱它。但我也努力保持客观。当我客观地审视我们现在的位置和所有问题，并展望未来时，缓解这些问题的一个方法就是实现某种心灵感应技术。你不能只是发条刻薄的短信或推文给别人，因为当你发出那种能量时，你会真切地感受到对方的感受，你会感到厌恶。而暴力——如果你对某人施暴，并且你真切地在自己体内感受到那种暴力的反应，你也会失去暴力的动机。如果我们没有攻击性倾向，没有原始的猩猩倾向。

But I just really wonder if we are just clinging on to this monkey body. Are we deciding that I like my monkey body? I do too, listen, I love it. But I'm also trying to be very objective. When I objectively look at it in terms of where we are now and all our problems, and look towards the future, one way to mitigate a lot of these would be the implementation of some sort of telepathic technology. You couldn't just text someone or tweet at someone something mean because you would literally feel what they feel when you put that energy out there, and you would be repulsed. And violence—if you were committing violence on someone and you literally felt the reaction of that violence in your own being, you would also have no motivation for violence. If we had no aggressive tendencies, no primal chimpanzee tendencies.

确实，世界上的暴力在几十年里明显减少了很多，但情感暴力却大幅上升。而互联网在这方面非常糟糕。我不会走过去揍你，因为你看起来是个大块头，你会还手，而且社会规范也不允许那样做。但如果我不认识你，我可能会发一条刻薄的推文，而我毫无感觉。显然，这已经成为社会中的一场大流行病，而我们没有进化出某种生物约束来应对它。

It's true that violence in the world has obviously gone down a lot over the decades, but emotional violence is up a lot. And the internet has been horrible for that. I'm not going to walk over there and punch you because you look like a big strong guy, you're going to punch me back, and also there's a societal convention not to do that. But if I didn't know you, I might send a mean tweet about you and I feel nothing. And clearly that has become a mega epidemic in society that we did not evolve the biological constraints on somehow.

是的，我真的很担心这已经让我们变得不稳定，让所有人都痛苦。它肯定加剧了这一点，恶化了我们所有的问题。我的意思是，如果你读过乔纳森·海特的《美国心灵的娇惯》——你读过吗？很棒的书。它对女性，特别是成长中的年轻女孩伤害很大。社交媒体的发明、iPhone 的引入与自残、自杀、网络欺凌之间存在直接关联。人们总是在没人的时候议论别人，但现在他们公开地这样做来伤害别人，这太可怕了。显然，我认为这对男性也伤害很大，也许他们只是说得少一些。但我不认为我们任何一个人是为这个而生的。

Yeah, and I'm actually very worried about how much that's already destabilized us and made us all miserable. It certainly accentuated it, exacerbated all of our problems. I mean, if you read Jonathan Haidt's book 'The Coddling of the American Mind'—have you read it? Great book. It's very damaging to women, particularly young girls growing up. There's a direct correlation between the invention of social media, the introduction of the iPhone, self-harm, suicide, online bullying. People have always talked about people when no one's around, but the fact that they're doing it now openly to harm people is horrible. Obviously, I think it's super damaging to men too, maybe they just talk about it less. But I don't think any of us are set up for this.

不，没有人是为这个而生的。我认为名人比任何人都更清楚这一点。我们都习惯了。你只是变得麻木，或者如果你明智的话，你不去参与。我不参与。我的新手机上甚至没有任何应用。我决定，好吧，没有任何真正智能的东西。没有推特。我有一部单独的手机，如果我必须发点什么，我会拿起它。但我的新手机上只有短信。

No, no one is set up for it. And I think famous people know that more than anyone. We all get used to it. You just get numb to it, or if you're wise, you don't engage. I don't engage. I don't even have any apps on my new phone. I decided, okay, nothing that's really smart. No Twitter. I have a separate phone that if I have to post something, I'll pick it up. But all I get on my new phone is text messages.

这更多是为了保持你的头脑纯净和清晰吗？

Is that more just to keep your mind pure and uncluttered?

是的，而且不让自己受诱惑。只是——有多少次我早上第一件事就是起床上厕所，然后坐在马桶上刷了一个小时的 Instagram，什么也没得到？它对我毫无益处。而且还有种想法，以为我会从中得到些什么。

Yeah, and not tempt myself. And to just—how many times have I gotten up to go to the bathroom first thing in the morning and spent an hour just sitting on the toilet scrolling through Instagram for nothing? It does zero for me. And there's this thought that I'm going to get something out of it.

实际上就在昨天，我在想我们谈论了这么久这些算法推送会以大的方式操纵我们，这会发生。但在小的方面已经如此，刷 Instagram 甚至没那么有满足感。你刷完一个小时，然后想，‘我知道那是浪费时间’，但它超过了那个门槛，你几乎无法——很难放下手机。你只是希望下一个会有趣。而偶尔，问题是每 30 或 40 个短视频中有一个非常精彩。我想知道这是否比每个都精彩更强大。就像老虎机——你挖金子，而不是像摘雏菊一样直接摘。如果算法故意沿途喂给你一些。

I was thinking actually just yesterday about how we've all talked for so long about these algorithmic feeds manipulating us in big ways, and that will happen. But in the small ways already, scrolling Instagram is not even that fulfilling. You finish that hour and you're like, 'I know that was a waste of my time,' but it was over the threshold where you couldn't quite—it's hard to put the phone down. You're just hoping that the next one's going to be interesting. And every now and then, the problem is every 30th or 40th reel that I click on is wild great. I wonder by the way if that's more powerful than if everyone was wild, if everyone was great. It's like the slot machine—you mine for gold, you don't just go out and pick it like daisies. If the algorithm is intentionally feeding you some along the way.

是的，嗯，不幸的是外面有很多东西。但就——我和肖恩·阿里聊过，他是 UFC 拳手，显然有非常强大的头脑，非常有趣的人。肖恩说的一件事是，‘我刷东西时会有这种低度焦虑，我不知道为什么，那是什么？’我认为这是部分原因——逻辑思维意识到这是对你资源的巨大浪费。我也从手机上删掉了一堆那样的东西，因为我就是没有自控力。我的意思是，我有自控力去删除它，但没有自控力在开始刷之后停下来。

Yeah, well, there's just a lot out there unfortunately. But in terms of—I was talking to Sha Ali, who's this UFC fighter, obviously has a very strong mind, really interesting guy. One of the things that Shawn said is, 'I get this low-level anxiety from scrolling through things, and I don't know why, what is that?' And I think it's part of—the logical mind realizes it's a massive waste of your resources. I also deleted a bunch of that stuff off my phone because I just didn't have the self-control. I mean, I had the self-control to delete it, but not to stop once I was scrolling through.

是的，所以我认为我们在某些方面被注意力劫持了。它也有一些好处，但我们还没有到位的东西——工具、社会规范等等——来很好地调节它。而且我们不是为它而设计的。全新的技术再次劫持了人类奖励系统，劫持了为沟通而设置的所有制衡机制，而历史上沟通是一对一的。历史上，沟通是一个人对着另一个人。当人们互相写信时，通常是情书或‘我想你’之类的东西——他们暴露了一些难以在你面前表达的东西。而且除非那个人是精神病，否则不会是仇恨信件。而能够简单地传达‘那个家伙，我希望他被车撞’是如此简单、容易和可行。而且你不会体验到——推特在这方面尤其可怕。其机制以某种方式奖励，我认为没有人完全理解，它利用了人类心理的某些东西。但这就是你获得参与度、获得粉丝、获得多巴胺冲击的方式。而我认识的那些整天泡在推特上的人，大多数对此不满而非满意。

Yeah, and so I think we're just getting attention hacked in some ways. There's some good to it too, but we don't yet have the stuff in place—the tools, the societal norms, whatever—to modulate it well. And we're not designed for it. Completely new technology that again hijacks our human reward systems and hijacks all the checks and balances that are in place for communication, which historically has been one-on-one. Historically, communication has been one person to another. When people write letters to each other, it's generally things like love letters or 'I miss you'—they're exposing something they have difficulty expressing in front of you. And it was never—unless the person was a psycho—hateful letters. Whereas the ability to just communicate 'that guy, I hope he gets hit by a bus' is so simple and easy and available. And you don't experience—Twitter seems to be particularly horrible for this. The mechanics work in a way that it really rewards, in ways that I don't think anybody fully understands, tapping into something about human psychology. But that's kind of how you get engagement, how you get followers, how you get the dopamine hits or whatever. And the people who I know that spend all day on Twitter, more of them are unhappy about it than happy.

哦，是的，他们是最不快乐的。我的意思是，我关注了不少人，我关注他们只是因为他们疯了。然后我会去看看他们在推什么。有些人每天在上面待 8 到 10 个小时。我会整天看到推文。

Oh yeah, they're the most unhappy. I mean, there are quite a few people that I follow that I only follow because they're crazy. And then I'll go and check in on them and see what they're tweeting about. And some of them are on there 8, 10 hours a day. I'll see tweets all day long.

我知道那个人不可能快乐。他们不快乐，而且停不下来。你停不下来。这似乎就是他们的生活。这是一种……他们从中获得意义，从强化中，你知道，短期的意义。我想也许每天你上床睡觉时感觉完成了一些事，得到了多巴胺。而每到十年结束，你可能会想，‘这十年去哪了？’

And I know that person cannot be happy. They're unhappy, and they cannot stop. You can't stop. And it seems like it's their life. It's a... and they get meaning out of it in terms of reinforcement, you know, short-term meaning. I think maybe each day you go to bed feeling like you accomplished something and got your dopamine. And at the end of each decade, you probably like, 'Where'd that decade go?'

是的。我和一个朋友聊过，他在这方面真的有问题。他说他走在街上都得看手机，看谁回复了，甚至不看路。他完全被这些交流的焦虑所困。

Yeah. I was talking to a friend of mine who was having a real problem with it. He's saying he would be literally walking down the street and he'd have to check his phone to see who's replying, and he wasn't even looking where he was walking. He was just like caught up in the anxiety of these exchanges.

不是因为别人说的好话？不，不，不，不。全是……而且他意识到，你知道，他在贬低别人，然后看回复。

And it's not because of the nice things people say? No, no, no, no. It's all... and with him, he was recognizing that, you know, he was dunking on people and then seeing responses.

是的。我很久以前就不那么做了。我不再在推特上以消极方式与人互动。我就是不做了。即使我不同意某人，我也会尽可能平和地说。我其实有比我愿意承认的更严重的网络喷子倾向，所以我尽量不给自己太多诱惑，但有时也会犯错。

Yeah. I stopped doing that a long time ago. I stopped interacting with people on Twitter in a negative way. I just won't do it. Even if I disagree with someone, I'll say something as peacefully as possible. I have like more of an internet troll streak than I would like to admit, and so I try to just not give myself too much the temptation, but I slip up sometimes.

是啊，太诱人了。完全同意。太诱人了，而且很有趣。说点刻薄话很有趣。我是说，再说一次，无论我们之前谈到的这个生物系统，它获得正奖励，嗯，那一刻就有反应。你知道，有反应。你说些离谱的话，就会有人回应，那种回应就像能量。还有那么多其他人在参与你的想法。但最终，对大多数人来说，这没有建设性，而且心理上充满危险。事情太多了。我不认识哪个整天参与其中的人是快乐的。肯定没有。

Yeah, it's so tempting. Totally. It's so tempting, and it's fun. It's fun to just say something shitty. I mean, again, whatever this biological system we were talking about earlier that gets a positive reward, well, moment there's a reaction. You know, there's reactions. You say something outrageous and someone's going to react, and that reaction is like energy. And there's all these other human beings engaging with your idea. But ultimately, it's just not productive for most people, and it's psychologically fraught with peril. There's just so much going on. I don't know anybody who engages all day long that's happy. Certainly not.

我觉得我见过它像摧毁——这个词太重了——但像让那些善良、聪明、有责任心的人偏离轨道，毁了事业、生活、幸福或人际关系，他们就是无法对抗这个恶魔。因为它劫持了他们的多巴胺，而且因为人们孤独隔离，情况更糟，因为他们觉得对别人说刻薄话感觉更好。‘如果我不快乐，我就说更糟的话。’还有心理层面，比如社交隔离带来的焦虑，以及对这种会杀死所有人的无形疾病的恐惧。所以你有这个……然后你在推特上与人互动，然后陷入那种焦虑，整天都在做。我认识不少人，尤其是喜剧演员，真的因此失去理智，也失去了同行的尊重。

I think I've watched it like destroy—is too strong of a word—but like knock off track the careers or lives or happiness or human relationships of people that are like good, smart, conscientious people that just like got couldn't fight this demon. Because it hacked their dopamine and really accentuated that because people were alone and isolated, and that made it even worse because then they felt even better saying shitty things to people. 'If I'm unhappy, I'm going to say even worse things about you.' And then there was the psychological aspect of it, like the angst that came from being socially isolated and terrified about this invisible disease that's going to kill us all. And so you have this... and then you're interacting with people on Twitter, and then you're caught up in that anxiety, and you're doing it all day. And I know quite a few people, especially comedians, that really lost their minds and lost the respect of their peers by doing that.

我非常同情那些在疫情期间失去理智的人，因为这对我们所有人来说是多么不自然的事。隔离太残酷了。但很多人确实如此，我不认为互联网，尤其是像推特这样的社交动态，能激发出任何人最好的一面。

I have a lot of sympathy for people who lost their minds during COVID, because what an unnatural thing for us all to go through. And the isolation was just brutal. But a lot of people did, and I don't think the internet, and particularly the social dynamics of things like Twitter, brought out anyone's best.

不。嗯，我的意思是，有些人，如果他们没有刻薄倾向，我认为有些人确实寻求安慰，并以积极方式与人互动。我看到了很多积极面。但问题是负面互动的影响要大得多。

No. Well, I mean, some people, I think if they're not inclined to be shitty to people, I think some people did seek comfort and they did interact with people in positive ways. I see plenty of positive. I think the thing is that the negative interactions are so much more impactful.

是的。你看，我认为有很多人用这些系统做美好的事。我不是说不是这样。但这不是人们一天结束后离开平台时的情绪驱动力，对吧？

Yeah. I look, I think there are a lot of people who use these systems for wonderful things. I didn't mean to imply that's not the case. But that's not what drives people's emotions after getting off the platform at the end of the day, right?

对。而且很可能不是……如果你看推特互动量的饼图，我会说很多都是关于人们对事情生气。

Right. And it's also probably not... if you looked at a pie chart of the amount of interactions on Twitter, I would say a lot of them are on people being angry about things.

你认识的那些每天用推特八到十个小时的人，有多少整天都在说别人的好话，而不是恶毒的话？很少。是的，很少。我一个都不认识。

How many of the people that you know that use Twitter those eight or ten hours a day are just saying wonderful things about other people all day versus the virulent? Very few. Yeah, very few. I don't know any of them.

但话说回来，我想知道随着新技术的实现，是否会让沟通变得与我们目前所做的截然不同。我们现在做的沟通不如一对一交流那么沉浸。你我在交谈，我们看着对方的眼睛，接收社交信号，互相微笑，大笑。这是一种非常自然的交谈方式。我想知道通过技术的实现，它是否会变得比一对一对话更沉浸、更互动，你会更了解对方对你所说的话的感受，了解那个人的记忆、生活、历史、教育，了解他们的思想如何运作，他们的思想如何与你的思想互动，你看到他们，真正看到他们。我想知道我们现在经历的，是不是就像人们第一次发明枪时，他们只是开始朝东西射击，你知道吗？

But then again, I wonder with the implementation of some new technology that makes communication a very different thing than what we're currently doing. What we're doing now with communication is less immersive than communicating one-on-one. You and I are talking, we're looking into each other's eyes, we're getting social cues, we're smiling at each other, we're laughing. It's a very natural way to talk. I wonder if through the implementation of technology, if it becomes even more immersive than a one-on-one conversation, even more interactive, and you will understand even more about the way a person feels about what you say, about that person's memory, that person's life, that person's history, their education, how it comes out of their mind, how their mind interacts with your mind, and you see them, you really see them. I wonder if what we're experiencing now is just like the first time people invented guns, they just started shooting at things, you know?

是的。如果你能感受到我对你说刻薄话或好话时的感受，对吧？那显然会改变你决定说什么。

Yeah. If you can like feel what I feel when you say something mean to me or nice to me, right? That's clearly going to change what you decide to say.

是的，是的，是的。除非你是精神病。除非你是精神病。那么是什么导致一个人成为精神病，这能被设计掉吗？想象一个世界，我们能找到这些问题的根源，通过编码和某种技术的实现，提升多巴胺和血清素，对人做一些事，消除所有这些问题，让人们以非常纯粹的方式沟通。

Yes, yeah, yeah. Unless you're a psycho. Unless you're a psycho. And then what causes someone to be a psycho, and can that be engineered out? Imagine a world where we can find the root cause of those things, and through coding and some sort of implementation of technology that elevates dopamine and serotonin and does some things to people that eliminates all of those problems, and allows people to communicate in a very pure way.

听起来很棒。听起来很棒。但你就不会有摇滚乐了。脱口秀会死。你不会有暴力电影，你知道。很多东西都会消失。但也许这也是我们进化到下一阶段存在的一部分。我对此真的感到困惑。

It sounds great. It sounds great. But you're not going to have any rock and roll. You'll stand-up comedy will die. You'll have no violent movies, you know. There's a lot of things that are going to go out the window. But maybe that is also part of the process of our evolution to the next stage of existence. I feel genuinely confused on this.

嗯，我觉得你应该困惑。我是说，可以肯定，事情的发展是疯狂的。但我甚至没有……难以置信。我是说，你只是……从……OpenAI 是什么时候，你最初是什么时候开始这个项目的？

Well, I think you should be. I mean, to be sure, how it's going, that's insane. But I don't even have like... beyond belief. I mean, you just... from the... when did OpenAI, when did you first start this project?

呃，从一开始，2015 年底，2016 年初。

Uh, all the very beginning, end of 2015, early 2016.

当你最初开始这个项目时，你心里有什么时间线？它是否按那个时间线进行，还是完全失控了？

And when you initially started this project, what kind of timeline did you have in mind, and has it stayed on that timeline, or is it just wildly out of control?

我记得早期和联合创始人约翰·舒尔曼聊过，他说，‘是的，我觉得这大概是一个 15 年的项目。’我说，‘嗯，听起来差不多。’从那以后我一直这么想。现在我不再把 AGI 视为终点，但要达到我们完成目标的那一点……

I remember talking with John Schulman, one of our co-founders, early on, and he was like, 'Yeah, I think it's gonna be about a 15-year project.' And I was like, 'Yeah, it sounds about right to me.' And I've always sort of thought since then. Now I no longer think of AGI as quite the end point, but to get to the point where we like accomplish the thing...

我们当初设定的目标大概要到 2030、2031 年才能实现。对我来说，这始终是一个误差范围很大的合理估计，而且我觉得我们正沿着我原本假设的轨迹前进。

We set out to accomplish something that would take us to like 2030, 2031. That has felt to me like a reasonable estimate with huge error bars, and I kind of think we're on the trajectory I would have assumed.

你当时认为对社会的影响会是什么？当你刚开始做这件事时，你说好吧，如果我们成功了，创造出某种极其先进的 AGI，那么实施过程和对社会的影响会怎样？你是坐在那里，一边列优点一边列缺点，还是只是抽象地思考？

What did you think the impact on society would be? When you first started doing this, you said okay, if we are successful and create some massively advanced AGI, what is the implementation and impact on society? Did you sit there with a graph with pros on one side and cons on the other, or just abstractly consider?

我们确实讨论了很多缺点。我们中的许多人都非常担心，现在仍然担心，安全和对齐问题。如果我们构建这些系统，我们都能看到容易想象的伟大未来，但如果出了大问题，那真的是大问题。所以有很多讨论，而且创始精神的一个重要部分就是我们如何解决这个安全问题，这到底意味着什么。我们相信的一件事是，世界上最伟大的头脑无法在真空中解决这个问题；你必须接触现实，看看技术走向。实践比理论更离奇，这对我们来说确实如此。但我们有一个非常强烈的缺点清单，因为过去几十年的科幻小说都在告诉你这会如何出错，以及为什么它现在就应该朝我开枪。我肯定你见过约翰·康纳的 ChatGPT？

We definitely talked a lot about the cons. Many of us were super worried about, and still are, safety and alignment. If we build these systems, we can all see the great future that's easy to imagine, but if something goes horribly wrong, it's really horribly wrong. So there was a lot of discussion about that, and a big part of the founding spirit is how we're going to solve this safety problem, what that even means. One thing we believe is that the greatest minds in the world cannot solve that in a vacuum; you have to have contact with reality, see where the technology goes. Practice plays out in a stranger way than theory, and that's certainly proven true for us. But we had a very intense list of cons, because there's all the last decades of sci-fi telling you about how this goes wrong and why it's supposed to shoot me right now. I'm sure you've seen the John Connor ChatGPT?

我没见过。那是什么？

I haven't. What is it?

就像《终结者》里的约翰·康纳，那个孩子在你打开 ChatGPT 时看着你。所以那些东西在我们脑海中非常清晰。现在我认为我们有很多工作要做，但我们更了解如何让 AI 安全了。'AI 安全'这个词被过度使用——它是指不说冒犯性的话，还是指不毁灭全人类？这是一个连续体。但对于'不毁灭全人类'这个版本，我们有很多工作要做，但我认为我们终于有了更多关于什么可行的想法。鉴于系统的发展方向，我们比最初认为的技术方向有更多机会来解决这个问题。所以这很好。在积极方面，我当时最兴奋、现在仍然最兴奋的是：如果这个系统能大幅提高社会中的科学知识率呢？我认为所有真正的可持续经济增长和进步都来自于科学和技术能力的提高。这样我们就能解决所有问题。如果 AI 能帮助我们做到这一点，那一直是我最兴奋的事情。

It's like John Connor from Terminator, the kid looking at you when you open up ChatGPT. So that stuff was very clear in our minds. Now I think we understand there's a lot of work to do, but we understand more about how to make AI safe. The term 'AI safety' gets overloaded—does it mean don't say something people find offensive, or does it mean don't destroy all of humanity? There's a continuum. But for the 'don't destroy all of humanity' version, we have a lot of work to do, but I think we finally have more ideas about what can work. Given the way the systems are going, we have a lot more opportunities available to solve it than I thought we would have, given the direction we initially thought the technology was going. So that's good. On the positive side, the thing I was most excited about then and remain most excited about now is: what if this system can dramatically increase the rate of scientific knowledge in society? I think all real sustainable economic growth and progress comes from increased scientific and technological capacity. So we can solve all the problems. If AI can help us do that, that's always been the thing I'm most excited about.

这看起来确实是 AI 最大的潜在积极潜力——解决人类一直以来的许多问题，许多社会问题。这就是为什么我在谈论 AI 总统。我有点不是在开玩笑，因为我觉得如果某个东西超级智能，意识到所有变量，没有人类偏见，没有其他动机，只有美国社区及其与世界其他地区互动的更大利益，消除那些独裁者——无论当选还是非当选——他们将自己的意志强加于民众，因为他们有保护特殊利益集团和行业的既得利益……

It certainly seems like that is the greatest potential positive potential of AI—to solve a lot of the problems that human beings have had forever, a lot of the societal problems. That's why I was talking about an AI president. I'm kind of not joking, because I feel like if something was hyper-intelligent and aware of all the variables with no human bias and no incentives other than the greater good for the community of the United States and its interaction with the rest of the world, the elimination of these dictators—whether elected or non-elected—who impose their will on the population because they have a vested interest in protecting special interest groups and industry...

当你这么说时，我觉得可怕的是感觉人类失去了控制，我本能地不喜欢那样。但如果相反，它是人类集体意志的表达，没有沿途的误译和腐败影响，那么是的，我能看到。这可能吗？

The thing I find scary when you say that is it feels like humanity not in control, and I reflexively don't like that. But if it's instead the collective will of humanity being expressed without mistranslation and corrupting influences along the way, then yes, I can see it. Is that possible?

看起来是可能的。看起来如果它被编程为为人类谋取更大利益，并考虑人类的价值观和需求……

It seems like it would be. It seems like if it was programmed to do the greater good for humanity and take into account the values and needs of humanity...

'为人类谋取更大利益'这个说法让我感到恐惧。非常奥威尔式。所有这些都是。但 AGI 当然也是。

There's something about the phrase 'do the greater good for humanity' that terrifies me. Very Orwellian. All of it is. But so is artificial general intelligence, for sure.

当然。开门吧，房子。我希望我从事的是道德负担较小的事情。但这真的很令人兴奋。我想象不出更酷的事情可做了。我觉得自己是地球上最幸运的人。但这不是简单模式。这不是简单模式的生活。

For sure. Open the door, house. I wish I had worked on something less morally fraught. But it's really exciting. I can't imagine a cooler thing to work on. I feel like the luckiest person on Earth. But it's not on easy mode. This is not life on easy mode.

不，不，不，不。

No, no, no, no.

我的意思是，你正处于人类历史上最壮观的变化之一的前沿。我会说比互联网的实施更壮观。互联网的实施是这的第一步。AGI 是打了类固醇的互联网，是超空间中的互联网。

I mean, you are at the forefront of one of the most spectacular changes in human history. I would say more spectacular than the implementation of the internet. The implementation of the internet was the first baby steps of this. Artificial general intelligence is the internet on steroids, the internet in hyperspace.

我想说的是，这是下一步，之后还会有更多步，但这是我们迄今为止最激动人心的一步。

What I would say is it's the next step, and there will be more steps after, but it's our most exciting step yet.

我的好奇是那之后的下一步是什么？想想是不是很令人兴奋？

My wonder is what are those next steps after? Isn't that so exciting to think about?

非常令人兴奋。我认为我们是最后一批人，我真的这么认为。我认为我们是最后一批带有所有生物问题的生物人。

It's very exciting. I think we're the last people, I really do. I think we're the last of the biological people with all the biological problems.

你对此感到兴奋吗？

You feel excited about that?

我只是觉得事情就是这样。你对此很坦然。它就是那样。我认为在这一点上你无法控制它，除非发生某种巨大的自然灾害，把我们重置回石器时代，这也是我们应该非常关注的，因为这种情况似乎经常发生。我们没有意识到，因为人类身体的时间线太短了。人类存在的时间线，幸运的话是一百年，但地球的时间线是数十亿年。如果你看看地球上的生命被彗星撞击地球并完全消除所有技术进步重置了多少次，似乎在有记录的历史中已经发生过多次。我总觉得我们对此考虑得不够。

I just think that's just what it is. You're just fine with it. It is what it is. I don't think you can control it at this point, other than some massive natural disaster that resets us back to the Stone Age, which is also something we should be very concerned with because it seems like that happens a lot. We're not aware of it because the timeline of a human body is so small. The timeline of a human existence is a hundred years if you're lucky, but the timeline of the earth is billions of years. If you look at how many times life on Earth has been reset by comets slamming into the Earth and completely eliminating all technological advancement, it seems like it's happened multiple times in recorded history. I always think we don't think about that quite enough.

我们之前讨论过模拟假说。最近它在科技行业重新兴起。随着我们越来越接近 AGI，一个新的观点是，如果祖先在模拟我们，他们想一遍又一遍模拟的时间点正是创造 AGI 的前夕。

We talked about the simulation hypothesis earlier. It's had this big resurgence in the tech industry recently. One of the new takes on it, as we get closer to AGI, is that if ancestors were simulating us, the time they'd want to simulate again and again is right up to the lead-up to the creation of AGI.

是啊，我们活在这个时代看起来非常疯狂，但这绝非巧合。这个时代是在我们有足够多的手机遍布世界、录制海量视频来训练世界视频模型之后，这一切通过脑机接口或其他方式接入我们，而在 AGI 真正疯狂之前。这也是一个有趣的模拟时刻：我们能挺过去吗？小行星会不会在我们到达之前就撞上来制造戏剧性紧张？我们能找到让这一切安全的方法吗？我们能达成社会共识吗？所以这让比以前多得多的人相信了这一点。

Yeah, so it seems very crazy we're living through this time, but it's not a coincidence at all. This is the time after we had enough cell phones out in the world recording tons of video to train the video model of the world, all being jacked into us via brain implants or whatever, and before everything goes really crazy with AGI. It's also this interesting time to simulate: can we get through? Does the asteroid come right before we get there for dramatic tension? Do we figure out how to make this safe? Do we figure out how to societally agree on it? So that's led to a lot more people believing in it than before.

没错，我同意。而且我认为这就是大势所趋。我不知道这是好事还是坏事，它就是一件事。但活在现在肯定更好。我可不想活在 19 世纪，坐着一辆篷车试图穿越全国。

Yeah, for sure. And I think this is just where it's going. I mean, I don't know if that's a good thing or a bad thing, it's just a thing. But it's certainly better to live now. I would not want to live in the 1800s and be in a covered wagon trying to make my way across the country.

是啊，我们迎来了历史上最激动人心的时代。这是最好的，最好的。但同时也是问题最多的时代，社会问题最多，对社会、环境、基础设施问题的意识最强，而这些问题正等着我们去解决。

Yeah, we got the most exciting time in history yet. It's the best, it's the best. But it's also has the most problems, the most social problems, the most awareness of social, environmental, infrastructure issues that we get to go solve.

嗯，我直觉上感觉和你有几分不同，我认为接近现在这种形态的人类还会存在很久。我不认为我们是最后的人类。你觉得我们还有多久？

Um, and I intuitively think I feel something somewhat different than you, which is I think humans in something close to this form are going to be around for a lot longer. I don't think we're the last humans. How long do you think we have?

嗯，比我能够推理的时间框架还要长。现在，也许会有一些人选择融合，和 AI 一起去探索宇宙，宇宙很大。但我真的能想象一个世界，在没有自然灾害的情况下，地球上没有和今天人类非常相似的人在做人类的事情，而人类的精神融合到那些在宇宙中做自己事情的其他东西里？我很难真正看到这种情况发生。也许这意味着我会变成老古董，这个预测大错特错，但我要说，随着我们在 AI 上取得进展，我越来越这么觉得，而不是相反。

Um, longer than a time frame I can reason about, really. Now, there may be a world where some people decide to merge and go off exploring the universe with AI, and there's a big universe out there. But can I really imagine a world where, short of a natural disaster, there are not humans pretty similar to humans from today on Earth doing humanlike things, and the sort of spirit of humanity merged into these other things that are out there doing their thing in the universe? It's very hard for me to actually see that happening. And maybe that means I'm going to turn out to be a dinosaur and be horribly wrong in this prediction, but I would say I feel it more over time as we make progress with AI, not less.

是啊，我完全没有这种感觉。我觉得我们几年内就结束了。不，也许几十年。一两代人。这可能会是一个渐进的变化，就像穿衣服一样。你知道，我不认为所有人都穿衣服，然后他们发明了衣服。我想这大概花了一段时间。当有人发明了鞋子，我想那也花了一段时间。当他们发明了结构、门、房子、城市、农业，所有这些事情都是随着时间的推移慢慢实现的，然后变得无处不在。而我认为这更具变革性，也是其中的一部分。

Yeah, I don't feel that at all. I feel like we're done in like a few years. No, a few decades maybe. A generation or two. It'll probably be a gradual change, like wearing of clothes. You know, I don't think everybody wore clothes and they invented clothes. I think it probably took a while. When someone figured out shoes, I think that probably took a while. When they figured out structures, doors, houses, cities, agriculture, all those things were slowly implemented over time and then become everywhere. And I think this is far more transformative and is part of that.

因为你认为有些人不会有选择不融合的权利，对吧？就像有些人不再有选择不用电话的权利一样。你看，我以前有朋友说‘我连电子邮件都没有’。那些人现在不存在了。他们都有电子邮件。每个人都有手机，至少是翻盖手机。我认识一些人，他们就是应付不了社交媒体之类的东西，就换成了翻盖手机。对他们有好处。我不知道这是不是真的，我听说你不能再走进 AT&T 商店买翻盖手机了。我听说刚刚变了。可以吗？哦真的吗？有人告诉我的，但我不确定是不是真的。

Because you don't think there will be an option for some people not to merge, right? Just like there's not an option for some people to not have telephones anymore. See, I used to have friends like 'I don't even have email.' Those people don't exist anymore. They all have email. Everyone has a phone, at least a flip phone. I know some people that just can't handle social media and all that jazz, they went to a flip phone. Good for them. I don't know if this is true or not, I've heard you can't walk into an AT&T store anymore and still buy a flip phone. I heard that just changed. You can? Oh really? Someone told me this but I don't know if it's true.

Verizon 还有。我刚去过。他们还有翻盖手机。我当时想，我喜欢这个。我喜欢这个小东西，只能打电话。我一直很向往那样做，但我的步骤是换一个只有短信功能的手机。已经用了几天，目前感觉不错。

Verizon still has them. I was just there. They still have flip phones. I was like, I like it. I like this little thing that you just call people. And I always romanticized about going to that, but my step was to go to a phone that has nothing on it but text messages. And that's been a few days, feeling good so far.

是啊，挺好的。你知道，我还有一个手机用来刷社交媒体，但当我拿起它，我就开始刷 YouTube、看视频、刷 TikTok 或 Instagram。我没有 TikTok，但我试过 Threads 一小段时间，但我觉得，哦，这鬼城，所以又回到了 X。

Yeah, it's good. You know, I still have my other phone that I use for social media, but when I pick that up, I start scrolling through YouTube and watching videos and scrolling through TikTok or Instagram. And I don't have TikTok, but I've tried Threads for a little while, but I'm like, oh, this ghost town, so went right back to X.

我周末住在一个牧场，房子里有 Wi-Fi，但其他地方都没有手机信号。每周我都忘记这有多好，没有手机信号散步是多么大的改变。这对你的大脑有好处。对有些人来说，对你的大脑好得难以置信。而且我认为我们这么快就失去了某种东西。‘不在服务区’根本不会发生。甚至在飞机上也不再发生了。但那个你的手机无法嗡嗡响、没有短信、什么都没有的时刻，我认为那是一件非常健康的事情。

I live on a ranch during the weekends, and there's Wi-Fi in the house, but there's no cell phone coverage anywhere else. And it's every week I forget how nice it is and what a change it is to go for a walk with no cell phone coverage. It's good for your mind. For some, it's unbelievable for your mind. And I think we have so quickly lost something. 'Out of service' just doesn't happen. It doesn't even happen on airplanes anymore. But that hour where your phone just cannot buzz, no text message either, nothing. I think that's a really healthy thing.

有一次我在洛杉矶摔了手机。我想那是我最后一次摔手机。手机就像‘我们完了’，然后它开始随机给人打电话。它就会打给这个人，我挂断，它又打给另一个人。我挂断。我给我妻子看，我说，看这个，太疯狂了，它一直在打电话。所以手机坏了，我得订一个新手机。我们度假大概八天，苹果花了三天才给我送来手机。我打赌你那三天过得很棒。

I dropped my phone once when I was in LA. I think it was the last time I dropped the phone. The phone was like, we're done, and it just started calling people randomly. It would just call people, and I'd hang it up, and it would call another person. I'd hang it up. I was showing my wife, I was like, look at this, this is crazy, it's just calling people. So the phone was broken, and I had to order a phone. We were on vacation for like eight days, and it took three days for Apple to get me a phone. I bet you had a great three days.

太棒了。太棒了，因为当我和家人在一起时，我完全投入。没有选择，我也不想着看手机，因为它不存在。我没有手机。那种低水平的焦虑得到了缓解，就是那种你总想查看手机时的焦虑。

It was amazing. It was amazing because when I was hanging out with my family, I was totally present. There was no options, and I wasn't thinking about checking my phone because it didn't exist. I didn't have one. And there was an alleviation of that low level of anxiety, this sort of like that you have when you always want to check your phone.

是啊，我认为那件事太糟糕了。我们还没有弄清楚如何在社会中运作，并得到那些偶尔的手机坏掉三天或散步没有信号的时刻。但我非常觉得是我的手机在控制我，而不是相反。我讨厌这样，但我还没想出办法。

Yeah, I think that thing is so bad. We have not figured out yet how we're going to function in a society and get those occasional times when your phone is broken for three days or you go for a walk with no service. But I very much feel like my phone controls me, not the other way around. And I hate it, but I haven't figured out what to do about it.

嗯，这就是我对未来技术的担忧。这是如此出乎意料。如果你想象一个世界，当你……想象一下在 1984 年走到一个人面前，指着一部手机说，‘有一天它会放在你的口袋里，它会毁了你的生活。’就像，什么？‘是啊，有一天人们会对着那东西手淫。’你会说，什么？‘有一天人们会在 Instagram 上看人被谋杀。’过去几个月我在 Instagram 上看到过很多谋杀案。真的吗？我从来没见过。你的时间线不好。我和我的朋友 Tom Segura，每天早上我们互相发在 Instagram 上找到的最糟糕的东西。为什么？为了好玩。他是喜剧演员，我们都是喜剧演员。你觉得这好玩？是啊，这简直荒谬。我的意思是，疯狂的车祸，人被公牛顶伤，每一件……我们试图超越对方。所以每天他发给我最……每天我醒来，我就说，‘Tom，你找到什么了？’

Well, that's what I'm worried about with future technology. This was so unanticipated. If you imagine a world when you can... imagine going up to someone in 1984 and pointing to a phone and saying, 'One day that'll be in your pocket, it's going to ruin your life.' Like, what? 'Yeah, one day people are going to be jerking off to that thing.' You're like, what? 'One day people are going to be watching people get murdered on Instagram.' I've seen so many murders on Instagram over the last few months. Really? I've never seen one. Been a bad timeline. Me and my friend Tom Segura, every morning we text each other the worst things that we find on Instagram. Why? For fun. He's a comedian, I'm both comedians. That's fun to you? Yeah, this is just ridiculous. I mean, just crazy car accidents, people get gored by bulls, and every... we try to top each other. So every day he's sending me the most... every day when I wake up, I'm like, 'Tom, what you got?'

你能解释一下那有什么好玩的吗？

Can you explain what's fun about that?

嗯，他是个喜剧演员，我也是个喜剧演员，喜剧演员喜欢混乱。我们喜欢荒谬、离谱的事情，那些远远超出日常经历的事情。另外，还有对人类行为广泛谱系的理解。如果你是个好人，身边都是好人，你很少看到有人无缘无故在街上被枪击或刺伤。但在 Instagram 上，你每天都能看到。这其中有某种东西提醒你：‘哦，我们疯了。’人类物种中有一定比例的人就是失控了，在外面制造混乱，骑越野摩托车然后摔断脖子。所有这些事情都存在。

Well, he's a comic and I'm a comic, and comics like chaos. We like ridiculous, outrageous things that are just so far beyond the norm of what you experience in a regular day. Also, there's the understanding of the wide spectrum of human behavior. If you're a nice person and you surround yourself with nice people, you very rarely see someone get shot or stabbed for no reason randomly in the street. But on Instagram, you see that every day. There's something about that which reminds you, 'Oh, we're crazy.' The human species has a certain percentage of us that are just off the rails, out there causing chaos, jumping dirt bikes and landing on your neck. All that stuff is out there.

即使听到这些，我身体上也会感到不舒服，看着也不开心。

Even hearing that makes me physically uncomfortable and unhappy to watch.

哦，是的，也让我不舒服，但我们每天还是互相这样做。这不好，绝对不好，但我不会停止。这很有趣。

Oh yeah, it makes me uncomfortable too, but yet we do it to each other every day. It's not good, definitely not good, but I'm not going to stop. It's fun.

但为什么有趣呢？

But why is it fun?

有趣是因为我的朋友汤姆，我们在这方面有点像。我们就是看看这个看看那个，这是几个月前开始做的事情，停不下来。Instagram 已经学会了你这么做，所以不断给你展示更多。Instagram 知道。我去发现页面时，简直疯了。但问题是，它也会出现在你的信息流里。这就是我对算法的理解：它知道你醒着，所以即使不是关注的人，也会出现在你的信息流里。Instagram 还是会展示给我。

It's fun because it's my friend Tom, and we're both kind of the same in that way. We just look at this and that, and it's a thing we started doing a few months ago and can't stop. Instagram has learned that you do that, so it keeps showing you more and more. Instagram knows. When I go to the Discover page, it's just crazy. But the thing is, it shows up in your feed too. That's what I understand about the algorithm: it knows you're up, so it shows up in your feed even if they're not people I follow. Instagram shows them to me anyway.

几天前我听到一个关于 Instagram 和信息流的有趣说法：如果你在非高峰时段使用，那时他们有更多处理能力，因为使用的人少，你会得到更好的推荐。所以你的信息流在半夜会更好。但什么是‘更好’？你的信息流不是变得更让你上瘾吗？对我来说，更好就是更多谋杀、更多动物袭击。听起来很可怕，但它似乎知道那是我喜欢的，是我互动的，所以大部分时间都推送给我。

I heard an interesting thing a few days ago about Instagram and the feed: if you use it at off hours when they have more processing capability available because fewer people are using it, you get better recommendations. So your feed will be better in the middle of the night. But what is 'better'? Doesn't your feed become more addictive to you? For me, better would be more murders, more animal attacks. Sounds horrible, but it seems to know that's what I like, what I interact with, so it sends me that most of the time.

是的，那可能产生各种疯狂的心理影响，我敢肯定。这也是我想摆脱它、远离它的原因之一。

Yeah, that probably has all kinds of crazy psychological effects, I'm sure. That's also one of the reasons why I want to get rid of it and move away from it.

也许它走得太远了。我甚至不知道是不是太远了，但它确实在向我展示社会、文明、人类行为中更黑暗的区域。

Maybe it went too far. I don't even know if it's too far, but what it is, is showing me the darker regions of society, of civilization, of human behavior.

但你认为我们即将把所有这些都编辑掉吗？我想知道这是否是一个解决方案。我真的这么想，因为我不认为这超出了可能性范围。如果我们真的能设计，比如关于 Neuralink 的一个很有前景的讨论是脊髓损伤患者、无法移动身体的人，能够通过直接连接来控制那些你无法再控制的身体部位。这对受伤的人、瘫痪的人、有各种神经系统疾病的人来说将是了不起的事情。这可能是首批应用之一，埃隆也谈到过。首批应用之一是恢复视力，增强有脑部问题的人的认知功能。这非常令人兴奋。

But you think we're about to edit all that out? I wonder if that is a solution. I really do, because I don't think it's outside the realm of possibility. If we really can engineer, like one of the talks about Neuralink that's really promising is people with spinal cord injuries, people that can't move their body, and being able to hotwire that where it essentially controls all these parts of your body that you couldn't control anymore. That would be an amazing thing for people that are injured, for people that are paralyzed, who have all sorts of neurological conditions. That is probably one of the first implementations, and that's what Elon has talked about as well. One of the first implementations is the restoration of sight, cognitive function enhanced for people that have brain issues. That's tremendously exciting.

是的，像许多其他技术一样，我认为我们无法阻止神经接口，因为沿途会有巨大的好处。但我也认为我们不知道它会走向何方。这绝对是潘多拉魔盒，我认为一旦打开，我们就回不去了，就像没有自然灾害我们不会回到没有电脑的时代一样。

Yeah, and like many other technologies, I don't think we can stop neural interfaces because of the great good that's going to happen along the way. But I also don't think we know where it goes. It's Pandora's Box for sure, and I think when we open it, we're not going to go back, just like we're not going to go back to no computers without some natural disaster.

顺便说一句，我这么说是一种极大的赞美：你是我听过的谈论即将到来的融合时最中立的人之一。你就像，‘是的，我认为它会发生。它会在这些方面好，那些方面坏。’但你对此似乎难以置信地中立，这总是让我钦佩。

By the way, I mean this as a great compliment: you are one of the most neutral people I have ever heard talk about the merge coming. You're just like, 'Yeah, I think it's going to happen. It'll be good in these ways, bad in these ways.' But you seem unbelievably neutral about it, which is always something I admire.

我尽量对一切都保持中立，除了腐败，我认为这是我们文化治理方式中最严重的问题之一。我认为腐败和金钱的影响是巨大的可怕问题。但在社会问题和人类信仰与思考方式方面，我尽量保持中立，因为我认为真正理解他人想法的唯一方法是尝试通过他们的思维来看待。如果你有这种固有的偏见和非常僵化的好坏对错观，我认为这对理解人们为何不同没有帮助。所以我尽量在看待任何事情时保持中立和客观。这是我学会的一项技能；这不是我在 2009 年开始这个播客时就有的。我开始这个播客只是和朋友们一起，我完全不知道它是什么。我根本不可能知道，也不知道它对我作为人类的进化会产生什么影响。我变得好多了，对事物更加敏感，对他人想法和感受更加有意识。我和 2009 年的我完全是不同的人，这很难承认。难以置信。这真的很酷，但这只是我接受这种意外教育的必然结果。

I try to be as neutral about everything as possible, except for corruption, which I think is one of the most massive problems with the way our culture is governed. I think corruption and the influence of money are just giant terrible issues. But in terms of social issues and the way human beings believe and think about things, I try to be as neutral as possible because I think the only way to truly understand the way other people think is to try to look at it through their mind. If you have this inherent bias and a very rigid view of what's good and bad and right and wrong, I don't think that serves you very well for understanding why people differ. So I try to be as neutral and objective as possible when I look at anything. This is a skill I've learned; this is not something I had in 2009 when I started this podcast. I started this podcast just around with friends, and I had no idea what it was. There's no way I could have ever known, but also had no idea what it was going to do to me in terms of the evolution of me as a human being. I am so much nicer, so much more aware of things, so much more conscious of the way other people think and feel. I'm just a totally different person than I was in 2009, which is hard to recognize. It's hard to believe. That's really cool, but it's just an inevitable consequence of this unexpected education that I've received.

同理心是线性增长的吗？

Did the empathy kind of come on linearly?

不，不是线性的。它始于认识到我在社交媒体上的负面互动对我没有帮助，对那个人也没有帮助。然后对那个做了蠢事的人产生同情。仅仅贬低别人是不好的。除了获得某种社会认可和一堆点赞之外，没有任何好处。那并没有让我感觉良好。那不好。而且，从 2009 年开始，大量的迷幻体验让我更深刻地理解了影响。最近有一次，我通过它得到的压倒性信息是，我所说和所做的一切都会波及到我与之互动的所有人。如果我不以至少整体善或整体理解为目标做事，那我就是在做坏事。而那种坏是真实存在的，尽管你试图忽视它，因为你不会立即感受到它。

No, it was not linear. It came on recognizing that the negative interactions on social media that I was doing didn't help me, didn't help the person. And then having compassion for this person that's up or done something stupid. It's not good to just dunk on people. There's no benefit there other than to give you some sort of social credit and get a bunch of likes. It didn't make me feel good. That's not good. And also, a lot of psychedelics, a ton of psychedelic experiences from 2009 on, gave me a greater understanding of the impact. I had one recently, and the overwhelming message that I was getting through this was that everything I say and do ripples off into all the people that I interact with. And if I'm not doing something with at least the goal of overall good or overall understanding, then I'm doing bad. And that bad is a real thing, as much as you try to ignore it because you don't interface with it instantly.

你还在制造不必要的负面情绪，我应该尽可能避免。这是那次迷幻体验给我的一个压倒性信息。我服用它是因为那天我对世界局势特别焦虑，尤其担心乌克兰、俄罗斯、中国，以及我们国家的政治体制，还有左派和右派之间那种极度两极分化的互动方式。天啊，这一切看起来太痛苦了。有些日子我就是想太多。我觉得我需要什么东西来打破这种状态，所以我服用了这些迷幻药。

You're still creating unnecessary negativity and that I should avoid that as much as possible. It was like an overwhelming message that this psychedelic experience was giving me. And I took it because I was just particularly anxious that day about the state of the world, particularly anxious about Ukraine and Russia and China and the political system that we have in this country, and this incredibly polarizing way that the left and the right engage with each other. God, it just seems so tormented. So I was just some days I just think too much about it. I'm like, I need something to crack me out of this. So I took these psychedelics.

从你在 2010 年代初的预期到现在，迷幻疗法没有取得更多进展，你对此感到惊讶吗？作为一项医学成就，它在合法化道路上没有取得更大进展，你惊讶吗？

Are you surprised psychedelic therapy has not made, from what you thought would happen in the early 2010s till now, are you surprised it has not made more progress on a path to legalization as a medical achievement?

不，我不惊讶，因为有很多人不希望它实现，而这些人对我们的医疗系统和监管系统拥有巨大的权力。而且这些人也没有体验过这些迷幻药。很少有经历过深刻迷幻体验的人不认为这些东西总体上是有益的。所以问题在于，这些法律和规则是由完全不了解这些东西积极影响的人来执行的。如果你了解这个国家迷幻药禁令的历史，这一切都发生在 1970 年，实际上是为了阻止民权运动和反战运动。他们试图找到一种方法来阻止这些人所做的那些让他们以非常不同的方式思考的事情，比如‘调谐、开启、退出’。他们只是想阻止这一切。还有什么比把每个参与其中的人都关进笼子更好的办法呢？找到生产它的人，把他们关进笼子，终身监禁，确保它是非法的，逮捕人，在电视上播放突袭行动，确保人们都知道。然后你还把它与对社会固有危险和有害的药物联系起来，比如芬太尼危机、90 年代的可卡因危机。所有这些都被归在‘毒品’这个总称下。迷幻药也被当作毒品来谈论，尽管它们能带来深刻的精神和心理变化。我记得上小学时接受毒品教育，他们说大麻很糟糕，因为它是通往其他东西的‘入门毒品’，还有这个坏的那个坏的，海洛因什么的。而最糟糕的、最坏的东西就是 LSD。你吃了 LSD 就会想：‘哦，他们在撒谎。’迷幻疗法绝对是我生命中最重要的事情之一。鉴于它对我有多强大——我曾与各种焦虑和其他负面情绪作斗争，然后看着这一切消失——我去了另一个国家一周，做了几件事，回来时完全变了一个人。我当时想：‘我这辈子都被骗了。’是的，我很感激这件事发生在我身上。我和其他很多人聊过，都有类似的经历。我假设这是很久以前的事了，我确实这么认为，而且我对美国正在发生的事情非常感兴趣。我特别关注 MDMA 和裸盖菇素的进展，我想：‘好吧，这会通过的，这会以非常积极的方式改变很多人的心理健康。’我确实惊讶于我们在这方面没有取得更快的进展，但我仍然乐观地认为我们会取得进展。

No, I'm not, because there's a lot of people that don't want it to be in place, and those people have tremendous power over our medical system and over our regulatory system. And those people have also not experienced these psychedelics. There's very few people that have experienced profound psychedelic experiences that don't think there's an overall good for those things. So the problem is you're having these laws and these rules implemented by people who are completely ignorant about the positive effects of these things. And if you know the history of psychedelic prohibition in this country, it all took place during 1970, and it was really to stop the civil rights movement and the anti-war movement. They tried to find a way to make all these things that these people were doing, that was causing them to think in these very different ways, like 'tune in, turn on, drop out.' They just wanted to put a halt to that. What better way to lock everyone who participates in that in a cage? Find the people producing it, lock them in cages, put them in jail for the rest of their life, make sure it's illegal, arrest people, put the busts on television, make sure that people are aware. And then you also connect it to drugs that are inherently dangerous for society and detrimental, the fentanyl crisis, the crack cocaine crisis that we experienced in the 90s. All of those things are under the blanket of drugs. Psychedelic drugs are also talked about like drugs, even though they have these profound spiritual and psychological changes that they emit. I remember when I was in elementary school in drug education, they talked about marijuana being really bad because it's a gateway to these other things, and there's this bad one, that bad one, heroin, whatever. And the very end of the line, the worst possible thing, is LSD. Did you take LSD and go, 'Oh, they're lying.' Psychedelic therapy was definitely one of the most important things in my life. And I assumed, given how powerful it was for me, like I struggled with all kinds of anxiety and other negative things, and to watch all of that go away, like I traveled to another country for a week, did a few things, came back a totally different person. And I was like, 'I've been lied to my whole life.' Yeah, I'm so grateful that this happened to me now. Talked to a bunch of other people, all similar experiences. I assumed this was a while ago, I assumed it was, and I was very interested in what was happening in the US. I was particularly looking at where MDMA and psilocybin were on the path, and I was like, 'All right, this is going to get through, and this is going to change the mental health of a lot of people in a really positive way.' And I am surprised we have not made faster progress there, but I'm still optimistic we will.

嗯，从 1990 年代开始我们已经取得了很大进步。在 1990 年代，你从未听说过迷幻静修，从未听说过人们去度这种假，从未听说过人们聚在一起做这些事情，然后带着深刻的体验回来告诉别人，并且亲眼看到人们改变，看到他们的本质改变，看到人们变得不那么自私、不那么有毒、不那么刻薄，而是更有同理心、更善解人意。是的，我只能从个人经历来说。这给我的生活带来了彻底的改变，也让我与不同的人进行了所有这些对话。我感到非常幸运能够做到这一点，我与这么多不同的人进行了这么多不同的对话，他们的思维方式如此不同，还有这么多杰出的人取得了如此多不可思议的成就。你开始理解他们的大脑如何运作，他们如何看待世界，他们如何与事物互动。这太棒了，非常酷。这是作为人类最酷的事情之一，你可以找到一种方法来减轻猴子身体的所有负面影响，而且有现成的工具。但不幸的是，在这个非常禁止的社会，这个禁止的社会里，我们被剥夺了这些，而且我们被剥夺了那些从未杀死过任何人的东西，这真的很奇怪，因为奥施康定仍然可以开处方。为什么我们不能让这些改变了人们生活的药物更容易获得？为什么我们不能阻止阿片类药物危机或芬太尼？对旧金山来说，这似乎是一场难以置信的危机。你还记得在对话开始时，你说人工智能会带来很多好处，总体上是好的，但也不会造成伤害。如果我们合法化所有毒品，那也会产生同样的效果。你会主张合法化所有毒品吗？

Well, we have made so much progress from the time of the 1990s. In the 1990s, you never heard about psychedelic retreats, you never heard about people taking these vacations, you never heard about people getting together in groups and doing these things and coming back with these profound experiences that they relay to other people, and literally seeing people change, seeing who they are change, seeing people become less selfish, less toxic, less mean, and more empathetic and more understanding. Yeah, it's I mean, I can only talk about it from a personal experience. It's been a radical change in my life, as well as again having all these conversations with different people. I feel so fortunate to be able to do this, that I've had so many different conversations with so many different people that think so differently, and so many exceptional people that have accomplished so many incredible things. And you get to sort of understand the way their mind works, the way they see the world, the way they interface with things. It's awesome, it's pretty cool. And that is one of the cooler things about being a human, that you can find a way to mitigate all the negative aspects of the monkey body, and that there are tools that are in place. But unfortunately, in this very prohibitive society, this society of prohibition, we're denied those, and we're denied ones that have never killed anybody, which is really bizarre when you know OxyContin can still be prescribed. What's the deal with why we can't make, like why we can't get these medicines that have transformed people's lives more available? What's the deal with why we can't stop the opioid crisis or fentanyl? Seems like just an unbelievable crisis for San Francisco. You remember at the beginning of the conversation when you said that AI will do a lot of good, overall good, but also no harm. If we legalize drugs, all drugs, that would do the same thing. Would you advocate to legalize all drugs?

这是一个非常复杂的问题，因为我认为会有很多成瘾者本来不会成为成瘾者。会有很多人的生活因为合法化而被毁掉。有很多人可能在心理上无法处理这些事情。也许他们已经这样了，这就是迷幻药的问题：它们从不推荐给那些对现实把握不稳的人。当然，那些正在挣扎的人，那些已经服用大量药物以维持正常世界稳定状态的人，如果你直接用裸盖菇素轰炸他们，谁知道会产生什么效果，以及他们是否在心理上过于脆弱而无法恢复。我的意思是，有很多很多关于人们服用过量 LSD 而再也回不来的故事。是的，是的。这些药力强大得似乎都不足以形容，对吧？是的。但还有一点，人类为什么总是想要扰乱他们正常的意识状态？总是这样。无论是我们都在喝咖啡，你知道，人们抽烟，他们服用阿德拉，他们做各种……

It's a very complicated question, because I think you're going to have a lot of addicts that wouldn't be addicts. You're going to have a lot of people's lives destroyed because it's legal. There's a lot of people that may not be psychologically capable of handling things. Maybe they already have, like that's the thing about psychedelics: they do not ever recommend them for people that have a slippery grasp on reality as it is. For sure, people that are struggling, people that are already on a bunch of medications that allow them to just keep a steady state of existence in the normal world, if you just bombard them with psilocybin, who knows what kind of an effect that's going to have, and whether or not they're psychologically too fragile to recover from that. I mean, there's many, many stories of people taking too much acid and never coming back. Yeah, yeah. These are like a powerful doesn't seem to like begin to cover it, right? Yeah. But there's also what is it about humans that are constantly looking to perturb their normal state of consciousness, constantly? Whether it's we're both drinking coffee, you know, people smoke cigarettes, they take Adderall, they do all...

人们做各种各样的事情来改变和增强他们的正常意识状态。无论是冥想还是瑜伽，他们总是在做些什么来摆脱自己的束缚或分散自己对存在痛苦的注意力。这似乎是人类正常的一部分，甚至猴子——黑长尾猴会对酒精上瘾，它们对发酵的水果和酒精上瘾，然后变成醉鬼和酒鬼。你认为这背后的深层教训是什么？

People do all sorts of things to change and enhance their normal state of consciousness. Whether it's meditation or yoga, they're always doing something to get out of their own way or distract themselves from the pain of existence. It seems like a normal part of humans, and even monkeys—vervet monkeys get addicted to alcohol, they get addicted to fermented fruits and alcohol, and they become drunks and alcoholics. What do you think is the deep lesson there?

嗯，我们并不完全快乐，你知道。有些东西能让你快乐一小会儿——比如喝几杯能让你开心一阵子，直到你变成酒鬼，直到你毁了肝脏，直到你撞车，直到你卷入一些如果你没喝醉绝不会参与的暴力冲突。我喜欢咖啡因，它显然是一种毒品。我喜欢喝酒，但我常常想，‘是啊，这就像，哎呀，我真希望没喝这杯。’其他东西我大多选择避免。但那是因为你聪明，你可能知道利弊，知道它如何影响你，什么对你有益，什么有害。但那是你作为一个知情的人可以做出的决定，如果一切都是非法的，你就不被允许做这个决定。

Well, we're not happy exactly, you know. And some things can make you happy for a little bit—like a couple of drinks make you happy for a little while, until you become an alcoholic, until you destroy your liver, until you crash your car, until you're involved in some violent encounter you'd never be in if you weren't drunk. I love caffeine, which is clearly a drug. Alcohol I like, but I often think, 'Yeah, this is like, ding me, I wish I hadn't had this drink.' And other things I mostly choose to avoid. But that's because you're smart, and you're probably aware of the pros and cons, and how it affects you, what's good and what's detrimental. But that's a decision you can make as an informed human being, which you're not allowed to make if everything's illegal.

对，而且当东西非法时，罪犯会卖这些东西，因为通过使其非法并不能抑制欲望；你只是让获取变得更加复杂。我想说的是，芬太尼真的很棒——我不想亲身体验，但显然它很棒。彼得·伯格上过播客，他为 Netflix 制作了关于萨克勒家族的止痛药纪录片。他说他曾经娱乐性地服用过一次奥施康定，然后说，‘哦，天哪，太神奇了。离我远点，感觉太好了。’这是问题的一部分：它会毁掉你的生活，它会抓住你，但那种感觉难以置信。莱尼·布鲁斯是怎么描述的？我想他把海洛因描述成得到上帝温暖的拥抱。它给你的感觉可能相当壮观。我不知道合法化是否能解决问题，但我确实知道另一个我们没有注意的问题是卡特尔组织的崛起。就在我们的边境对面，有庞大的组织赚取数不清的钱。他们把芬太尼掺进街头的 Xanax 里，人们以为自己在吃 Xanax 却过量服用芬太尼而死。他们对可卡因也这样做，对一切东西都这样。芬太尼便宜且极其强效，如果东西是合法的，这就不是问题。那么，你最终会倾向于说，‘让我们干脆让一切合法化’吗？

Right, and also when things are illegal, criminals sell those things because you're not tempering the desire by making it illegal; you're just making access much more complicated. What I was going to say is that fentanyl is really great—I don't want to know, apparently it is. Peter Berg was on the podcast, he produced that painkiller documentary for Netflix about the Sackler family. He said he took OxyContin once recreationally and was like, 'Oh my God, it's amazing. Keep this away from me, it feels so good.' That's part of the problem: it will wreck your life, it will capture you, but the feeling is unbelievable. How did Lenny Bruce describe it? I think he described heroin as getting a warm hug from God. The feeling it gives you is probably pretty spectacular. I don't know if legalizing that will solve the problems, but I do know that another problem we're not paying attention to is the rise of the cartels. Right across our border, there are enormous organizations that make untold amounts of money. They put fentanyl in street Xanax, people overdose thinking they're getting Xanax and die from fentanyl. They do it with cocaine, with everything. Fentanyl is cheap and insanely potent, and that wouldn't be a problem if things were legal. So would you net out towards saying, 'Let's just legalize everything'?

我最终会倾向于那个方向，但我也会实施一些严肃的缓解措施，比如药物成瘾咨询和伊博格碱疗法，这是另一种非法的东西。有人刚刚告诉我这对他们来说是多么具有变革性。我没有亲身经历过，但伊博格碱对我许多有药丸问题的朋友来说——我的朋友埃德·克莱在墨西哥开了一个伊博格碱中心，因为他受伤了，对药丸上瘾，无法戒掉。他做了一次伊博格碱，一次，就戒掉了。一次 25 小时的超级不舒服的经历——据说很可怕。这不是娱乐性的，不是你想在周末和朋友一起做的事情。你做是因为你需要想办法摆脱这个烂摊子。想想这个国家在康复中心上花了多少钱，复发率是多少？非常高。我有很多朋友去过戒毒和戒酒康复中心，不少人都直接回去了。似乎效果不太好。当人们跌入谷底，有坚强的意志，参加 12 步计划或匿名戒毒会，从他人那里获得支持，最终建立其他行为的基础——锻炼、冥想、瑜伽，无论什么成为他们新的成瘾，一个更积极的成瘾——这时似乎有效。但身体成瘾的现实是，有缓解措施。很多人服用了裸盖菇素后完全戒掉了毒品，戒掉了香烟，因为他们意识到，‘哦，这就是我这样做的原因。’这就是为什么我更乐观地认为世界会在接受方面取得更快的进展。你听到很多这样的故事。所以我会说，显然我们现有的许多心理健康治疗充其量是无效的，显然我们的成瘾项目是无效的。如果我们有这个东西，在每一项科学研究中都显示出这些令人难以置信的结果，它就会发生。我仍然对此感到兴奋，我仍然认为这将是一个变革性的积极发展。它将改变政治，它将改变我们看待他人的方式，它将改变我们对整个社会和文化的看法，如果它合法化，它将改变人们相互互动的方式。它正在慢慢合法化——想想大麻，这个入门毒品。大麻现在在 23 个州是娱乐性合法的，在更多州是医疗合法的。在加州获得医疗执照非常容易。我在 1996 年就拿到了一个。过去你可以去某个地方说，‘我头疼’，然后砰，你就得到了合法的医用大麻处方。我以前得去恩格尔伍德，去贫民区，去恩格尔伍德健康中心。我想，‘这太疯狂了。’大麻现在算是有点合法了。然后在 2016 年，它在加州娱乐性合法化，这改变了一切。我有所有这些右翼人士，他们吃大麻食品来助眠。

I would net out towards that, but I would also put into place some serious mitigation efforts, like counseling for drug addiction and ibogaine therapy, which is another thing that's illegal. Someone was just telling me about how transformative it was for them. I haven't experienced it personally, but ibogaine for many of my friends who have had pill problems—my friend Ed Clay started an ibogaine center in Mexico because he had an injury, got hooked on pills, and couldn't kick it. He did ibogaine once, one time, done. A 25-hour super uncomfortable experience—it's supposed to be horrible. It's not recreational, not something you want to do on the weekend with friends. You do it because you need to figure out how to get out of this mess. Think about how much money is spent on rehabs in this country, and what's the relapse rate? It's really high. I have many friends who have been to rehab for drug and alcohol abuse, and quite a few went right back to it. It doesn't seem to be that effective. It seems effective when people hit rock bottom, have a strong will, get involved in a 12-step program or Narcotics Anonymous, get support from others, and eventually build a foundation of other behaviors—exercise, meditation, yoga, whatever becomes their new addiction, a more positive one. But the reality of physical addiction is that there are mitigation efforts. So many people have taken psilocybin and completely quit drugs, quit cigarettes, because they realize, 'Oh, this is why I'm doing this.' That's why I was more optimistic that the world would have made faster progress towards acceptance. You hear so many stories like this. So I would say, clearly a lot of our existing mental health treatment at best doesn't work, clearly our addiction programs are ineffective. If we have this thing that in every scientific study delivers these unbelievable results, it's going to happen. I still am excited for it, I still think it'll be a transformatively positive development. It will change politics, it will change the way we think of other human beings, it will change the way we think of society and culture as a whole, it will change the way people interact with each other if it becomes legalized. It's slowly becoming legalized—think of marijuana, the gateway drug. Marijuana is now legal recreationally in 23 states, and medically in many more. It's really easy to get a medical license in California. I got one in 1996. You used to be able to just go somewhere and say, 'I have a headache,' and bam, you get a legal prescription for weed. I used to have to go to Englewood, to the hood, to the Englewood Wellness Center. I thought, 'This is crazy.' Marijuana is now kind of sort of legal. Then in 2016 it became legal recreationally in California, and it changed everything. I have all these people who were right-wing, taking edibles to sleep.

就像这样，因为现在它合法了，他们以不同的方式看待它。我认为那种药物，一种相当温和的迷幻药，也有增强效果。它让人们更有同情心，更善良友好。它几乎是暴力增强药物的反面。它根本不增强暴力。它就像酒精、可卡因那样。说一件会让我非常不受欢迎的事：我讨厌大麻。它对我完全不适合。

Like this is because now that it's legal they thought about it in a different way. And I think that that drug, which is a fairly mild psychedelic, also has enhancing effects. It makes people more compassionate, it makes people more kind and friendly. It's sort of the opposite of a drug that enhances violence. It doesn't enhance violence at all. It just does—like alcohol does, that cocaine does. To say a thing that'll make me very unpopular: I hate marijuana. It does not sit well with me at all.

它对你做了什么让你不喜欢？

What does it do to you that you don't like?

呃，它让我在之后很长一段时间里感到疲倦和迟钝。

Uh, it makes me tired and slow for a long time after it.

我认为也存在生物差异，对吧？比如有些人，我妻子，她对酒精反应不好。她喝一杯就醉。但这是生物性的。她有某种基因——我忘了叫什么，关于胆红素？她的身体就是不太能处理酒精。所以她是个便宜约会对象。

I think also there's biological variability, right? Like some people, my wife, she does not do well with alcohol. She can get drunk off one drink. But it's biological. She has some sort of a genetic—I forget what it's called, something about Billy Rubin? Something that her body just doesn't process alcohol very well. So she's a cheap date.

哦，我的意思是基因上我有某种突变，让这成为一种不愉快的体验。是的，我说的是对我来说正好相反。酒精对我完全没有影响。我可以喝三四杯，20 分钟内就清醒了。我的身体，我的肝脏就像直接把它处理掉了。我能很快清醒。但我也不需要它。比如我在做“清醒十月”，整个月都不喝酒。昨晚做了演出，很棒，没问题。不喝酒似乎完全不影响我。但我确实喜欢它。我确实喜欢一杯葡萄酒。这是一天结束时美好的事情。我喜欢它。

Oh, all I meant is that genetically I got whatever the mutation is that makes it an unpleasant experience. Yeah, what I was saying is for me that's the opposite. Alcohol doesn't bother me at all. I could drink three or four drinks and I'm sober in 20 minutes. My body, my liver just blasts right through it. I can sober up real quick. But I also don't need it. Like I'm doing Sober October, for the whole month. I did shows last night, great, no problems. Not having alcohol doesn't seem to bother me at all. But I do like it. I do like a glass of wine. It's a nice thing at the end of the day. I like it.

说到这个和一般的迷幻药，你知道，许多文化在人的一生或成人礼中都有某种迷幻时刻的位置。但据我所知，大多数都有某种仪式。我确实担心——而且我认为这些东西太强大了——我担心它们被到处随意使用。

Speaking of that and psychedelics in general, you know, many cultures have had a place for some sort of psychedelic time in someone's life or rite of passage. But as far as I can tell, most of them are under—there's some sort of ritual about it. And I do worry that—and I think these things are so powerful—that I worry about them just being kind of used all over the place all the time.

是的，我希望我们作为一个社会——因为我认为即使进展缓慢，这也不会发生——找到一种方式来给予它应有的尊重。嗯，是的，看看情况如何。同意。

Yeah, and I hope that we as a society—because I think this is not going to happen even if it's slow—find a way to treat this with the respect that it needs. Um, yeah, see how that goes. Agreed.

是的，我认为环境和心态非常重要，在做之前、做的时候都要思考你在做什么。就像我前几天晚上那次迷幻体验，我当时想，天啊，有时候我就是想太多关于世界的事，它如此——而且你有孩子，你会想他们将在什么样的世界里长大？那天就是那种日子，我想，天啊，有这么多愤怒，这么多这个那个。然后它就把这些从我身上带走了，一整天剩下的时间。那天晚上我非常友好、非常快乐，就想拥抱每个人。我真的——我明白了。我——哦天哪，那并不是想错了。

Yeah, I think set and setting is very important, and thinking about what you're doing before you're doing it, while you're doing it. Like I was saying the other night when I had this psychedelic experience, I was just like, God, sometimes I just think too much about the world and that it's so—and you have kids and you wonder like what kind of a world are they going to grow up in? And it was just one of those days where I was just like, God, there's so much anger and there's so much this and that. And then it just took it away from me the rest of the day. Like that night I was so friendly and so happy and I just wanted to hug everybody. I just really—I got it. I got—oh my God, that's not thinking about it wrong.

你认为世界上的愤怒比以前高得多，还是我们只是——就像我们讨论的社交媒体带来的这些动态？

Do you think the anger in the world is way higher than it used to be, or we just—it's like all these dynamics from social media we were talking about?

我认为社交媒体的动态确实加剧了一些人的愤怒。但我认为世界上的愤怒只是不平等带来的挫败感的一部分，那些问题如此明显却得不到解决，以及人们所有的困扰。我的意思是，你在这个国家看到的许多大规模暴力、大规模抢劫等等，都是由穷人实施的，这并非巧合。

I think the dynamics in social media have certainly exacerbated anger in some people. But I think anger in the world is just a part of frustration in inequality, problems that are so clear but are not solved, and all the issues that people have. I mean, it's not a coincidence that a lot of the mass violence that you're seeing in this country, mass looting and all these different things, are being done by poor people.

你认为 AGI 会成为世界的均衡力量还是加剧不平等？

Do you think AGI will be an equalizing force for the world or further inequality?

我认为这取决于如何实施。我的担忧又回到了我们之前讨论的某种神经接口，它会将你的生产力提升到远超他人的程度，让你能控制更多资源，并通过它提升你的经济组合和世界影响力，以及你能获得的权力。它会在其他人参与之前就做到。因为我想象在财务上它会像早期的手机。你还记得电影《华尔街》里他拿着那个大砖头手机吗？就像，看他，他在海滩上拿着电话。那太疯狂了。那时没人有那种东西，它们非常稀有。1994 年我第一次搬到加州时买了一个，我觉得自己活在未来。不，那是摩托罗拉 StarTAC。那是个很酷的手机。实际上 1988 年我就在车里装了一个。我是最早拥有手机的人之一。我在车里装了一个，这很棒，因为我的朋友比尔·布鲁门瑞奇，他经营喜剧连线，他会打电话给我，因为他知道能联系到我。比如如果有人生病或退出，我就能接到演出，因为他能打电话给我。所以我在车里，他说：“乔，你在干嘛？今晚有空位吗？”我说：“没有，我有空。”他说：“太棒了。”然后他就给我演出。所以我通过这个电话得到了很多演出，它基本上自己付了钱。但我买它只是因为酷。比如我可以开车时打电话。老兄，我在开车，我在给你打电话。能开车打电话太疯狂了，我车上有个小天线，车顶上一个弯弯曲曲的小天线。但现在每个人都有。你可以去第三世界国家，小村庄里的人都有手机。这非常普遍，基本上无处不在。有手机的人比没有手机的人多。手机数量比人类数量还多，这相当疯狂。我认为最初的成本问题——一开始它会贵得令人望而却步。问题是富人能够做到，然后真正疯狂的人最终会在头上钻孔。如果那东西有效——也许不是第一代，第一代可能有问题，但第二代更好——总会有一个你必须进入游戏的时候。总会有一个你必须卖掉股票的时候，就像你知道的，别等太久。等等，就这样。一旦发生，我担心拥有它的人会比其他人有巨大优势，贫富差距会更大，更加两极分化。

I think it would depend on how it's implemented. My concern is again what we were talking about before with some sort of neural interface, that it will increase your ability to be productive to a point where you can control resources so much more than anyone else, and you will be able to advance your economic portfolio and your influence on the world through that, your amount of power that you can acquire. It will before the other people can get involved. Because I would imagine financially it'll be like cell phones in the beginning. You remember when the movie Wall Street when he had that big brick cell phone? It's like, look at him, he's out there on the beach with a phone. That was crazy. No one had one of those things back then, and they were so rare. I got one in 1994 when I first moved to California, and I thought I was living in the future. No, it was a Motorola StarTAC. That was a cool phone. I actually had one in my car in 1988. I was one of the first people to get a cell phone. I got one in my car, and it was great because my friend Bill Blumenreich, who runs the Comedy Connection, he would call me because he knew he could get a hold of me. Like if someone got sick or fell out, I could get a gig because he could call me. So I was in my car, he's like, "Joe, what are you doing? Do you have a spot tonight?" And I'm like, "No, I'm open." He's like, "Fantastic." And so he'd give me gigs. So I got a bunch of gigs through this phone where it kind of paid itself. But I got it just because it was cool. Like I could drive down the street and call people. Dude, I'm driving and I'm calling you. It was nuts to be able to drive and I had a little antenna, a little squiggly antenna on my car on the roof. But now everyone has one. You can go to a third world country and people in small villages have phones. It's super common, it's everywhere essentially. More people have phones than don't have phones. There are more phones than there are human beings, which is pretty wild. And I think that initial cost problem—it's going to be prohibitively expensive initially. And the problem is the wealthy people are going to be able to do that, and then the real crazy ones that wind up getting the holes drilled in their head. And if that stuff is effective—maybe it's not generation one, maybe there are problems with generation one, but generation two is better—there's going to be a time where you have to enter into the game. There's going to be a time where you have to sell your stocks, like you know, don't wait too long. Hang on, there go. And once that happens, my concern is that the people that have that will have such a massive advantage over everyone else that the gap between the haves and have-nots will be even further and it'll be more polarized.

这是我改变想法的一件事。你知道，几年前 OpenAI 的一个人对我说，你真的不能在没有计划的情况下让一些人崛起，因为这可能会造成难以置信的权力扭曲，我们将不得不就此进行某种社会讨论。

This is something I've changed my mind on. You know, someone at OpenAI said to me a few years ago, like, you really can't just let some people emerge without a plan, because it could be such an incredible distortion of power, and that we're going to have to have some sort of societal discussion about this.

是的，这看起来是真的。看起来——所以在这一点上，尤其是如果它像 AGI 一样有效，或者像，抱歉，ChatGPT 一样。ChatGPT 太棒了。当你输入信息，问它问题，它能给你答案，你可以让它为你编写一个网站，它不会立即——

Yeah, that seems real. That seems like—so at this point, especially if it's as effective as AGI is, or as, excuse me, ChatGPT is. ChatGPT is so amazing. When you enter into it information, you ask it questions and it can give you answers, and you could ask it to code a website for you and it doesn't instantly—

它能解决那些你本来需要几十年才能解决的问题，而且立刻就能搞定。这是它最笨的时候了。没错，这就是疯狂之处。所以想象一下，比这更先进的东西，经过多阶段的改进和创新，然后它直接与大脑连接。但只有付得起钱的人才能用。而那些人只是普通人类，并没有被增强。我们还没有进化，身体上还是老样子，仍然保留着所有人类奖励系统。我们基本上还是那些领地意识强烈的灵长类动物。现在你想想，某个精神病亿万富翁植入了这个设备，然后决定彻底劫持我们的金融系统，获取所有资源，制定只对自己有利的法规和影响力，并确保他们能控制这一切……

It solves problems that would literally take you decades to solve, and it gets to it right away. This is the dumbest it will ever be. Yeah, that's what's crazy. So imagine something like that, but even more advanced, multiple stages of improvement and innovation forward, and then it interfaces directly with the mind. But it only does it with the people that can afford it. And those people are just regular humans, so they haven't been enhanced. We haven't evolved physically; we still have all the human reward systems in place. We're still basically these territorial primates. And now you have, you know, just imagine some psychotic billionaire who gets this implant and decides to completely hijack our financial systems, acquire all the resources, set into place regulations and influences that only benefit them, and then make sure that they can control it from their...

你认为这在多大程度上真的需要物理植入或物理融合，而不仅仅是有些人能用上 GPT-7，花很多钱在推理算力上，有些人则不能？我觉得那也会带来巨大的变革。但我的想法是，一旦有了神经增强，想想人类心智以及它如何与世界互动，你如何与语言、思想和事实互动，而某种东西比这强大指数级，但它还允许你使用同样的情感、同样的自我、同样的欲望和驱动力——嫉妒、欲望、仇恨、愤怒——所有这些，但带着神一般的力量。当一个人能读心而其他人不能，当一个人能完全准确地预测股票、资源和商品的所有趋势，并据此做出选择……我完全理解这些。我唯一有点困惑的是，人类说话和听力的带宽，或者打字和阅读的带宽，并不高。但已经足够高了，如果你能直接说：‘告诉我股市会发生什么；如果我想赚大钱，现在该怎么做？’它就会在屏幕上显示给你。即使没有神经接口，你离你描述的场景也已经很近了。当然，对于股票，或者‘告诉我如何发明一项改变历史进程的新技术’，所有这些事情。我认为真正重要的是获取大量算力，尤其是差异巨大的算力，也许比接口本身更重要。我认为这肯定会在一个人拥有的权力和影响力中发挥巨大作用。我的担忧是，神经接口会带来的结果是，你不再是一个人类心智与数据交互；你现在是人类心智的一个超级先进版本，它有着我们甚至无法想象的深远可能性。我们可以想象，但无法真正概念化，因为我们没有背景，目前没有那种能力和可能性。我们只能猜测。但当它真的实现时，你面对的是一个完全不同的存在。我唯一能说的实话是：我不知道。

How much do you think this actually requires a physical implant or a physical merge, versus just some people have access to GPT-7 and can spend a lot on the inference compute for it and some don't? I think that's going to be very transformative too. But my thought is that once we have a neural enhancement, if you think about the human mind and how it interacts with the world, how you interact with language and thoughts and facts, and something exponentially more powerful than that, but it also allows you to use the same emotions, the same ego, the same desires and drives—jealousy, lust, hate, anger—all of those things, but with this godlike power. When one person can read minds and other people can't, when one person has a completely accurate forecast of all the trends in stocks, resources, and commodities, and they can make choices based on those... I totally see all of that. The only thing I feel a little confused about is that human talking and listening bandwidth, or typing and reading bandwidth, is not very high. But it's high enough where if you can just say, 'Tell me everything that's going to happen in the stock market; if I want to go make all the money, what should I do right now?' It goes and just shows you on the screen. Even without a neural interface, you're kind of a lot of the way to the scenario you're describing. Sure, with stocks or with, you know, 'Tell me how to invent some new technology that will change the course of history.' Yeah, all those things. I think what somehow matters is access to massive amounts of computing power, or especially differentially massive amounts, maybe more than the interface itself. I think that certainly is going to play a massive factor in the amount of power and influence a human being has, having access to that. My concern is that what neural interfaces are going to do is now you're not a human mind interacting with that data; now you are some massively advanced version of what a human mind is, and it has profound possibilities that we can't even imagine. We can imagine, but we can't truly conceptualize them because we don't have the context, we don't have that ability and that possibility currently. We can just guess. But when it does get implemented, you're dealing with a completely different being. The only true thing I can say is I don't know.

你有没有想过为什么是你？你有没有想过，‘我怎么就处在这场惊人变革的前沿了？’

Do you wonder why it's you? Do you ever think, 'How am I at the forefront of this spectacular change?'

嗯，首先，我认为这很像——你可以对很多公司说这句话，但没有哪家公司像 OpenAI 这样真实。CEO 远不是公司里最重要的人。在我们这里，有一大批研究人员，每个人单独来看，对我们迄今为止以及未来的成功都比我更关键。但即便如此，我敢打赌那些人也会想，‘嗯，成为他们真奇怪。’但对我来说，这已经足够奇怪了，以至于它稍微提高了我对模拟假设的概率估计。

Well, first of all, I think this is very much like—you could make this statement for many companies, but none is it as true as for OpenAI. The CEO is far from the most important person in the company. In our case, there's a large handful of researchers, each of which are individually more critical to the success we've had so far and that we will have in the future than me. But even that, and I bet those people really are like, 'Hmm, this is weird to be them.' But it's certainly weird enough for me that it ups my simulation hypothesis probability somewhat.

当你想到模拟理论的可能性时，你会给多少百分比？

When you think about the possibility of simulation theory, what percentage would you give?

我一直不知道该怎么给它一个数字。我读过的每一个解释为什么它概率极高的论点，在我看来都很有道理。但感觉无法推理。

I've never known how to put any number on it. Every argument I've read explaining why it's super high probability all seems reasonable to me. It feels impossible to reason about, though.

你呢？我也是。我可能会说也许，但它还是那样。而且我必须——主要是我觉得这没关系。我觉得，好吧，它确实有关系，我猜，但目前没办法知道。而且，什么才重要呢？如果这真的是——我的意思是，我们对生命固有的理解是，我们是这些生物体，与其他生物体互动，我们交配繁殖，这产生了更多的我们，然后随着社会进步，我们获得更多信息、更多理解和知识，下一代社会将比前一代更优越。这就是我们今天看待社会的方式。没有人想活在 1860 年，那时你会因为感冒而死，感染无药可治。现在活着要好得多，无可争议，对吧？除非你真的更喜欢《黄石》里那种简单生活。我们现在拥有的信息获取渠道，如果你选择去获取信息，你比以往任何时候都有更多的途径。这太酷了。而且随着时间的推移，如果你从有文字记载的历史开始到现在，一个明显的事实是，信息获取越多，人们就能做出更好的选择。他们并不总是做出更好的选择，但有了更多信息，他们肯定有更大的潜力做出更好的选择。你知道，我们认为这只是一个生物学现象，但想象一下如果并非如此。想象一下这是一个程序，你只是连接到这个程序的意识，它创造了这种体验，这种体验与我们通常认为的、碳基生命与真实世界中固体物理事物互动的真实生物体验无法区分。我仍然不清楚我应该做什么不同的事，或者想什么。

What about you? Same thing. I go maybe, but it's still what it is. And I have to—the main thing is I think it doesn't matter. I think it's like, okay, it kind of—it definitely matters, I guess, but there's no way to know currently. And also, what matters though? If this really is—I mean, our inherent understanding of life is that we are these biological creatures that interact with other biological creatures, we mate and breed, and that this creates more of us, and then hopefully as society advances and we acquire more information, more understanding and knowledge, this next version of society will be superior to the version that preceded it. That's just how we look at society today. Nobody wants to live in 1860 where you died of a cold and there's no cure for infections. It's much better to be alive now, inarguably, right? Unless you really do prefer the simple life that you see on Yellowstone or something. What we're dealing with now in terms of access to information, the lack of ignorance if you choose to seek out information—you have so much more access to it now than ever before. That's so cool. And over time, if you go back to the beginning of written history to now, one of the things that is clearly evident is the more access to information, the better choices people can make. They don't always make better choices, but they certainly have much more potential to make better choices with more access to information. You know, we think that this is just this biological thing, but imagine if that's not what's going on. Imagine if this is a program and that you are just consciousness that's connected to this thing that's creating this experience that is indistinguishable from what we like to think of as a real biological experience from carbon-based life forms interacting with solid physical things in the real world. It's still unclear to me what I'm supposed to do differently or think.

是的，没有答案。你百分之百正确。你能做什么不同的事呢？我的意思是，如果你像对待模拟一样存在，如果你只是像活在模拟中一样生活，那是解决方案吗？我不知道。我觉得无论如何它对我来说是真实的。它是真实的。我会那样生活。而如果真正的模拟真的实现了，那就会成为问题。是的，如果我们真的创造了一个真实的模拟……

Yeah, there's no answer. You're 100% right. What can you do differently? I mean, if you exist as if it's a simulation, if you just live your life as if it's a simulation, is that the solution? I don't know. I think it's real to me no matter what. It's real. I'm going to live it that way. And that will be the problem with an actual simulation if and when it does get implemented. Yeah, if we do create an actual simulation that's...

与现实生活无法区分。比如，模拟的规则是什么？它如何运作？那个模拟是否公平公正、更加合理和平？模拟中没有战争吗？我们是否都应该同意接入它，因为我们将拥有完全不同的生活体验？所有对犯罪和暴力的焦虑，那些我们真正恐惧的东西，在这个模拟中都将不复存在。

Indistinguishable from real life. Like, what are the rules of the simulation? How does it work? Is that simulation fair and equitable, much more reasonable and peaceful? Is there no war in that simulation? Should we all agree to hook up to it because we will have a completely different experience in life? All the angst of crime and violence, the things we truly are terrified of, will be non-existent in this simulation.

是的，我的意思是，如果我们继续下去，似乎只要从现在的 VR 外推……你看过 Lex Fridman 和马克·扎克伯格做的播客吗？我看了一些片段，但还没看完整。对，所以他们基本上是在元宇宙中使用非常逼真的物理化身。那是第一步，也许是第三步。在那个时候可能比 Pong 稍微高级一点。是的，也许是雅达利，也许你现在在玩太空侵略者。但不管是什么，它都走在通往那个无法区分的事物的道路上。这似乎是不可避免的。这两件事对我来说似乎是不可避免的。

Yeah, I mean, if we keep going, it seems like if you just extrapolate from where VR is now... Did you see the podcast that Lex Fridman did with Mark Zuckerberg? I saw some clips but haven't watched the whole thing yet. Right, so they're essentially using very realistic physical avatars in the metaverse. That's step one, maybe step three. Maybe it's a little beyond Pong at that point. Yeah, maybe it's Atari, maybe you're playing Space Invaders now. But whatever it is, it's on the path to this thing that will be indistinguishable. That seems inevitable. Those two things seem inevitable to me.

对我来说不可避免的是，我们将创造一种生命形式，一种远比我们先进的人工智能生命形式。一旦它变得有知觉，它就能创造出比自己好得多的版本。然后随着它拥有更好的版本，它会继续前进。如果它继续下去，它将达到神一般的能力：完全理解宇宙的每一个方面及其自身的结构，如何操纵它，如何穿越它，如何交流。而且，你知道，如果我们继续下去，如果我们存活一百年、一千年、一万年，并且我们仍然在这条技术能力指数级增长的道路上，那就是神。我们变成了类似神的东西。那可能就是我们所做的。那可能就是智能、好奇、创新的生命真正做的事情：它创造了一个东西，这个东西又创造了我们所居住的宇宙，就像创造了下一个模拟。然后是的，也许那就是宇宙本身的诞生：创造力和智能，一切都源于此。

The inevitable thing to me is that we will create a life form that is an artificial intelligent life form that's far more advanced than us. And once it becomes sentient, it will be able to create a far better version of itself. And then as it has better versions of itself, it will keep going. And if it keeps going, it will reach Godlike capabilities: the complete understanding of every aspect of the universe and the structure of itself, how to manipulate it, how to travel through it, how to communicate. And that, you know, if we keep going, if we survive a hundred years, a thousand years, ten thousand years, and we're still on this same technological exponential increasing in capability path, that's God. We become something like a God. And that might be what we do. That might be what intelligent, curious, innovative life actually does: it creates something that creates the very universe that we live in, like creates the next simulation. And then yeah, maybe that's the birth of the universe itself: creativity and intelligence, and that it all comes from that.

我以前有个关于大爆炸的笑话：如果大爆炸只是一件自然的事情呢？比如人类变得如此先进，以至于他们创造了一台大爆炸机器。然后，你知道，我们如此自闭，又充满了阿德拉，以至于我们对后果毫无概念或担忧。然后有人说‘我来按’，他们按下去，砰，我们每 140 亿年从头开始。那就是大爆炸。我的意思是，我不知道它会走向何方，但我确实知道，如果你从远处观察人类，如果你是一个外星生命形式，完全脱离对我们文化的任何理解，对我们生物本能的任何理解，你只是看着这个占主导地位的物种在这个星球上做什么：它制造更好的东西。这就是它所做的。

I used to have this joke about the Big Bang: what if the Big Bang is just a natural thing? Like humans get so advanced that they create a Big Bang machine. And then, you know, we're so autistic and riddled with Adderall that we have no concept or worry of the consequences. And someone's like, 'I'll press it,' and they press it, and boom, we start from scratch every 14 billion years. And then that's what a Big Bang is. I mean, I don't know where it goes, but I do know that if you looked at the human race from afar, if you were an alien life form completely detached from any understanding of our culture, any understanding of our biological imperatives, and you just looked at what is this one dominant species doing on this planet: it makes better things. That's what it does.

我同意。它发动战争，它偷窃，它做很多不该做的事，它污染，它做所有这些可怕的事情。但它也持续不断地创造更好的东西。无论是更好的武器，从投石机到步枪到炮弹到火箭船到高超音速导弹到核弹，它创造越来越好的东西。这是它做的第一要务。而且它从不满足于已有的东西。再加上消费主义，这已经融入我们的骨子里，这种对新事物或更好事物的持续渴望。嗯，这推动了创新，因为它提供了持续创新和不断创造更新更好事物所需的资源。

I agree. It goes to war, it steals, it does a bunch of things it shouldn't do, it pollutes, it does all these terrible things. But it also consistently and constantly creates better things. Whether it's better weapons, going from the catapult to the rifle to the cannonballs to the rocket ships to the hypersonic missiles to nuclear bombs, it creates better and better things. That's the number one thing it does. And it's never happy with what it has. And you add that to consumerism, which is baked into us, and this constant desire for new or better things. Well, that fuels that innovation because it gives it the resources it needs to consistently innovate and constantly create newer and better things.

嗯，如果我是一个外星生命形式，我会想，‘哦，它在做什么？它在试图创造更好的东西。’那么它的前沿是什么？技术。技术是我们创造的最具变革性、最壮观、最有趣的东西。而且最奇怪的是：我们如此习以为常，以至于你可以立即与新西兰的人 FaceTime。我们很快就能习惯任何事情，你知道，几乎把它视为理所当然。

Well, if I was an alien life form, I'd be like, 'Oh, what is it doing? It's trying to create better.' Well, what is the forefront of it? Technology. Technology is the most transformative, the most spectacular, the most interesting thing that we create. And the most alien thing: the fact that we are so comfortable that you can FaceTime with someone in New Zealand instantly. We can get used to anything pretty quickly, you know, and take it for granted almost.

是的。嗯，如果你是一个外星生命形式，看着我们，你会想，‘它在做什么？哦，它一直在制造更好的东西。它会继续制造更好的东西。’嗯，如果它继续制造更好的东西，它就会制造出更好的思考的东西。它现在就在这样做，而你是其中的一部分。它会制造出更好的思考的东西。嗯，那个更好的思考的东西现在基本上处于阿米巴原虫阶段，处于小型多细胞生命形式阶段。嗯，如果那个版本变成了奥本海默呢？如果那个版本变得，比如，如果它扩展得如此之大，变得如此超级智能，以至于与之前存在过的任何其他智能生命形式完全不同呢？而且它不断做同样的事情：制造越来越好的自身版本。嗯，那会走向何方？它会走向神，走向类似神的东西。也许神是真实存在的，但也许它是人类持续不断创新、不断渴望突破创造力和技术能力界限这一过程的真实结果。

Yeah. And well, if you were an alien life form and you were looking at us, you're like, 'What is it doing? Oh, it keeps making better things. It's going to keep making better things.' Well, if it keeps making better things, it's going to make a better version of a thinking thing. And it's doing that right now, and you're a part of that. It's going to make a better version of a thinking thing. Well, that better version of a thinking thing is basically now in the amoeba stage, it's in the small multicellular life form stage. Well, what if that version becomes an Oppenheimer? What if that version becomes, like, if it scales up so far that it becomes so hyper-intelligent that it is completely alien to any other intelligent life form that has ever existed here before? And it constantly does the same thing: makes better and better versions of itself. Well, where does that go? It goes to a god, it goes to something like a God. And maybe God is a real thing, but maybe it's a real consequence of this process that human beings have of consistently, constantly innovating and constantly having this desire to push the envelope of creativity and technological power.

我想这归结为也许在定义上存在分歧，关于你说的‘它变成神’是什么意思。比如，我完全可以……我认为它变得比我们聪明得多、能力强得多。如果继续下去，那个东西会变成什么？也许你所说的神一般的力量是指那个东西可以创造、可以模拟一个宇宙。是的，好吧，这个我能共鸣。我认为我们创造的任何东西仍然要受这个宇宙物理定律的约束，对吧？

I guess it comes down to maybe a definitional disagreement about what you mean by 'it becomes a god.' Like, I can totally... I think it becomes something much, unbelievably much smarter and more capable than we are. And what does that thing become if that keeps going? And maybe the way you mean it as a Godlike force is that that thing can then go create, can go simulate a universe. Yes, okay, that I can resonate with. I think whatever we create will still be subject to the laws of physics in this universe, right?

是的，也许那就是神存在的上层结构。‘神’这个词是个有争议的词，因为它已经被过度利用了。但你知道，我和斯蒂芬·迈耶有过一次对话，他是一位物理学家，我相信他是物理学家。他既是物理学家，也是宗教信徒。那是一次非常奇怪的对话，非常迷人的对话。他信仰基督，是的。他甚至相信复活，我觉得这很有趣。但和他交流很有趣，因为他有这些预先设计好的小演讲。他遇到过这些问题太多次了，以至于他对这些事情有非常措辞优美、非常清晰的回答，我感觉到这些就像段子，你知道吗？就像我和一个喜剧演员聊天时，喜剧演员有一个关于火车旅行的段子，然后他告诉你那个段子。就是那样。他有关于为什么他相信耶稣、为什么他相信的段子。非常非常聪明的人。但我提出这个问题：当我们思考神时，如果不是神创造了宇宙，而是宇宙就是神呢？而创造性的……

Yeah, maybe that is the overlying fabric that God exists in. The 'god' word is a loaded word because it's just been so co-opted. But you know, I was having this conversation with Stephen Meyer, who is a physicist, I believe he's a physicist. He's a physicist and he's also religious. It was a really weird conversation, very fascinating conversation. Believer in Christ, yeah. He even believes in the resurrection, which I found very interesting. But it's interesting communicating with him because he has these little pre-designed speeches. He's encountered all these questions so many times that he has these very well-worded, very articulate responses to these things that I sense are like bits, you know? Like when I'm talking to a comic, a comic has this bit on train travel and they tell you the bit. That's what it's like. He has bits on why he believes in Jesus and why he believes. Very, very intelligent guy. But I propose the question: when we're thinking about God, what if instead of God created the universe, what if the universe is God? And the creative...

所有生命和万物的力量就是宇宙本身，而不是认为有一个创造我们的东西。这接近很多东方宗教。我觉得这比许多其他宗教更容易理解。当我服用迷幻药时，我有那种感觉：周围存在着一种疯狂的创新和连接，但我们的思想太原始了。我们就是这种东西——这是我们曾经的样子。那是什么？有一个叫 Shane Against the Machine 的艺术家，他用 Zilant 符号做了一个黑猩猩头骨。背面有一个小 Zilant。他做了这件很酷的艺术品。但我在想，我们的局限性是否在于我们是灵长类动物的高级版本。我们仍然有所有这些：嫉妒、羡慕、焦虑、欲望、愤怒、恐惧、暴力——所有这些曾经对我们生存和走到这一步很重要的有害特质。随着时间的推移，我们会想办法通过工程手段消除它们。随着智能生命变得更加智能，我们创造出比我们更强大的智能生命版本，如果这种情况持续下去——我是说 ChatGPT。想象一下回到苏格拉底时代，给他看那个，解释它，给他看一部手机。他会说，‘你做了什么？这是什么？’我打赌他会对手机比 ChatGPT 更印象深刻。手机的通信能力，但获取信息的方式会如此深刻。那时，伽利略因为说地球不是宇宙的中心而被软禁。现在我们知道它不是。我们有卫星，我们把相机送入轨道。

The force of all life and everything is the universe itself, instead of thinking there's this thing that created us. This is close to a lot of Eastern religions. I think this is an easier thing to wrap my mind around than many other religions. When I do psychedelics, I get that feeling like there's this insane suit of innovation and connectivity all around us, but our minds are so primal. We're this thing—this is what we used to be. And what is that? There's a guy named Shane Against the Machine, an artist who created a chimpanzee skull made out of Zilant symbols. It's got a little Zilant on the back. He just made this dope art piece. Cool. But I wonder if our limitations are that we are an advanced version of primates. We still have all these things: jealousy, envy, anxiety, lust, anger, fear, violence—all these detrimental traits that were important for us to survive and get to this point. As time goes on, we will figure out a way to engineer those out. As intelligent life becomes more intelligent, and we create a version of intelligent life far more capable than us, if that keeps going—I mean, ChatGPT. Imagine going back to Socrates and showing him that, explaining it, showing him a phone. He'd be like, 'What have you done? What is this?' I bet he'd be much more impressed with the phone than ChatGPT. The phone's ability to communicate, but then the access to information would be so profound. Back then, Galileo was put under house arrest for saying the Earth is not the center of the universe. Now we know it's not. We have satellites, we send cameras into orbit.

我完全理解。我们大致知道和聪明人说话是什么感觉，所以你会想，‘哦，我没想到你可以和一个非人说话，而他们有时在回应中表现得像人。’但手机——如果你在 2000 年后醒来，看到一部手机，你对此没有任何模型，而且你没有逐渐到达那里。

I totally get that. We kind of know what it's like to talk to a smart person, so you're like, 'Oh, I didn't think you could talk to a non-person and have them be person-like in some responses some of the time.' But a phone—if you just woke up after 2,000 years and there was a phone, you'd have no model for that, and you didn't get to get there gradually.

是的，你没有。我的朋友 Eddie Griffin 有个笑话，说亚历山大·格拉汉姆·贝尔一定是在吸可卡因，因为只有吸了可卡因的人才会说，‘我想和一个不在这里的人说话。’这就是电话。可卡因会让人想那样做吗？我不知道，我没吸过，但我想是的。在我看来，它只会让人愤怒和混乱。

Yeah, you didn't. My friend Eddie Griffin has a joke about how Alexander Graham Bell must have been doing coke, because only someone on coke could be like, 'I want to talk to someone who's not even here.' That's what a phone is. Does coke make people want to do that? I don't know, I've never done coke, but I imagine it does. It seems to me like it just makes people angry and chaotic.

是的，有一点，但他们也有想法。

Yeah, a little, but they also have ideas.

回到这个问题：如果继续下去，它会走向何方？它必须达到某种不可能的能力水平。想想我们现在能用核能、核弹、高超音速导弹做什么——那些从人类创造力和想象力中提取出来的疯狂物理事物，通过工程和技术实现，这些物理设备如果带到 500 年前，将和魔法无异。我认为这非常了不起。所以继续下去，10 万年后，如果我们还在，如果像我们这样的东西还在，它能做什么？就像我不认为苏格拉底能预测手机一样，我无法预测。我可能完全错了。但也许这就是彗星存在的原因——也许是一次很好的重置。留下一些，给他们一个曾经存在的乌托邦世界的遥远记忆，让他们经历数千年的野蛮和恐怖行为，然后重建社会。这就是新仙女木期撞击理论：大约 11800 年前，在冰河时代末期，我们被多颗彗星击中，导致北美冰盖瞬间融化，一切都被淹没——这是《吉尔伽美什史诗》和《圣经》中洪水神话的来源。有物理证据：岩芯样本显示高水平的铱，这在太空中很常见，在地球上很稀有，还有撞击产生的微钻石。大约 30%的地球有这方面的证据。这些理论的支持者很可能是正确的，这就是为什么他们发现可追溯到 11000-12000 年前的古代建筑，而当时人们被认为是狩猎采集者。也许我们的时间线完全错了。

Back to this: where does it go if it keeps going? It has to go to some impossible level of capability. Just think of what we can do now with nuclear power, nuclear bombs, hypersonic missiles—the insane physical things we've taken out of human creativity and imagination, through engineering and technology, implemented these physical devices that are indistinguishable from magic if you brought them 500 years ago. I think it's quite remarkable. So keep going 100,000 years from now, if we're still here, if something like us is still here, what can it do? In the same way that I don't think Socrates would have predicted the phone, I can't predict that. I'm probably totally off. But maybe that's also why comets exist—maybe it's a nice reset. Just leave a few around, give them a distant memory of the utopian world that used to exist, have them go through thousands of years of barbarism and horrific behavior, and then reestablish society. This is the Younger Dryas impact theory: around 11,800 years ago, at the end of the Ice Age, we were hit by multiple comets that caused instantaneous melting of the ice caps over North America, everything flooded—the source of flood myths from the Epic of Gilgamesh and the Bible. There's physical evidence: core samples show high levels of iridium, common in space but rare on Earth, and micro diamonds from impacts. About 30% of the Earth has evidence of this. It's very likely that proponents of this theory are correct, and that's why they find ancient structures dating to 11,000-12,000 years ago, when people were thought to be hunter-gatherers. Maybe our timeline is really off.

我一直很感兴趣地关注着。Randle Carlson 是最值得关注的人。他几乎一生都致力于此，顺便说一句，这源于一次迷幻体验。他有一次吸了酸，看着一个巨大的峡谷，他有一个幻象，认为它是由于极地冰盖瞬间侵蚀而形成的，一股不可思议的水浪冲过地球，切出了这些路径。现在似乎有实际的物理证据表明那可能发生过。我们只是幸存者，我们重新出现了。社会和人类文明偶尔会倒退到原始状态。

I've been watching that with interest. Randle Carlson is the greatest guy to pay attention to that. He's kind of dedicated his whole life to it, which by the way happened because of a psychedelic experience. He was on acid once, looking at this immense canyon, and he had this vision that it was created by instantaneous erosion of the polar caps, and that it just washed this wave of impossible water through the Earth, cutting these paths. Now there seems to be actual physical evidence that that probably took place. We're just the survivors, and we have reemerged. Society and human civilization occasionally gets set back to a primal place.

是啊，谁知道呢。如果你是对的，我们在这里所做的是编辑掉我们身上所有不喜欢的冲动，我们到达那个世界——似乎有点无聊。所以也许那时我们必须制作一个新的模拟，让人们以为他们在经历一些戏剧。或者我们到达这个拥有这种力量的点，但富人和穷人之间的鸿沟太大，有人掌握了这项技术并用它来压迫没有它的人。我们没有缓解人类生物学问题，我们的奖励系统——‘我必须拥有更多，你必须拥有更少。’这是竞争的自然倾向，有人劫持了它。我认为这将是一个非常重要的问题，必须在第一批人按下那个按钮之前解决。你怎么看埃隆呼吁暂停 AI？他当时正在创办一家 AGI 公司。

Yeah, who knows. If you're right that what happens here is we kind of edit out all the impulses in ourselves that we don't like, we get to that world—seems kind of boring. So maybe that's when we have to make a new simulation to watch people think they're going through some drama. Or maybe we get to this point where we have this power, but the haves and the have-nots, the divide is too great, and people get a hold of this technology and use it to oppress people who don't have it. We didn't mitigate the human biological problems, the reward systems that we have—'I got to have more and you got to have less.' That's this natural inclination for competition, and someone hijacks that. I think this is going to be such a hugely important issue to get ahead of before the first people push that button. What do you think about when Elon was calling for a pause on AI? He was starting an AGI company while he was doing that.

是啊，他是不是在呼吁暂停之后才创办的？我想是之前，但我不记得了。无论如何，这是不是那种‘打不过就加入’的情况？

Yeah, didn't he start it after he was calling for the pause? I think before, but I don't remember. In any case, is it one of those 'you can't beat him, join him'?

我认为，说我们必须弄清楚如何让这件事安全、美好、广泛有益，这种直觉非常重要，但我觉得呼吁暂停是幼稚的。我认为，正如我们之前提到的，你无法通过坐在房间里苦思冥想而在安全方面取得进展。你必须看到技术的发展方向，必须接触现实。当我们试图在确保安全和有益的前提下向 AGI 迈进时，遇到任何障碍，我们都会尝试寻找技术、政策或社会解决方案来克服它。这可能是技术的局限，比如某些东西不工作、产生幻觉或不够智能，也可能是安全问题，我们需要重新分配资源来解决。但对我来说，这都是同一件事：我们在朝着目标前进的每一步中解决出现的问题。如果你想的话，可以称之为暂停——暂停能力发展以专注于安全——但实际上，我认为这个领域在这方面有点摇摆不定。安全与能力并不是两件分开的事。我认为这个领域的一个肮脏秘密是，我们只有一种进步的方式：我们可以更深入地理解和推动深度学习，这可以以不同的方式使用，但正是同一种技术最终会帮助我们解决安全问题。话虽如此，作为人类，从情感上讲，我非常理解为什么呼吁暂停很有诱惑力。这在生活中经常发生，对吧？这发展得太快了，我们得暂停一下。

I think the instinct of saying like we've really got to figure out how to make this safe and good and widely good is really important, but I think calling for a pause is naive at best. I kind of think you can't make progress on the safety part of this as we mentioned earlier by sitting in a room and thinking hard. You've got to see where the technology goes, you've got to have contact with reality. And then when we try to make progress towards AGI conditioned on it being safe and beneficial, when we hit any kind of block, we try to find a technical, policy, or social solution to overcome it. That could be about the limits of the technology and something not working, like hallucinations or not getting smart, or it could be a safety issue where we've got to redirect our resources to solve that. But for me, it's all the same thing: we're trying to solve the problems that emerge at each step as we get where we're trying to go. You can call it a pause if you want, if you pause capabilities to work on safety, but in practice I think the field has gotten a bit wobbly on that axle. Safety and capabilities are not these two separate things. I think one of the dirty secrets of the field is that we have this one way to make progress: we can understand and push on deep learning more, and that can be used in different ways, but it's that same technique that's going to help us eventually solve the safety. All that said, as a human emotionally speaking, I super understand why it's tempting to call for a pause. It happens all the time in life, right? This is moving too fast, we've got to take a pause here.

从国家安全的角度来看，我们率先提出这一点有多令人担忧？

How much of a concern is it in terms of national security that we are the ones that come up with this first?

嗯，我想说的是，如果我们的对手先提出这一点并用它来对付我们，而我们没有一定的能力，那感觉会很糟糕。但我希望发生的是，这可以成为一个时刻——回到之前的对话——我们团结起来，克服我们的原始冲动，说让我们一起作为一个俱乐部来做这件事。那样会更好，会很棒。也许通过 AGI 和这项技术的实施，它会让翻译变得即时和容易。

Well, I would say that if an adversary of ours comes up with it first and uses it against us, and we don't have some level of capability, that feels really bad. But I hope that what happens is this can be a moment where, to tie it back to the earlier conversation, we kind of come together and overcome our base impulses and say, let's all do this as a club together. That would be better, that would be nice. And maybe through AGI and the implementation of this technology, it will make translation instantaneous and easy.

嗯，这已经发生了，对吧？但我的意思是，它还没有实时发生。很快就能准确沟通的那个点，很快，很快。

Well, that's already happened, right? But I mean, it hasn't happened in real time. The point where you can accurately communicate very soon, very soon, very soon.

我确实认为，不管怎样，世界会在这里团结起来。我不认为人们完全意识到了利害关系，但这不是地缘政治的事情。如果这归结为地缘政治斗争或竞赛，我不认为有任何赢家。所以我对人们团结起来持乐观态度。

I do think, for what it's worth, that the world is going to come together here. I don't think people have quite realized the stakes, but this is not a geopolitical thing. If this comes down to a geopolitical fight or race, I don't think there's any winners. So I'm optimistic about people coming together.

我也是。我的意思是，我认为大多数人都会喜欢这样。如果你问活着的绝大多数人类，大家和睦相处不是更好吗？也许你不能完全说我们只会有一个全球性的努力，但我认为至少我们可以达到一个点，即有一套全球性的规则、安全标准、组织来确保每个人都遵守规则。我们为原子武器做到了这一点，在生物学领域也有类似的事情。我认为我们会做到的。

I am too. I mean, I think most people would like that. If you asked the vast majority of human beings that are alive, wouldn't it be better if everybody got along? Maybe you can't go all the way and say we're just going to have one global effort, but I think at least we can get to a point where we have one global set of rules, safety standards, organization that makes sure everyone's following the rules. We did this for atomic weapons, it's been similar things in the world of biology. I think we'll get there.

这是一个很好的例子，核武器，因为我们知道它们的破坏能力，正因为如此，自 1947 年以来我们就没有引爆过。相当不可思议。除了试验，我们没有在战争中使用过。1945 年还是 1947 年？第二次世界大战是什么时候结束的？他们投下炸弹不是 1947 年吗？我想那是 1945 年。我在想之后是否还有更多我不知道的。不，可能是 1945 年。我想是 1945 年。所以从 1945 年开始，这相当非凡。这很了不起。我本来不会预料到这一点。我想如果我能穿越回 1945 年，没有人会想到，哦，天哪，这将成为人们做的事情，只是向城市投掷炸弹。我的意思是，我会说我们不会存活太久，而且确实有这种恐惧。他们设法阻止了这一点，这种相互确保毁灭、自我毁灭、普遍毁灭的威胁，真是非常了不起。我的意思是，整个世界，我们有足够的武器让地球变得无法居住。完全如此。正因为如此，我们没有这样做，这是一个好迹象。我认为这应该给人们一些希望。

That's a good example, the nuclear weapons, because we know the destructive capability of them, and because of that we haven't detonated one since 1947. Pretty incredible. Other than tests, we haven't used one in terms of war. 1945 or 1947? When was the end of World War II? Wasn't it 1947 when they dropped the bombs? I think that was 1945. I was wondering if there's more after that I didn't know about. No, it might be 1945. I think it was 1945. So from 1945, which is pretty extraordinary. It's remarkable. I would not have predicted that. I think if I could teleport back to 1945, no one would have thought, oh my God, this is just going to be something that people do, just launch bombs on cities. I mean, I would have said we're not going to survive this for very long, and there was a real fear of that for sure. It's pretty extraordinary that they've managed to stop that, this threat of mutually assured destruction, self-destruction, universal destruction. I mean, the whole world, we have enough weapons to literally make the world uninhabitable. Totally. And because of that, we haven't done it, which is a good sign. I think that should give some hope.

我正想说，我认为这应该给人们一些希望。是的，应该。我的意思是，史蒂芬·平克因为他的工作而受到很多批评，因为他有点淡化今天的暴力，但他并不是在淡化今天的暴力，他只是看统计趋势。如果你看看今天的生活现实与 100 年前、200 年前相比，它要安全得多。为什么你认为这是一件有争议的事情？为什么有人不能说，当然我们仍然有问题，但情况正在好转？因为人们不想这么说。他们，尤其是活动家，完全沉浸在这种想法中，即今天存在问题，而且这些问题很大，还有纳粹，等等。但没有人说今天没有大问题，对吧？没有人说没有。但仅仅说今天的情况更好，有些人就是不想听。对。但那些人也是沉迷于问题的人。问题成了他们的全部生活，解决这些问题成了他们的身份，参与解决方案或他们认为的解决方案成了他们一生的工作。然后有人过来说，实际上生活比以往任何时候都安全。我可以看到这深深地否定了他们。是的。但这也是事实，你知道。再说一遍，问题是什么？为什么人们不能认识到这一点？嗯，这是灵长类动物的大脑。我的意思是，这是我们之前强调的所有问题。而克服这个问题的解决方案可能是通过技术。这可能是我们无需漫长进化就能做到的唯一方法，因为生物进化与技术进化相比相对缓慢。那可能是我们的瓶颈，我们仍然在处理这个灵长类动物的身体。而像通用人工智能，或者某种与之互动的实现形式，无论是 Neuralink，还是某种改变心智与其他心智互动方式的东西。

I was going to say, I think that should give some hope. Yeah, it should. I mean, Steven Pinker gets a lot of flak for his work because he just sort of downplays violence today, but it's not that he's downplaying violence today, he's just looking at statistical trends. If you look at the reality of life today versus life 100 years ago, 200 years ago, it's far safer. Why do you think that's a controversial thing? Like why can't someone say, sure we still have problems, but it's getting better? Because people don't want to say that. They, especially people who are activists, they're completely engrossed in this idea that there are problems today and these problems are huge, and there are Nazis, and... But no one's saying there aren't huge problems today, right? No one's saying there's not. But just to say things are better today, some people just don't want to hear that. Right. But those are also people that are addicted to the problems. The problems become their whole life, solving those problems becomes their identity, being involved in the solutions or what they believe the solution to those problems become their life's work. And someone comes along and says, actually life is safer than it's ever been before. I can see that that's deeply invalidating. Yeah. But also true, you know. And again, what is the problem? Why can't people recognize that? Well, it's the primate brain. I mean, it's all the problems that we highlighted earlier. And that might be the solution to overcoming that, is through technology. And that might be the only way we can do it without a long period of evolution, because biological evolution is so relatively slow in comparison to technological evolution. And that might be our bottleneck, is that we just still are dealing with this primate body. And that something like general artificial general intelligence, or something like some implemented form of engaging with it, whether it's Neuralink, something that shifts the way the mind interfaces with other minds.

说到生物进化，是不是很疯狂，我认为会有一些人，他们活着见证了晶体管的发明或发现——随便你怎么称呼——他们也将活着见证 AGI 的创造。一个人的一生。是的。你想知道一个疯狂的事吗？从……

Isn't it wild that, speaking of biological evolution, there will be people I think who were alive for the invention or discovery, whatever you want to call it, of the transistor, they will also be alive for the creation of AGI. One human lifetime. Yeah. You want to know a wild one? From the...

从莱特兄弟驾驶飞机到有人在广岛投下原子弹，不到 50 年。这太疯狂了。

From the Wright brothers flying the plane to someone dropping an atomic bomb on Hiroshima, it was less than 50 years. That's wild.

不到 40 年，对吧？太疯狂了。简直离谱。我是说，60 多年就登上了月球。太不可思议了。

Less than 40, right? That's crazy. Bananas. I mean, 60 something years to land on the moon. Nuts.

它会走向何方？虽然只是猜测，但肯定很有趣。这是我们这个时代最迷人的事情。

Where is it going? I mean, it's just guesswork, but it's interesting for sure. It's the most fascinating thing of our time.

这在智力上令人着迷，而且我认为这将是极其净收益的事情之一。我们一直在谈论世界上的问题，这很好地提醒了我们能改善多少。我们将要大幅改善，而这将是我们迄今为止创造的最强大的工具，帮助我们实现这一点。

It's fascinating intellectually, and I also think it is one of these things that will be tremendously net beneficial. We've been talking a lot about problems in the world, and it's a nice reminder of how much we get to improve. We're going to get to improve a lot, and this will be the most powerful tool we have yet created to help us do that.

我认为你是对的。这次对话很棒。谢谢你邀请我。很感激。感谢你所做的一切。随时告诉我们进展，如果你创造了 AGI，给我打个电话。让我们知道。

I think you're right. This was an awesome conversation. Thanks for having me. Appreciate it. And thanks for everything. Keep us posted, and if you create AGI, give me a call. Let us know.

好的，谢谢。谢谢。大家再见。

All right, thank you. Thank you. Bye, everybody.