Sam Altman 谈 AI 如何革新创业环境

今天，我和 OpenAI 联合创始人 Sam Altman 坐在一起。很久以前，你在斯坦福做过一场关于如何创办初创公司的演讲。从那以后，过去 10 年里最大的变化是什么？

Today I'm sitting down with Sam Altman, the co-founder of OpenAI. A long time ago, you gave a talk at Stanford about how to start a startup. What is the biggest thing that has shifted over the past 10 years since those talks?

显然是 AI。现在一个小团队能做的事和做事的速度已经完全变了。一个成立 10 周的初创公司现在看起来令人难以置信，相比之下 10 年前的同龄公司会很糟糕。你还要在一个完全不同的世界里保持竞争力。

Clearly AI. What is possible to do now with a small team and the speed at which you can do it have totally changed. A 10-week-old startup now is incredible compared to what one looked like 10 years ago. And you have to be competitive in a very different world.

如果你现在想到一个成立 10 周的初创公司，它移动速度极快的最佳例子是什么？

If you think of a 10-week-old startup today, what is the best example of one that moved extremely quickly?

我甚至不知道它叫什么，但我遇到了一个才成立 2 周左右的初创公司，它完全重新设计了办公生产力套件，为的是让 AI 成为文档、演示文稿、电子表格的一等消费者。这最近可能还相当于一年的工作量。

I don't even know what they're called, but I met a startup that is about 2 weeks old that has entirely redone an office productivity suite for a world where AI is a first-class consumer of documents, presentations, spreadsheets. That would have been a year's worth of work recently.

我会假设，如果创办初创公司的门槛和工具变得更普及，那么更难的初创公司会变得更容易起步。

I would assume that if the barriers to starting startups and the tools become more widespread, then harder startups become easier to start.

破解难的初创公司意味着什么变化得如此之快，以至于我对多年后什么才是真正困难且有价值的事情没有完美的心理模型。我听过很多人说任何物理世界的东西都额外有价值，因为软件会变得免费，但不久机器人就会变得很好，火箭也可能变得更容易。我喜欢这样的时代。当地面变化最剧烈、成本和周期时间迅速下降时，初创公司拥有最大的优势。现在很多地方都在发生这种情况。大多数初创公司说“我要为企业垂直领域 X 构建 AI 智能体”。这在很多情况下会成功，但它很可能不是这个时代最具定义性的初创公司。鉴于格局如此动荡，我很惊讶没有更多人利用全新的工具去挑战疯狂的事情，并真正内化缩放定律会继续的事实。他们应该为这个月不可能但 2 年或 4 年后可能的事情做计划。现在是一个不可思议的初创时间，但有一种诱惑是用今天的智能体去追求容易的胜利。

What it means to be a hard startup is changing so incredibly quickly that I don't have a perfect mental model for what will be really hard and valuable over many years. I've heard people say anything physical is extra valuable because software will become free, but soon robots will be good and rockets may get easier. I love times like this. Startups have the biggest edge when the ground is shifting the most, when costs and cycle times are dropping rapidly. That's happening in many places now. Most startups say 'I'm going to build AI agents for enterprise vertical X.' That will work in many cases, but it probably won't be the most defining startup of the era. Given how much is in flux, I'm surprised more people aren't taking on crazy things with new tools and truly internalizing that scaling laws will continue. They should plan for things that aren't possible this month but will be in 2 or 4 years. It's an unbelievable time for startups, but there's a temptation to go for easy wins with today's agents.

你曾说过，每次遇到新的人，你会试着在他们身上绘制你的心理模型，下次见面时看看他们进步了多少。在数千人身上这样做是否影响了你这样说的能力：“模型今天这么智能，三个月后那么智能”，然后你可以绘制出来并相信增长？

One thing you've said is that whenever you meet someone new, you try to plot where they are in your mental model, and the next time you see how far they've progressed. Did doing that across thousands of people impact your ability to say, 'The model is this intelligent today, this intelligent three months later,' and then plot it and believe in the growth?

我认为这两件事背后有一个共性：我对人、公司或模型的指数增长建立了深深的信任。主观上，观察模型的进步和观察创始人的发展感觉不同，但关于世界的基本信念是一样的。如果我现在还在给初创公司创始人做顾问，这是我最想让他们理解的事情。这显然很难，这就是为什么押注高增长年轻创始人还有“免费钱”留在市场上。市场还没有完全适应这一点，也没有适应模型进步的指数级持续。现在开始研究那些需要更智能或更便宜模型的事情是完全可以的。

I think there's a general thing behind both of those: I developed a great trust in exponentials in people, companies, or models. Subjectively, watching model progress doesn't feel the same as watching a founder develop, but the underlying belief about the world is the same. If I were still advising startup founders, this is the most important thing I would try to get them to understand. It's evidently hard, which is why there's free money in betting on high-growth young founders. The market hasn't fully adapted to that, nor to the exponential of model progress continuing. It's okay to start working on things now that require smarter or cheaper models.

这种不断变化的混乱，你怎么才能擅长在混乱环境中运作？

This idea of constantly shifting chaos, how do you get good at operating in a chaotic environment?

我认为这只是练习。有些东西不管你在理性上多理解，都需要很多次才能在情感上承受。在混乱中运作，相信你会搞定，这不会要你的命，而你目前还不知道怎么做但你会找到办法。只有亲身经历才能学会。我认为这是年轻创始人的一个真正弱点：他们没有职业经验来在这方面达到情感宁静。他们最终会学会，但付出巨大痛苦和代价。如果你最终从事高压、有影响力的工作，你肯定会学会应对混乱。但这是教不来的，只能学。

I think it's just practice. There are a handful of things that no matter how much you intellectually understand, they take a lot of reps to emotionally handle. Operating in chaos, trusting you'll figure it out, that this won't kill you, and you don't know how yet but you will. This is only learned by going through it. I think this is a real weakness of young founders: they haven't had the career experience to achieve emotional peace with it. They eventually learn it, but at great pain and cost. You definitely will learn to deal with chaos if you end up in a high-stress, impactful job. But it's not teachable, only learnable.

你曾说过，第一次经历公司生死攸关的事件时，感觉世界在崩塌。但到第十次时，就没那么糟了，你会想“我挺过了前九次，这次大概也没那么糟。”你还说你得出的结论是事情总会出错，所以你必须内化享受痛苦或不好事情的感觉。

You once said that the first time you experience a company-killing event, it feels like the world is falling apart. But after the tenth time, it's not as bad and you think 'I survived the first nine, this one's probably not that bad.' You also said you came to the conclusion that things are always going to go wrong, so you have to internalize the enjoyment of experiencing pain or not-great things.

你是如何实现这种转变的？我早些时候在听你说话时反思，我在 YC 做办公时间时，总能通过创始人面对糟糕事情的反应判断出他们是新手还是老手。我常常不知道公司有多新，但通过他们谈论问题时的情绪状态，我能分辨出这是当批的创始人还是已经做了几年的老创始人。我认为大多数人认为糟糕经历的反面是好经历，他们宁愿有好经历，因为那似乎更有趣或更愉快。但如果你换个角度想，糟糕经历的反面是没有经历，而在不远的将来，你会进入没有经历的状态。所以你也可以对糟糕经历心存感激。Naval Ravikant 曾说过一句我很喜欢的话：如果你的生活遥控器上有一个快进按钮，你的生命就会结束。所以无聊的部分、糟糕的部分，远比没有经历要好。它都是生活趣味性、情感深度和广度的一部分。我觉得对糟糕的日子心存感激是相当容易的。当你思考难题以及是什么激励人们兴奋地去攻克难题时，你如何在公司内部决定哪些事情要解决、何时解决，以及何时专注于核心能力，何时拓展范围？

How did you make that transition? I was reflecting earlier when you were saying this about how when I would do office hours at YC, I could always tell who the new founders were because of the way that they reacted to stuff going horribly wrong versus the veteran founder. I often didn't know how new a company was, but I could tell if it was a founder in the batch versus a founder that had been doing this for a couple of years by their emotional state when they talked about a problem. I think most people think about the opposite of a bad experience as a good experience and they would rather have the good experience because it seems like more fun or more pleasant. But if you think about it, the opposite of a bad experience is no experience, and at some point in the not-too-distant future you'll be in no experience land. So you can be grateful for the bad experience too. Naval Ravikant used to say this thing that I loved: if you had a fast forward button on a remote for your life, your life would be over. So the boring parts, the bad parts, are much better than no experience. It's all part of the interestingness and the emotional depth and range. I find it fairly easy to be grateful for the bad days. When you think about hard problems and what rallies people to get excited to go work on hard problems, how do you decide internally at the company which things to tackle and when, and when you want to focus on your core competencies versus expanding the scope?

我认为清晰的使命和对问题的深刻理解共同很好地指出了你应该做什么。你不会完全做对。在某些时候，你仍然会过度扩张或不够雄心勃勃。但我们非常专注于这一点：它将极大地赋能人们。它会起伏不定但很棒，对我们来说极其重要的是让世界上的权力更加去中心化和分散。事实上，我现在担心的最大 AI 风险之一就是 AI 威权主义。少数人或公司认为他们需要控制世界，那将非常非常糟糕。考虑到这个使命，我们觉得需要想办法让 AI 变得极其丰富、极其便宜、极其强大，把它交到每个人手中，并有很多可供使用。当我们审视实现这一目标的障碍时，有一大堆新的约束：芯片、能源、数据中心、机器人，以及所有需要整合起来构建这个平台的部件，以实现这一使命。你可以在这个平台上构建许多东西，比如每个垂直领域、每个创业公司。我们对此没有兴趣。我认为去中心化经济是重要且有益的，我希望我们只是非常擅长生产这些智能单元，让世界变得极其强大、极其划算，并将其注入每个产品和服务。所以我认为这很好地指导了我们下一步该做什么。有趣的是，实现极其丰富、高质量、低成本的智能的关键输入——比如能源和机器人——也是你在智能丰富的世界中真正需要的东西。如果想法丰富，有所有这些好东西，我们仍然生活在物理世界中，仍然希望事情发生。所以我们需要能够在这个世界上推动事情发生。我在想，能源和机器人对于保持在这条智能基础设施斜坡上如此重要，而且之后立即需要的东西也是能源和机器人，这是否有什么深刻含义，还是仅仅是最无聊的显而易见：要生产任何东西，包括智能，你都需要操纵物质。但我们现在如此需要这些新领域，之后又立即需要更多，这确实有点意思。

I think a clear mission and a deep understanding of the problem together do a fairly good job of pointing to what you should do. You won't get it all right. You'll still at some points overly expand or not be ambitious enough. But we are very focused on this: it's going to massively empower people. It's going to be choppy but wonderful, and it's extremely important to us that power in the world gets more decentralized and more spread out. In fact, one of the biggest AI risks I am worried about right now is AI authoritarianism. A small number of people or companies thinking they need to control the world would be very, very bad. Given this mission, we feel we need to figure out how to make AI extremely abundant, extremely cheap, extremely powerful, put it in everyone's hands, and have a lot of it to use. As we look at what's in the way of that, there are a whole bunch of new constraints: chips, energy, data centers, robots, and all the pieces that need to come together to build this platform to enable this mission. There are a bunch of things you can build on top of that platform, like every vertical, every startup. We have no desire to do that. I think a decentralized economy is important and good, and I would like us to just be really great at producing these units of intelligence and make the world incredibly capable and an incredibly great deal, and we're going to imbue it in every product and service. So I think that's very instructive to us about what we have to go do. Interestingly, the key inputs to extraordinarily abundant, high-quality, low-cost intelligence—things like energy and robots—are also things that you really want in a world of abundant intelligence. If ideas are plentiful and there's all this great stuff, we still live in a physical world and we still want stuff to happen. So we need to be able to make stuff happen in that world. And I have wondered if the fact that energy and robots are so important to staying on this intelligence infrastructure ramp, and also the things you need immediately after it are energy and robots, says anything deep at all, or is just the most boring obvious thing: to produce anything, intelligence included, you need to manipulate matter. But it is kind of interesting that we need these new areas so much for what we're doing now and then so much more immediately after.

你认为最大的瓶颈是什么？如果我们只想让这一切持续 Scaling 不减，你看到的最大瓶颈是什么？

What do you think your biggest bottleneck is? If we just say we want this to continue scaling unabated, what is the biggest bottleneck that you can see?

晶体管，然后是电子，按此顺序。

Transistors and then electrons, in that order.

假设你像 Jensen 一样，他非常擅长让所有供应商都配合他对未来的愿景。我认为你在这方面也非常厉害。你如何看待不仅让 OpenAI 保持在 OpenAI 的时间线上，并且也让世界其他部分也保持在 OpenAI 的时间线上？

So let's say you take someone like Jensen, who is incredibly good at aligning all his suppliers towards his vision for the future. I think you're also extremely good at this. How do you think about not only keeping OpenAI on the OpenAI timelines, but also keeping the rest of the world on OpenAI timelines as well?

你要和他们多沟通。你必须这样做，而且不能只是说‘我需要你在这一天交付这个涡轮或这个芯片’。因为人们会说‘好的，随便吧’，他们还有很多其他优先事项。你真的需要向他们展示：这是即将推出的模型，它能实现什么，这是我们的研究。对于我们的核心供应商，我们会向他们展示很多关于我们正在做的事情，我们为什么相信我们所相信的，它将实现什么。你要弄清楚——最重要的是让他们相信这个使命，明白为什么他们需要优先考虑它，但同时你也要尽可能将他们的激励机制与你的对齐。

You talk to them a lot. You have to, and you don't just say, 'I need you to deliver this turbine or this chip on this date.' Because people will say, 'Okay, whatever.' And they have a bunch of other priorities. You really have to show them: here is the upcoming model, here's what it enables, here's our research. For our core suppliers, we really show them a lot about what we're doing, why we believe what we believe, what it's going to enable. You figure out — the most important thing is to get them to believe in the mission and why they need to prioritize it, but you also try to figure out how to align their incentives with yours as much as you can.

Charlie Munger 有一句很棒的话：‘每次我以为我理解了激励的力量，我都低估了它。’你如何成功地将不同合作伙伴以及 OpenAI 内部人员的激励对齐？

Charlie Munger has this awesome line: 'Every time I've thought that I understood the power of incentives, I've underestimated them.' How do you successfully align all the incentives of the different partners as well as the people inside OpenAI?

嗯，公司是启动这一过程的绝佳载体。小时候，我对工业革命非常着迷，我把它理解为一系列恰好在同一时期出现的技术，并且出于某种令人费解的原因，它们几乎同时实现了规模化。我会和朋友或父母讨论这项技术是否是最重要的发明，或者那项技术。以我现在花了更多时间思考商业的视角来看，我认为重要的发明是股份公司。在那之前，只有基于信任的家庭企业，你必须认识每个人，没有公司或股东的概念。然后突然之间，这个新事物被发明出来了。世界各国的主权者授予了一种新的地位——一种以前不存在的东西。

Well, companies are amazing vehicles for this to start. As a kid, I was very fascinated by the industrial revolution and I understood it as a bunch of technologies that happened to come around the same time and for some puzzling reason happened to scale at around the same time. I would talk to my friends or my parents about whether this technology was the most important invention or that one. My read of it from my current lens as someone who has now spent more time thinking about business is that the important invention was the joint stock corporation. Before that was invented, you had family businesses that ran on trust, you had to know everyone, and you didn't have the idea of a corporation or stockholders. Then all of a sudden this new thing was invented. Countries, the sovereigns of the world, granted a new kind of status — something that didn't exist before.

我们允许这些新实体出现，不赋予它们国家的权力，但给予远超个人的力量。从根本上，这关乎激励对齐、责任保护——另一种激励对齐。能够积累资本，突然就可以做家族企业做不到的事。这种环境让人们筹集资金发展技术，带有高度投机性。汇集资本，想办法让不同公司以不同方式专业化并相互互动。严肃的金融体系和新型工具。这真是不可思议。公司的概念在协调庞大群体的激励方面做了大量工作。现在，数百万人可以成为公司股东。还有很多其他事情，但我不想贬低这个惊人的发明以及资本主义在人类社会中令人难以置信的超常表现。人们应该更多关注的一张图表是过去一百年极端贫困的下降、婴儿死亡率等所有相关指标。如果你回顾一下，这是过去一百年最聚焦的视图。但如果你看经济增长、生活质量改善等所有这一切，画出整个人类历史，在公司发明的地方画一条线，你会看到曲线形状发生了相当有趣的变化。除此之外，我认为我们有一个超越任何经济力量的使命，这也许更重要。

This very powerful thing that we're going to let these new entities emerge and not give them the power of a state but give them like much more than the power of a person. And fundamentally this was about you know incentive alignment, liability protection which is another kind of incentive alignment. Um the ability to like amass capital and all of a sudden you could do things beyond what a family business could do. And that environment let people raise money to develop technologies. Very speculative nature. Uh pool capital figure out ways to get different companies to sort of specialize in different ways and interact with each other. Serious financial systems and new kinds of instruments. This was really an incredible thing. Uh and the idea of a company has done a huge amount to align incentives of very large groups of people. Now, you know, millions of people can be shareholders of a company. There's a bunch of other things we do, but I don't want to diminish what has been this incredible invention and like the just ridiculous overperformance of capitalism in human society. Like a chart that I think people should look at much more than they do is the like fall of extreme poverty over the last hundred years like infant mortality and all the other things any of them that you want any of the associated things. Um you know if you like kind of look back at that that's like the most zoomed in view of the last hundred years. But if you look back at kind of economic growth, quality of life improvements, uh all of these things and you just like drew out all of human history and you put the line when like the company was invented, you would see like a really quite interesting change in the shape of that curve after it. Beyond that, I think we have a mission that transcends any amount of economic power anything else could ever do. Uh, and that's probably more important.

我觉得最有趣的一件事是，你有没有看过那种统计数据，比如 CEO 中反社会人格和精神病态的比例？非常高。

One of the things that I've thought is the most interesting is have you ever seen one of those stats where it's like I don't know what percentage of like CEOs are sociopaths and psychopaths? It's like very very high.

人们谈论过这个。

People talk about this.

所以我想知道，资本主义是不是第一个允许这些极度自私的人重新与社会激励对齐，只是创造伟大产品，然后你帮助社会，同时自己也赢的系统？

So I wonder if capitalism is the first system that allowed these people that are like super kind of self- serving to like their incentives are realigned with societies and just create great products and then suddenly you help society but you also win.

嗯，我当然认识很多反社会人格的 CEO，但我认识的最好的那些我不会归入那类。有很多自我极高的 CEO，人们对自己有难以置信的高评价，这可能纯粹令人讨厌，但现在因为人们可以买他们的股票，也许你更能容忍他们的烦人和古怪。所以肯定有一些对齐作用。我不知道。我认为很多大公司 CEO 根本不属于反社会人格。

Um, I definitely know a lot of sociopathic CEOs, but I would say the best ones I know I would not put in that category. There's a lot of very high ego CEOs, uh, people just like incredibly high opinions of themselves on a way that maybe like would have been just like purely annoying, but now because people can like, you know, buy their shares, maybe you like tolerate their annoyance and antics a little more. So, there's definitely some aligning thing there. I don't know. I think there are a lot of CEOs of very large companies that I would not put in the sociopath category at all.

你认为是什么驱动了那些人？

What do you think drives those people?

我的意思是，因人而异。显然，看到自己在游戏中能变得多好，或者每年能比过去进步多少，是很大一部分原因。这很有智力上的刺激。至于所有其他问题，就像在玩最有趣的战略游戏，那可以很有趣。

I mean, it's different for different ones of them. Like definitely seeing how good they can get at the game or like how much they can get better than themselves every year I think is a huge part of it. You know, it's like intellectually quite stimulating. Um for all of the other problems like you're kind of playing the most interesting strategic game and that can be fun.

如果你必须思考你今天运作的方式，与 10 年前如果处于完全相同位置时你会运作的方式，这两个人之间最大的区别是什么？

If you had to think about the way that you operate today versus the way that you would have operated 10 years ago if placed in the exact same position that you are currently what are the biggest differences between those two people?

嗯，回到之前的问题，驱动我最大的是，这是我能想象到的最有趣、最重要的事，我们现在正处于奇点，这就是那个时刻。10 年前，那顶多是一个遥远的梦，看起来非常不可能。现在我们实际上处于我们过去在午餐桌上非严肃地讨论的那一刻。所以驱动我的东西是我一生都在等待这个，我认为它会对世界产生不可思议、极其积极、很棒的影响。我很兴奋能投身其中。我也认为其他公司描绘的一些替代愿景相当可怕。我要确保与之对抗，不让它发生。但与 10 年前相比，主要区别是我们真的身处其中了。这是真实的。我认为既可以说这一切都是一个疯狂的指数增长，任何时刻都不是转折点，同时我们又处于另一个决定性时期，曲线可能走向一方或另一方，就像 10 年前我们开始时那样。

Well, on the to the previous question, the thing that drives me the most is like this is the most interesting important thing I can imagine doing and we are now like in the singularity like this is the moment for the la 10 years ago this was like a kind of far off dream at best. Um seemed very improbable and now we're like actually in the moment that we used to like talk about at the lunch table in a very not serious way. Uh, and so the thing that drives me is I've been waiting for this my whole life and I think it's going to be incredible, hugely positive, awesome for the world. Um, I'm excited to get to work on that. I also think some of the alternative visions painted by other companies are like quite terrifying. I'm going to make sure that gets pushed against and is not what happens. Um, but the main thing of what's different than 10 years ago is like we're actually in it. like this is the real I think it is both true that it is all one crazy exponential and any one moment is not like the tipping point and also that we are somehow in another one of those decisive periods where the curve can go one way or another like it was when we started 10 years ago.

你认为让曲线的这个部分走对的最大影响是什么？

What do you think the biggest impacts are on getting this part of the curve right?

我认为仍然有重大的 AI 对齐问题需要解决，安全问题和重大的经济问题、工作未来等等。但我认为当下的斗争是：我们要走向 AI 威权主义还是自由？我们要因为真实的安全和经济问题，决定让单一模型成为机器神或相关公司来做这件事，还是说：你知道吗，可能会有点混乱，但每次人类用自由换取安全，长期来看都是净损失，所以我们将这件事交到人们手中，赋予他们权力。我们将让社会表达想法，按自己的方式使用这项技术。当然，要设置一些护栏，但拥有大量权力和潜力。我真的很关心我们这样去做。

I think like there are still major AI alignment issues to solve and safety issues and there are still major economic issues, future of jobs, all of that kind of stuff. But I think the fight of the current moment is are we going to head to a world of AI authoritarianism or liberty? Are we going to decide that because of the very real safety issues and economic issues that you know we want one single model to be the machine god or the company associated to do that or are we going to say like you know what it's going to be it may be a little messy but every time that humanity has traded off its liberty for safety it's been a long-term net loss and so we are going to put this in the hands of people we're going to empower power them. We are going to let society express its ideas and use this technology in the way they want. Of course, according to some guard rails with some guardrails, but you know, with a lot of power and potential. Uh, and I really care that we do it that way.

几个月前你和 Patrick Hollson 在台上对话时，你说了一件很有意思的事：你每天内部会和 300 到 400 人交谈或发短信。

When you were talking with Patrick Hollson on stage, must have been like a couple months ago, there was this fascinating thing that you said, which is basically like you will talk or text with 3 to 400 people a day internally.

我发短信的人比我认识的任何人都多。我不认为这是好事，实际上是坏习惯。

I text more people than anybody else I know. I don't think this is like a good thing. I think this is actually like a bad habit.

为什么？

Why?

因为它占用了大部分时间，我现在注意力持续时间很短。

Um because it takes up like most of my time and I get like into I have a very short attention span now.

有意思。那让你能做到什么否则做不到的事？

Interesting. What does that enable you to do that like otherwise?

大量的上下文。我的意思是，我就像一个不太聪明的 AI 模型，但对当下发生的事情有大量的短期上下文，然后忘记。那是否让你能做出别人做不到的决策？比如公司决策更好？我不知道。它让我做出和别人不同的决策。

Huge amount of context. I mean I'm like a AI model that's not that smart but just has like a huge amount like a shortterm context on like whatever's happening right now and then forget. Does that enable you to make decisions that other people wouldn't have? Like what company decision better? I don't know. I mean, it lets me make different decisions on other people would.

过去 12 个月里，如果你没有那么多的上下文信息，你会做出哪些不同的决策？

What different decisions would you have made in the last like 12 months that you wouldn't have made if you didn't have so much context?

我不知道有没有特别具体的例子能公开分享，但有很多时候，因为我知道某个即将到来的研究突破会影响到客户或供应商之类，我就能在当下做出稍微不同的决定。如果那天早上我不知道，我就会做出不同的决定，结果会更糟。

I don't know if I have like a specific one I can share publicly, but there's like a lot of things where, you know, because I know about some random research breakthrough impending that is going to like affect some customer or some supplier or whatever, I can make a slightly different decision in the moment. And if I hadn't known it that morning, I would have made a different decision. It would have been worse.

我觉得你最擅长的之一就是长期思考，但似乎事情发展得太快，世界变化太快，以至于规划 20 年后然后回头看，我甚至不知道这有没有意义。你怎么看待这个？你的时间定价在缩短吗？

One of the things that I think you are the best at is kind of like long-term thinking and yet it also seems like things are moving so quickly and the world is changing so fast that planning 20 years out and then like looking backwards I don't even know if that makes sense. Um how do you think about this? How is like is your time pricing shortening?

我其实不会试图向后规划。我会对未来抱有少数几个坚定的信念，把它们当作前进的方向，然后从当前状态向前规划，思考我们现在能做什么、今年能做什么。有时你也需要规划一些 5 年或 10 年的事情，但我是在少数信念的引导下向前规划。我觉得很多人对未来有太多信念，抱着一种僵化的世界观试图套用，结果他们最终去追逐，比如看到太空公司变成 AI 公司之类的。只对未来深信少数几件事，其余保持灵活，并且坚持核心，这很有帮助。

I actually don't try to plan backwards. I try to have like a a small number of strongly held convictions about the future that are like directions to head towards but then I try to like plan forwards from the current state about like what can we do now what can we do this year and sometimes you do have to like plan a few things on a you know 5 or 10 year horizon but I try to plan forward guided to a small number of beliefs about the future I think there are a lot of people who have way too many beliefs about the future and a kind of rigid worldview that they try to fit and then but what happens is they end up like chasing, you know, you see like space companies turn into AI companies or whatever. Um, and like having just a small number of things that you deeply believe about the future and being very flexible on the rest. Um, and staying very true to the core is is helpful.

我最了解的一个人，他的公司核心价值观就是关键路径，只专注于关键路径，专注于任何最大障碍的关键驱动因素。搞定它，然后解决下一个，如此反复。你在自己做这些决定时，是如何看待关键路径的？

One of the people that I know the best, his top, you know, company core value is just critical path and just focusing on the critical path and like staying focused on the key drivers of whatever the biggest roadblock is. just unfuck that and then go to the next one and do that again. How do you think about critical path in your own life when you're kind of like trying to make these decisions?

很久以来，我一直专注于这个单一目标：实现丰富的智能（abundant intelligence），并坚信只要没有奇怪的权力集中和新型威权主义，人类将因此获得难以置信的繁荣。在任何时候，我都很清楚通往那里的关键路径是什么。我从来没有被诱惑去考虑是否应该重新审视目标或思考其他事情。但最近我稍微多想了一些：如果我们真的接近超级智能（superintelligence），接下来是什么？这是我十多年来第一次感觉我们接近了，并思考下一步。

I mean, for so long now, I have felt focused on this singular goal of abundant intelligence and a belief that like incredible human prosperity will come from that as long as we don't have a weird power concentration and kind of a new kind of authoritarianism. It has been fairly clear to me at any given time what's on the critical path to get there. And I have not been tempted by, you know, should I reconsider the goal or should I think about these other things. I have been thinking about that a little bit more recently of like, okay, if we really are close to superintelligence, what's next? But but it's this that's kind of how I subjectively feel we are close is the first time in like more than a decade I've thought about what is the next thing.

就是那个牧场吗？

And is it the ranch?

呃，最终可能途中还会有几件事，但没错，那是最终计划。

Uh I mean eventually there might be a few more things on the way, but that is yes, that is the eventual plan.

如果你必须猜一猜，Sam 在去牧场之前还想完成什么？

If you had to like try to guess what else does Sam want to accomplish before he ranches.

嗯，你看，AI 超级智能一定会被构建出来，但还没完成，我们正在滑行轨道上，通往那里的路就是富足、去中心化、广泛访问，确保最终实现广泛共享的繁荣。我真的很想实现这一点。所以也许这就是我一直在思考的下一件事：技术上我们如何完成使命。我认为我们的很多世界观以及关于安全和影响的信念也都已经实现了。但你知道，非常广泛的共享繁荣，以及真正专注于赋能世界，这似乎非常重要。

Um look AI superintelligence is going to get built like that you know not done but like going to we are we are on the glide path path to the get there is very abundance and decentraliz and like wide access to it and making sure that it ends up with like broadly shared prosperity. I would really like to accomplish that. So maybe that that's the thing that I've been like thinking about as the next thing like how does technologically we're going to accomplish our mission. I think a lot of our worldview and a lot of our beliefs about safety and the impact is going to have have also we've gotten those things done too. Um but you know very broadly shared prosperity and uh a real focus on enabling the world here. Um that seems very important.

我反复听你说过这个想法：在临界情况下上飞机。我为你做过很多次了。

I've heard you say again and again uh this idea of get on planes in marginal situations. I've done this a huge number of times for you.

呃，到目前为止效果很好。不错。我觉得这是非常可靠的建议。

Uh so far it's working pretty well. Good. I think it's very solid advice.

你最后一次在临界情况下上飞机是什么时候？

When was the last time that you got on a plane in a marginal situation?

嗯，我不能说是为什么，但就在最近，而且结果很好。

Well, I can't say what for, but very recently and it worked out.

恭喜。

Congratulations.

谢谢。我其实真的不想去。那是一个非常不方便的两次过夜行程，不想去。新生儿，一团糟，但我还是去了。

Thank you. And I really didn't want to. It was like a very inconvenient two overnights. Didn't want to do it. New baby, whole mess, but I did.

好吧。实际上，带我进入你决定做这次世界巡演的心理过程。我是说，很难回到 3 年前，因为现在感觉 AI 好像一直存在，但 3 年前我们刚刚推出 GPT-4。世界在崩溃，每个人都在恐慌。大家都觉得天哪，没人知道该想什么。我能感觉到风暴正在酝酿，人们变得越来越愤怒，或者世界领导人都在说，我们是否需要控制并关停它？这东西是不是觉醒了？那是一个奇怪的时期，因为一切都发生得太快，人们没有时间消化。所以我们不断收到升级的消息，很多世界领导人都说，你能来见我们吗？我们正在考虑做这个那个。我当时想，如果我不去，或者公司里其他人去，但公司当时不大，所以很可能得我去，我不去露面跟人们谈谈。我有种感觉事情会变得非常糟糕，我考虑做一系列短期旅行，但我讨厌长途飞行，讨厌时差，讨厌所有这一切。所以我想，干脆在短时间内搞定它。Brian Chesky 建议我这么做。他为 Airbnb 做过类似的事，大概去了 8 个城市。然后你就说：“去他的，我们上吧，直接做 30 个。” 我们最后去了 28 个国家，35 天。我就像住在飞机上一样，非常奇怪。你有床吗？我有，我很舒服，但我有床。我坐经济舱，所以有点不同。这比那个好多了，对吧？但你知道，旅行还是很糟糕。你可以让它变得尽可能舒适，但你仍然想念自己的床、自己的时区、自己的办公室等等。我学到的一件事是，无论有多少人有时差疗法，它就是很糟糕。你的大脑不喜欢。我从来没有因为去了太多地方而适应。

Okay. actually take me into your like mental process for deciding to do this world tour. That is just I mean it again it's like hard to rewind 3 years cuz now like it feels like AI has been here forever but 3 years ago we had just launched GPT4. The world was melting down. Everybody was freaking out. Um everybody was just like oh man. And no one knew what to think. And I could just feel these like storm clouds brewing and like people just getting angry or it was like world leaders were like do we need to like take control and shut things down? Like is this thing waking up and about it was it was a weird time and because it had all happened so fast people had not had time to like wrap their head around it go through this process whatever. Uh so you know we were like getting all of this escalating stuff and all these like world leaders were like will you come meet us we're like you know thinking about doing this and that and I was like uh I think if I or someone from the company but at the time company was not very big so I was like probably going to have to be me does not go like show up and talk to people. I had like a sense it was about to go very badly and I was thinking about doing like a bunch of short trips but I hate long flights. I hate jet lag. I had the whole thing. So, I was like, I'm just going to like get this over with in a short period of time. Um, Brian Chesky advised me to do it. He had done something for Airbnb, similar for Airbnb, but he had done like eight cities or something. And you were like, "Fuck it, we ball. Let's just do 30." I think we did 28 countries, 35 days. Uh, I like lived on an airplane. It was a very strange. Did you have a bed? I had like a Yeah, I was comfortable, but yeah, I had a bed. Um, I fly coach, so it's a little bit different. This is better than that, right? This is a lot better than that. But like, you know, it's like travel still sucks. Like you can make it as comfortable as you want and you still like miss your own bed and your own time zone and your office and whatever. One thing that I've learned is no matter how many people have like these cures for jet lag, they're all like it just sucks. Your brain doesn't like it. I never in because I went to so many places.

我觉得整趟旅行最大的时差变化也就 4 小时，而且大多是一次 1 小时的跳跃。所以我不太有倒时差的感觉，只是很累，而且很奇怪。能这么快看世界还挺酷的，因为你真的能感受到从一个地方到另一个地方时那些很细微的文化差异。但到行程最后几天，我开始有那种半梦半醒的奇怪梦境，感觉自己回到了童年时的床上。

I think my biggest time zone change was 4 hours the whole trip and mostly it was like a 1 hour at a time hop. So, I wasn't that jetlagged. I was just exhausted and it was very strange. It was kind of cool to get to see the world so quickly because you really do get a sense for cultural differences that are very subtle when you go from place to place. But by the end of it, like the last few days, I started having this weird, half asleep, half awake kind of dream. I started having this vision or feeling of being in my childhood bed.

有意思。

Interesting.

那之后我再也没有过那种感觉。我会迷迷糊糊醒来，以为自己还躺在童年房间的床上。我就想，'好吧，这大概是某种'该回家了'的深层信号。'

Which I never had before since. But I would sort of vaguely wake up and think I was in my childhood bed in my childhood room. And I was like, 'Okay, this is some deep it's time to go home thing.'

就像是潜意识在反抗什么的？

You're like subconscious fighting back or something.

差不多吧。

Something like that.

有意思。当你处于'僵尸模式'，还在努力运转，或者用你喜欢的话说'哥布林模式'，你会有哪些变化？你有什么不同？你怎么决定继续前行？

Interesting. When you're in zombie mode and you're kind of trying to function, or goblin mode, whatever you prefer, how do you change? What is different about you? How do you decide to keep on going?

到最后几天我确实到了数着日子等回家的地步。每天差不多 14 小时连轴转，而我不是一个外向的人。有时候一天要见几百个人。紧张感在后期有所缓解，但一开始整个氛围都非常紧张。一旦平静下来，有帮助的一点是我知道自己什么时候能回家，也知道接下来还得再熬一阵子。后来我也做过一些短途旅行，大概一年一两次。我学到的是，我宁愿把一堆国际旅行集中到 7 到 10 天内完成，而不是分散成很多次一日游。所以我一直都尽量这样安排。

I definitely got to this point near the end where I was hunting down the days till I could be home. Each day was about 14 hours nonstop and I'm not an extroverted person. I was meeting hundreds of people a day in some cases. Tensions lowered by the end, but at the beginning the world was very nervous. Once it calmed down, one thing that helped was knowing when I was going to get home and knowing I would have to do it again for a while. I've done smaller trips since, probably once or twice a year. I learned that I would much rather put a bunch of international travel together in a 7 or 10 day chunk than a bunch of one day trips spread out. So I always do try to do it that way.

我觉得很多人高估了大多数行动所涉及的风险。如果你只是在 Robinhood 上买一堆看涨期权，那确实风险很大，但坐飞机之类的风险可能就小得多。

Now I think a lot of people overestimate the risk involved in making most actions. There is a lot of risk if you're just buying a bunch of call options on Robinhood, but there's probably a lot less risk on getting on planes and stuff.

有哪些最好的例子，你做的事情在别人看来风险很大，但你知道并非如此，或者你相信风险不大？

What have been the best examples of times where you basically did something that seemed really risky to other people, but you knew wasn't or you believed wasn't?

我觉得最明显的例子就是我总是非常确信地大量购买算力。

I guess the obvious example would be that I was always extremely high conviction buying a lot of compute.

你有没有看过那个梗图：'一切都有高风险，如果你是个……'

Have you ever seen the meme everything's high risk if you're a...

没有，好梗。我花了很多时间试图和人们谈论，为什么他们认为某个决定是高风险或低风险的。我发现，只要让他们大声说出来，往往就能突破他们在这方面的思维障碍，因为人们通常总有一个方向是错的。不总是情绪部分，但有时也包括情绪。

No. Good meme. I spend a lot of time trying to talk to people about why they think a specific decision is high risk or low risk. I have found that just getting people to speak about it out loud will often get through their intellectual blocks in either direction because people are usually wrong one way or the other. Not always the emotional part of it, but sometimes also.

如果你回想一次这样的会议，某人坚决反对你想做的某个决定，你用什么最有效的方式帮助他们看到你所看到的？

If you had to think through one of these meetings where someone is firmly against some decision that you want to do, what is the most effective way to help them see what you see?

说实话，我不太擅长这个。正确的做法是花很多时间解释，让别人理解。我以前更擅长。但随着生活变得忙碌，我更多是变得烦躁或沮丧，然后说'我们就这么做吧'。我不认为这是个好特质。我希望回到那种真正把事情谈透的状态。

I don't think I'm very good at this to be honest. The right answer is to spend a lot of time trying to explain it and get people there. I used to be better at that. As life has gotten so busy, I'm more just like I get upset or frustrated and say, 'This is what we're going to do.' I don't think it's a positive trait. I would like to get back to more of the let's really talk it through.

你说过你以前更难以共事，大概在 2008、2009、2010 年左右，比现在更难。

You said that you used to be much harder to work with, I don't know, 2008 910 somewhere around there than you are today.

我不知道人们会不会用'共事愉快'来形容我。我倾向于认为共事起来有趣、高效、能激发雄心壮志、非常专注于使命，但日常真的有趣吗？我觉得自己总体上不是一个特别有趣的人。

I don't know if people would describe me as fun to work with. I like to think interesting to work with, effective, ambition raising, very mission focused, but day to day really fun? I don't think I'm a particularly fun person in general.

我觉得如果你自己享受其中，那种情绪是会传染的。而且如果你通常在做自己想做的事，那本身就是一种享受。

I feel like if you are having a good time, it tends to be infectious. And if you are generally working on things you want to work on, it is having a good time.

也许我错了。你可以去问问我合作过的很多人。我不觉得你会得到清一色的'共事愉快'这样的评价。

Maybe I'm wrong. You could go ask a bunch of people I work with. I don't think you would get a resounding 'fun to work with'.

可能不是这个措辞。

Maybe not that language.

我觉得人们会说一些像'创造性的问题解决方案''新的见解''非常高的雄心壮志'，比如'让我做到了我以为自己做不到的事'之类的。

I think people would say stuff like creative solutions to problems, new insights, very high level of ambition, like got me to do something I didn't think I could do, that kind of thing.

当你和某人开会时，发现他们不够有雄心，你需要帮助他们变得更有雄心，典型的情况是什么样的？

What does it typically look like when you go into a meeting with someone and you realize that they are not ambitious enough and you need to try to help them become more ambitious?

我想到的是和新 YC 创始人第一次办公时间。你看到一个在美国大公司工作过几年的普通人，他们的雄心壮志、规模思维、自信等等，都低得离谱。这个人一生中从未没有过老板，从未没有过父母、老师或经理告诉他们该做什么或允许做什么。他们还因为太有创意、想要太多或想得太大而受到惩罚。大多数国家或文化都有某种说法来形容'过于雄心勃勃'。

The version that came to mind was first office hours with new YC founders. You take an average person who's worked in corporate America at a big tech company for a few years. The level of ambition, scaled thinking, self-belief, whatever, it's catastrophically low. This person has never not had a boss in their life, never not had a parent or teacher or manager telling them what to do or what they were allowed to do. They also got punished for being too creative or wanting too much or thinking too big. Most countries or cultures have some sort of phrase for being too ambitious.

高罂粟综合征。

Tall poppy syndrome.

对，就是那个。你会意识到这在人们心中有多根深蒂固。他们从小就从父母、老师、朋友那里听到这种话。让人们不再那样想需要花不少时间。

That's one. You realize how deep this is in people. They heard it from parents, teachers, friends from when they were quite little. Getting people to just not think that way takes some real time.

让一个人从不相信自己到相信自己，最有效的路径是什么？

What's the most effective path to get someone from not believing in themselves to believing in themselves?

小规模的重复成功。你做了自己原本认为做不到的事，或者尝试了觉得过于雄心勃勃或期望过高的事，而且成功了。然后你对自己说，'好的，再试一次。'

Small repeated wins. You do something you didn't think you could do, or try something that felt too ambitious or too high expectation, and it works. Then you say, 'Okay, try it again.'

你有没有像史蒂夫·乔布斯那样，盯着别人的眼睛说……不，没有那样的。

Do you do anything like Steve Jobs where he would stare into their eyes and say... No, nothing like that.

我不觉得有。

I don't think so.

就像是'你能行'那种？

It's like you can do it.

我不做励志演讲，也不做强势推销，那些都不做。

I don't do inspirational speeches. I don't do the hard sell, none of that.

你最喜欢的策略游戏是什么？

What's your favorite strategy game?

除了和朋友打扑克之夜，我很久没玩过游戏了。真希望我玩过。

I have not played a game other than my poker night with friends in so long. I wish I had.

我想我只能说扑克了，但我很怀念那些好玩的桌游。

I guess I'll have to say poker, but I miss great board games.

你最喜欢的桌游是什么？

What was your favorite board game?

没有哪一款桌游是我会想反复玩的。我不是那种会说“卡坦岛是有史以来最好的桌游”的人。但所有桌游都很有趣，因为它们是新的，你必须去搞清楚一套新的挑战和规则等等。

There's no one board game that I want to play again and again. I'm not someone who says Settlers of Catan is the best board game of all time or whatever. But they're all fun because they're new and you have to figure out a new set of challenges and rules and everything else.

在商业世界里，我认为你构建公司的方式几乎与任何其他人所能或愿意做的都不一样。是什么让你能做到这一点？

When you're thinking about the business world, I think you build your company differently than almost anyone else is able or willing to. What allows you to do that?

我其实觉得每家大型公司都很不一样。我不认为这是独特的。如果你列出前 20 大科技公司，它们在战略、运营和文化上都惊人地不同。我们确实与众不同，但这并不是什么深刻的见解。

I actually think every big company is pretty different. I don't think this is unique. If you make a list of the 20 biggest tech companies, they're all shockingly different from each other in terms of strategy, operations, and culture. We are quite different, but that's not a deep insight.

你认为你的超能力是什么？

What do you think your superpower is?

我认为我们做得非常好的一点是，在某个没人相信、但显然是应该相信的事情上抱有原则性的信念，然后围绕它整合所有资源和人才，使之成为现实。

I think we did a really good job of having principled conviction on something that no one else believed, even though it was the obvious thing to believe, and then putting together all the pieces and talent around that to make it happen.

是什么让你拥有那种别人没有的信念？

What allowed you to have that conviction where other people just didn't?

我想过很多。我不知道。这对我们所有人来说都如此显而易见，以至于我更担心我们在自欺欺人，而不是别人都错了。但事后看来，我们显然是对的。我不理解其他人那种深层的思维障碍和信念缺失。这很奇怪。我花了很多时间试图理解这一点。

I have thought about this a lot. I do not know. It seemed so clearly obvious to all of us that I was more worried we were drinking our own Kool-Aid than that everybody else was wrong. But with hindsight, we were clearly right. I don't understand the deep mental block and lack of conviction from everybody else. It's very strange. I've spent a lot of time trying to understand this.

你有没有在，你知道，《活出生命的意义》里找到什么解释？还是没有？

Did you find any explanation in your, you know, 'Man's Search for Meaning' or no?

有那些显而易见的东西，比如群体思维、重大变革不常出现、指数曲线不会长期保持强劲。但也许在 2015 年到 2018 年之间，我们不得不非常幸运，这需要很大的信念，但到了 2019 年，肯定到 2020 年，我们本不应该还存在。谷歌本应轻松取胜。杰夫·贝佐斯称他们构建 AWS 并在七年内没有遇到激烈竞争是一个商业奇迹。我觉得 OpenAI 是另一个商业奇迹。显然，大公司为何变得僵化、固步自封，这也在很多方面对它们有帮助。但大公司不会永远保持唯一的主导力量，这很好。那会非常糟糕。但显然，这些商业奇迹发生了，而它们本不该发生。

There's all the obvious stuff about groupthink and big transformations not coming along often, and exponential curves not staying powerful for long. But maybe between 2015 and 2018 we had to get really lucky and it required a lot of belief, but by 2019, certainly by 2020, we shouldn't have really gotten to exist. Google should have just run away with it. Jeff Bezos called it a business miracle that they built AWS and didn't have serious competition for seven years. I feel like OpenAI is another business miracle. There's clearly something about why big companies get sclerotic and set in their ways, which helps them in many ways too. But it's wonderful that big companies don't stay the only dominant forces forever. That would be really bad. But there's clearly something where these business miracles happen and they shouldn't.

那是正确的吗，还是它们应该发生？

Is that right, or should they?

嗯，我认为它们发生对世界是好事。

Well, I think it's great for the world that they do.

对吧？但回头解释 2019 年为什么 OpenAI 被那些拥有巨额资本、人才等一切的巨头允许做我们所做的事情——这很难解释。

Right? But to go back and explain in 2019 why OpenAI was allowed by the giants with huge amounts of capital and talent and everything else to do what we've done — it's hard to explain.

为什么微软给了你们钱？

Why did Microsoft give you the money?

我认为那更容易解释：谷歌有 DeepMind，而微软没有在人工智能上的赌注，而且大公司觉得它们应该在所有主要技术领域都有所布局。

That I think is easier to explain: Google had DeepMind and Microsoft did not have an AI bet, and big companies feel like they should have something in play in all the major areas of technology.

是不是有点像他们觉得这会很大，但低估了会有多大？

Was it kind of like they thought it was going to be big but underestimated how big it would be?

我认为这对他们来说很棒。我不知道我们给微软增加了 1 万亿、2 万亿还是多少美元的市值，但我认为数额很大。而且我认为他们获得了不可思议的技术和云增长，并将继续下去。所以是的，我确信这比他们预想的要大，但我认为他们下了一个好赌注，并且做得非常好，我对此非常高兴。

I think it's been great for them. I don't know if we've added one trillion, two trillion dollars or whatever of market cap to Microsoft, but I think it's been a lot. And I think they have gotten incredible technology and cloud growth and will continue to. So yeah, I'm sure it's bigger than they thought it was going to be, but I think they made a good bet and have done super well with it and very happy about that.

OpenAI 有过不同的版本。第一个是研究实验室。下一个是产品公司，你曾说过你把产品公司嫁接在研究实验室上，这与大多数公司做的完全相反，而现在你基本上要成为一家大型基础设施公司，OpenAI 几乎所有的价值都将来自这种低利润的基础设施。你还说过，为了过渡到这个业务的下一阶段，你的大脑运作方式并非天生就适合那种模式。你必须改变什么？业务的下一阶段是什么？

There's been different versions of OpenAI. The first one was the research lab. Next one was the product company and you kind of said that you bolted on the product company onto the research lab which is the exact opposite of what most companies do, and now you're basically going to be like a massive infrastructure company and almost all the value from OpenAI is just going to come from this low margin infrastructure. And then you also said that in order to transition into the next phase of this business it's not necessarily something the way your brain works is naturally designed to operate in that kind of model. What do you have to change and what is that next phase of the business?

我不回答这个问题，因为我认为这涉及到我们还没准备好谈论的事情。但我对下一阶段非常兴奋，我想我已经找到了如何将我非常擅长和热爱的事情与我们未来 10 倍或 100 倍的增长结合起来。

I don't answer this because I think this is getting into something we're not quite ready to talk about. But I am very excited for the next phase, and I think I have figured out how to align something I am very good at and very passionate about with what will be the next 10x or 100x of our growth.

从研究实验室负责人 Sam 到产品公司负责人 Sam 的转变是如何发生的？对你来说最大的转变是什么？

How did the transition from Sam the research lab head to Sam the product company head happen? What were the biggest transitions for you?

整个过程非常不同。感觉像是两个几乎完全无关的工作。并不是说我完成了前一个现在做后一个，因为我仍然负责研究实验室。但它们几乎是分开的，而运营研究实验室并没有让我为运营产品公司做好任何准备。

The whole thing was extremely different. It felt like two almost completely unrelated jobs. It's not that I did one and now I do the other, because I still am responsible for the research lab. But they are almost separate things, and running the research lab did not prepare me for running the product company in any way.

YC 的经历是否让你准备得更充分一些？

Did the YC experience prepare you more?

我曾目睹很多人不得不扩大公司规模。

I had watched a lot of people have to scale big companies.

终于坐上驾驶座有趣吗？

Was it fun to finally get in the driver's seat?

不好玩。ChatGPT 上线后的第五天，我们突破了 100 万用户。我看到它第一天上升，下午回落；第二天上升到更高的峰值然后回落；第三天上升到更高的峰值然后回落。每一天研究人员都说：‘哦，那只是一次疯狂的昙花一现的公关事件。结束了。’但我在 YC 见过足够多的情况，知道当某样东西完全有机地那样增长时，它具备了所有即将爆发的特征。然后在第五天，它达到了 100 万用户。我回到家，才真正意识到我们将要成为一家公司，而且会发展得非常快，我们很快就会变成一家大公司。我见过创始人经历过的。我回到家，Ollie 说：‘哦，我看到你们突破了 100 万用户。’

No. The fifth day after ChatGPT launched was the day we crossed a million users. I had watched it go up the first day, then come down in the afternoon, then go up the second day to a higher peak and come down, go up the third day to a higher peak and come down. Each day the researchers were like, 'Oh, that was some crazy flash-in-the-pan PR thing. It's over.' But I had seen enough at YC to know that when something grows like that completely organically, it had all the spectral signature of something that was going to happen. Then on the fifth day, it crossed a million users. I came home and it really hit me that we were about to become a company, and it was going to go very fast, and we were just going to turn into a big company real quick. I had watched what founders went through. I came home and Ollie was like, 'Oh, I saw that you crossed a million users.'

“恭喜。”我当时想，“你根本不知道这有多糟糕。你不知道将要发生什么。这不仅仅对我不好，对你也不好。我们过着这么安静美好的生活，真的很好。但它马上就要经历一场炮击。”而事实正是如此。这是一次非常突然的转变。我知道它会发生，我也目睹过其他人经历这一切。我知道那不是什么愉快的体验，但我想，“我们正在被大炮发射。开始吧。”你是否有在潜意识里准备了多年，因为你知道最终智能会变得足够好，以至于会出现一个神奇的产品？

Congratulations. I was like, 'You have no idea how bad this is. You have no idea what's about to happen. It's not just bad for me, it's bad for you, too. We have this nice quiet life, it's really wonderful. It's about to go through a cannon.' And that is what happened. It was a very abrupt transition. I knew it was going to come, and I had watched other people go through it. I knew it wasn't a pleasant experience, but I thought, 'We're being shot at the cannon. Here we go.' Were you subconsciously preparing yourself for years, knowing that eventually the intelligence would get good enough that there would be a magical product?

我的看法是，我在理智上知道它迟早会发生。我不知道具体时间。当然，我也不知道那就是那一刻。但我决定自欺欺人，并且成功地说服自己假装它不会发生。说实话，经营这个研究实验室是我能想象到的最酷、最有趣、最了不起的工作和生活方式。

My take on it is that I intellectually knew it was going to happen at some point. I had no idea when. And I didn't know that was going to be the moment, of course. But I had decided to pretend to myself, and I had successfully deluded myself into acting like it wasn't going to happen. Truly, running the research lab was the coolest, most fun, most amazing job and lifestyle setup I could possibly have imagined.

怎么说？

How so?

我的意思是，它一点也不压力山大。在智力上，它极其令人满足。那是最聪明的一群人，做着可能是上个世纪甚至更久以来最重要的工作。而我就坐在前排。那是几代人一遇的时刻。难以置信。那是最酷的事情。

I mean, it was not stressful at all. It was intellectually incredibly satisfying. It was the smartest group of people with probably the most important work that has happened in the last century, maybe longer. And I had this front row seat. That was a once-in-many-generations moment. It was unbelievable. It was the coolest thing.

你们也经历了同样的增长吗？我总看到 Tibo 在推特上发帖，说又来了一个重置，又一个重置。你们是否又看到了那种类似 ChatGPT 的体验？

Are you guys experiencing that same kind of growth? I keep on seeing Tibo post on Twitter that basically here's another reset, here's another reset. Are you seeing that same ChatGPT-esque experience again?

完全认同。我认为到目前为止，有两个巨大的形态因子及其相关的增长。一种是聊天机器人，然后是编码智能体。编码智能体正在疯狂发展。我认为很快就会有第三个，那就是持久智能体、幕僚长、同事，不管我们怎么称呼它们。我认为这很快就会到来。所以我们将很快经历第三波浪潮。

Totally. I think there have been two giant form factors and associated growth so far. There were the sort of chatbots, and then the coding agents. The coding agents are just going totally nuts. There will be a third one soon, I think, which will be this idea of the sort of persistent agents, chiefs of staff, co-workers, colleagues, whatever we call them. And I think that'll come pretty soon. So we're going to go through the third of these waves pretty fast.

很久以前，大概是去年八月，我看到你发了一条推文，我心里想，这对名声可没好处。嗯，让我想想。为什么？你为什么那么做？

So, a long time ago, apparently in August last year, I saw you tweet something and I thought to myself, this is not going to help the reputation. Um, let's see. Why? Why did you do that?

首先，我喜欢电影里的那个场景。死星从超空间跳出来，伴随着戏剧性的音乐。那是一个很棒的场景。我觉得那很好笑。我当时在刷推特，大概是深夜，我不太记得了。

Well, first of all, I love that scene from the movie. The Death Star is coming out of hyperspace and there's this dramatic music. It's a great scene. I thought it was funny. I was scrolling Twitter. I think it was late at night. I don't really remember.

这就是阴谋论产生的地方。你这是在火上浇油。

This is where the conspiracy theories happen. Like you're feeding the fire here.

我只是觉得当时它让我觉得好笑。老实说，这不是一条好推文，明确一下。

I just thought it amused me at the time. It was not a great tweet, to be honest, to be clear.

我其实觉得那很棒。

I actually thought it was great.

我觉得那很好笑。

I thought it was funny.

我觉得那很好笑。

I thought it was funny.

我想我觉得那很好笑。

I think I thought it was funny.

是啊。

Yeah.

我确实觉得世界越来越有这种感觉了。

I do feel like this is more and more what the world feels like.

呃。

Uh.

是的。

Yes.

呃，是的。

Uh yes.

好吧。是的。这是个好梗。

Okay. Yes. That's a good meme.

有人这样解释 OpenAI 和整个行业的营销挑战：第一，我们接近创造出我们需要向世界解释的东西。第一，我们接近创造出一个能实现任何愿望的精灵。第二，我们要确保我们的第一批愿望能广泛造福人类，让世界变得更多人能拥有更多愿望。第三，你可以许愿的空间非常庞大且富有创意，用真正的人类价值观和偏好来探索它将非常令人兴奋。但还有第四点，那就是：你开始许愿，电脑实现了它们，然后你说，“我没想到这能行。”

The way that somebody explained the marketing challenge for OpenAI and the industry in general: number one, we are close to creating what we need to explain to the world. Number one, we are close to creating a genie that can grant any wish. Number two, we are going to make sure that our first wishes broadly benefit humanity and that we get the world to a place where a lot more people get to have a lot more wishes. Number three, the space of what you can wish for is incredibly big and creative, and it'll be quite exciting to figure that out with real human values and preferences. But there's a fourth thing, which is exactly that: you start making these wishes, the computer grants them, and then you're like, 'I didn't think that was going to work.'

是的。

Yeah.

然后呢？这是一种奇怪的感觉。

What now? It's a weird feeling.

是的。

Yeah.

这就像那个问题，人们花了一百年试图证伪雅各比猜想。然后有人直接问了 Claude，几天之内就解决了。所以我认为未来会有很多好工作。我真的这么认为。我们会有很多智力上的成就感。大多数工作也会比表面上看起来更容易适应。但数学，我认为，是一个非常重要的例子，需要我们现在密切关注，它可能不会朝那个方向发展。事实上，我认为很可能不会。似乎即使系统变得更好，我认识的那些人也并没有变得更清闲。每个人都更忙了。每个人都在做更多的事情，除了那些说“哦，我花了一万亿个 token”的人。我看到有人说，“我花了数不清的 token。”然后有人说，“你为什么没有更成功？”长期以来，技术一直承诺人们会减少工作，享受更多的闲暇。确实朝那个方向发展了。人们确实比历史上许多时期有更多的闲暇时间和更高的生活质量。但不知何故，我们从未实现全社会范围的每周工作四小时的承诺。我不期望 AI 会改变这一点。显然，社会对很多人来说并不奏效，实际发生在人们身上的生产力提升会是件好事。我怀疑这会发生。但我非常钦佩人类的一点是，我们的期望会不断提高。我们总是想要更多。我们想出新的东西来做，为彼此创造，为自己追求。这是一个相对的游戏。人们非常关注自己相对于他人的表现。驱动经济的竞争，以及那些希望为他人有用、创造东西、服务他人的美好愿望——我期待这些会继续下去。我认为在后超级智能的世界里，我们都会比想象中更忙碌，我们仍然会抱怨，但内心深处我们会感到快乐。

It's just like the problem where people spent a hundred years trying to disprove the Jacobian conjecture. And then some guy just asked Claude, and within a few days, it was solved. So I think there's going to be lots of great jobs in the future. I really do. I think we're going to have lots of intellectual fulfillment. And most jobs are going to adapt more than it seems. But math, I think, is a very important example for us to study very closely right now, of something that may not go that way. In fact, I would say probably won't go that way. It seems like even as the systems get better, the people I know are not less busy. Everyone is more busy. Everyone's just doing more stuff, except for the people that are like, 'Oh, I spent a trillion tokens.' I saw someone say, 'I spent ungodly numbers of tokens.' And someone said, 'Why aren't you more successful?' Technology has been promising people they're going to work less and have all this leisure. And it has gone in that direction. People do have more time for leisure than at many previous points in history, and a higher quality of life in many ways. But somehow we never get the promise of the 4-hour work week at mass scale. I don't expect AI to change that. Clearly society is not working for a lot of people, and more productivity gains that actually occur to people would be a great thing. I suspect that will happen. But a thing I really admire about people is our expectations go up. We always want more. We think of new things to do, to create for each other, to want for ourselves. It's a relative game. People are very focused on how they're doing relative to others. The competition that drives the economy and the very wonderful desires about wanting to be useful to others, to create something, to be of service—I expect that to keep going. I think we're all going to be much busier than we thought we were supposed to be in a post-superintelligence world, and we're still going to complain about it, but secretly we're going to be happy.

你认为随着我们的直接投入与价值创造之间的相关性越来越低，地位争夺会如何变化？

How do you think the status games change as less and less of our direct input is correlated with value creation?

我觉得这完全不像我们的直接投入。我们看重的东西，我猜会是那些非常人性化的事物。

I don't think it will feel at all like our direct input. The things that we value, I suspect will be things that are very human.

嗯，比如为别人做饭。你并没有创造很多经济价值，但这能表达爱。

Well, like one thing is cooking for people. You're not creating a bunch of value in the world, but it shows love.

你是在创造价值，那些是体验。你不是在创造传统意义上的经济价值。为别人做饭是个很好的例子。我猜在机器人能做出很棒的食物之后很久，人们一起做饭、吃饭这件事仍然很重要。

You are creating value in that those are the experiences. You're not creating economic value in the traditional sense. Cooking for people is a great example. I suspect that people cooking and eating together will remain important long after robots can do a great job cooking food.

你觉得一个好的启发式方法是不是看看那些流传最久的伟大作品？比如《圣经》，它过去 2000 年都有影响力，很可能一千年后仍然如此。如果要推断人类还会在意什么，那应该是最原始、最古老的东西。我们热爱冒险，热爱一场精彩的探索。

Do you think a good heuristic would be to look at the greatest works that survive the longest? For example, the Bible odds are it's still going to be impactful a thousand years from now because it has been impactful for the past 2,000 years. If you had to extrapolate what things humans will still care about, it's like the most primal, earliest thing. We love adventure, we love a good quest.

做饭。跟进化生物学对赌通常是坏主意。

Cooking. Betting against evolutionary biology is usually a bad bet.

是啊。

Yeah.

所以我认为那些事情会持续下去。

And so I would assume those things continue.

你上一次意识到自己不够有野心是什么时候？

When was the last time that you realized that you weren't being ambitious enough?

嗯，我在算力投资上严重不足。

I mean, I definitely badly undershot on the compute investments.

如果用对的思维模型，你当初能预见到吗？

Could you have known going in with the right mental model?

能，但我被金融市场之类的东西唬住了。我不确定。那显然是个错误。

Yes, but I got psyched out by the financial markets or something. I don't know. That was clearly a mistake.

将来你怎么纠正这个？

How do you correct that in the future?

希望我能从中吸取教训，然后犯一些新错误，但不会再犯那个了。

Hopefully I'll learn from it and I'll make some new mistakes, but I won't make that one again.

在算力方面，这可能是史上最大的基础设施项目，或者即将成为。是啊。你是打算在 OpenAI 内部垂直整合从发电到 token 生成再到服务 ChatGPT 或 Codex 用户的一切？还是打算找一堆合作伙伴？

On the compute side, this is kind of the biggest infrastructure project of maybe all time, or is about to be. Yeah. Are you going to try to vertically integrate under OpenAI's hood everything from power generation to token generation to serving a person in ChatGPT or Codex? Or are you going to try to have a bunch of partners?

我们显然不打算在 OpenAI 内部包揽一切，但会努力把供应链运转得比现在更好。我们已经把芯片设计和模型设计结合起来了，效果不错。我认为不需要用同样的方式把电子制造也整合进来，因为那更像一种大宗商品。不过，当你真正考虑生产智能所需的全链条时，我觉得人们过于关注算法本身，而忽略了能建造更多数据中心的数据中心。在一个有机器人和真正自动化供应链的世界里，完全可以想象这一点：你可以用一个数据中心的算力驱动一群机器人，去建造更多数据中心。如果我们对额外算力最终能解锁什么的判断是正确的，那么相比于其他用途，这可能是非常棒的一件事。

We're not literally going to try to do it all inside of OpenAI, obviously, but we will try to do a better job of a well-functioning supply chain than we have done so far. We have brought chip design and model design together, and that's been good. I don't think we need to bring electron production together in the same way because that's more of a commodity. However, when you really think about the whole supply chain that has to come together to produce intelligence, I would say there is relatively too much focus on algorithms that create better algorithms and not enough focus on data centers that can create more data centers. In a world of robots and a truly automated supply chain, you can totally imagine doing that. You can spend a data center's thinking power to drive a fleet of robots to make more copies of the data center. That is probably a very wonderful thing to do relative to other uses of that compute, if we're right about what that additional compute will eventually unlock.

你有多少时间在思考如何成功设计那个完全机器化的终局——让智能能够自我复制？

How much of your time is spent thinking about how to successfully design that fully robotic end future where intelligence can just produce more intelligence?

几乎全花在执行上。这个想法很明显，很简单。你很快就能说出需要哪些环节。但要整合整个机器和所有公司，就没有宏大设想的荣光和思想碰撞，只有大量艰辛工作。

Almost all goes into execution. The idea is very obvious, very easy. You can relatively quickly say here are the pieces that need to come together. But then to get that whole machinery and all those companies working together, there's none of the glory of the big idea and trying to think the big thoughts, but a lot of grinding.

说到执行，对你来说通常意味着什么？

When you think of execution, what does that typically mean for you?

没有固定的模式。如果说有的话，就是按要求完成问题中的每一步，设法实现它。

There's no typical. To the degree that there is, it's that you just do whatever step is required in the problem and figure out how to make it happen.

但你知道，如何为新的晶圆厂融资，与如何组建优秀的芯片设计团队、让他们与研究团队合作、再让供应链运转起来，这些完全不同。每一个都需要非常不同的方法。我知道 Brian Chesky 有一个做法：他会厚着脸皮去找任何领域的专家，然后直接问他想问的问题，就像在快速学习。如果你需要快速掌握新东西，你的做法是什么？

But you know, figuring out how to finance a new fab build-out is very different from figuring out how to get a great chip design team together and then get them to work with the research team and then actually get the supply chain running. Each of those is a very different approach. One thing I know that Brian Chesky does is he shamelessly tries to find whoever the expert is in any given thing and then basically just goes and asks them whatever question he has, as if he's trying to come up to speed on something. What is your process for doing something like that if it's something new that you need to become good at in a very rapid period of time?

肯定是找专家聊，尽可能多读。我在这方面没什么特别高明的策略。我一直很感激，也总是惊喜地发现，只要开口问，专家们都非常愿意帮忙。这是人性中非常美好的一点。

Definitely finding the experts and talking to them, reading as much as I can. I don't have a lot of great strategies here. I've always been really grateful and pleasantly surprised by how much experts are willing to help people if they ask. It's a very nice thing about humanity.

我听过你说过一句话：你应该总是去争取你想要的东西。

One thing I've heard you say is you should always ask for what you want.

嗯，因为不是每次都能得到，但有时候真的能得到。

Um, because not all the time you get it, but sometimes you do.

有时候确实如此。

Sometimes you do.

而当你说到做到时，奇迹就会发生。

And when you do, amazing things can happen.

你上一次提出一个在别人看来完全疯狂的要求并得到了它，是什么时候？

When was the last time that you asked for something that would seem absolutely insane to someone else and you got it?

其实，最近有一次我上了一架飞机，但我还不能说…

Actually, it was this recent time I got on a plane that I still can't talk about, but it...

好吧，那再往前一个。说一个你能讲的。

Okay, go back one more. Go back one more that you can talk about.

我的意思是，在某种意义上 Codex 就是一个例子。这不是最近的一个，但很有启发性。我们当时远远落后于 Claude Code，试图用一个编程应用击败他们看起来像一场疯狂的敢死队任务。共识是这种事从来不会成功，所以应该放弃。但我们觉得它非常重要。我们要求一个团队去做。他们完成了一件令人难以置信的事情，在商业史上非常罕见。现在它是我认识的大多数最优秀程序员都在用的产品。这看起来不可能，但如果我们没有对团队说‘嘿，我们有一个非常重要但又极其艰巨的任务交给你们’，这一切就不会发生。

I mean, in some sense Codex was an example. This is not the most recent one, but it's instructive. We were way behind Claude Code and it seemed like a crazy kamikaze mission to try to beat them with a coding app. The consensus is that this kind of thing never works and you just move on. But we decided we thought it was really important. We asked a team to do it. They performed a legitimate, unbelievable thing, very rare in the history of business. And now it is the product that most of the best coders I know use. That felt like an impossible thing, but if we hadn't asked the team, 'Hey, we have a really important but extremely hard mission for you,' it just wouldn't have happened.

你为什么做出那个决定？为什么不干脆放弃？

Why did you make that decision? Why didn't you just give up?

这感觉像是少数几个会很快发生的战略领域之一，而且编码对 RSI 至关重要，更不用说经济价值了，所以我们觉得不能放弃。在 YC，你基本上得到了使命最浓缩的版本：只是试图增加世界上的创新量。我的直觉是，如果你真的思考为什么决定去 OpenAI 而不是 YC，你几乎可以论证通过从事 AI 工作可以在增加世界创新量上产生更大的影响。

It felt like one of these few very strategic areas that was going to happen very quickly, and coding is so important to RSI, to say nothing of the economic value, that we just didn't feel like one we could give up on. At YC you basically had the most distilled version of what the mission was: just trying to increase the amount of innovation in the world. My hunch is that if you actually thought about why you decided to go work on OpenAI instead of YC, you could almost make the argument that you could have a bigger impact on increasing the amount of innovation in the world through working on AI.

那你为什么决定去 OpenAI 而不是 YC 工作？

So why did you decide to go work on OpenAI instead of YC?

当时没有那么理性化。我只是隐约知道我一生都想从事 AI 工作，而且我知道这将是我能接触到的最重要的事情。对我来说这是我的热情。我确实认为它很重要，但我就是想做。

It wasn't that intellectualized at the time. It was just like I kind of knew that my whole life I wanted to work on AI, and I kind of knew that it would be the most important thing I could ever touch. For me it was my passion. I did think it'd be really important, but I just wanted to do it.

最初几周是什么样子？我知道有一个——我不知道你是否称之为营地——但人们基本上都聚在一起搞某种 retreat。那时甚至没打算成立一家公司。

What were those first few weeks like? I know that there was this — I don't know if you describe it as a camp — but people just basically all got together on some kind of retreat. It wasn't even like meant to be a company at the time.

我们开始后的头几周吗？我们在 Greg Brockman 的公寓里。大概有 10 或 12 个人。之前做了很多准备工作，然后有一天我们到了那里。第一天大约是 1 月 4 日。当时的感觉是：“好了，我们到了。接下来做什么？我们应该弄块白板，开始讨论想法。也许我们该写论文。”我心想：“哦，我们做了什么？”那真是一个疯狂的时刻。

Like the first few weeks once we started? We were in Greg Brockman's apartment. There were maybe 10 or 12 of us. And it had been all of this work to get it going, and then we showed up one day. The first day I was like, January 4th or something like that. And it was like, 'So here we are. What are we going to do? We should get a whiteboard. We should start talking about ideas. Maybe we should write papers.' And I had this feeling of, 'Oh, what have we done?' It was a really crazy moment.

你当时已经向世界承诺了，而且很多其他人也在做吗？

Had you already committed to the world and a whole bunch of other people were doing it?

我们花了好几年才真正找到节奏并确定要做什么。我的意思是，我们知道想要弄清楚如何构建 AI，但除此之外，真的很不清楚。

It took a couple of years to really get our groove and figure out what we were going to do. I mean, we knew we wanted to figure out how to build AI, but beyond that, it was really unclear.

当你处于这种曼哈顿计划式的摸索阶段时，如何最高效地穿过树林，快速找出哪些方向不该做？

When you're in the stumbling through the woods phase of this type of Manhattan project, how do you most efficiently just stumble through the trees to rapidly figure out what not to work on?

如果某个方向不奏效且你没了思路，你可以砍掉它。这相对容易。真正难的是当某个方向效果极好时——你何时决定砍掉其他好的方向来让最好的方向更好？所以，OpenAI 历史上有很多重要时刻，当一个方向开始奏效时，我们砍掉了其他好的方向，让最好的方向变得更好。当 GPT-3 开始奏效时，我们关闭了机器人等让我们非常兴奋的项目，以便真正聚焦于此。然后当编码智能体最近开始奏效时，我们关闭了同样令我们兴奋的 Sora 和浏览器，以全力推进这个方向。这很难做到。

If something's not working and you run out of ideas, you can kill it. That's kind of easier. The really hard thing is when something is working super well — when do you decide to kill the other things to make it work even better? So, there have been a lot of quite important moments in OpenAI history where one thing started to really work and we killed other good things to make the best thing work better. When GPT-3 started to work, we shut down things like robotics stuff that we were really excited about to really focus on this. And then when coding agents started to work recently, we shut down things like Sora that we were also really excited about and the browser to really work on this. And that is difficult to do.

假设你向 Sora 这样的新业务投入了十亿美元。你如何判断它是否会成为 OpenAI 的支柱？

Let's say you invest a billion dollars into some new business like Sora. How do you decide whether or not that business is going to become a OpenAI pillar?

哦，它本来会非常成功。只是把算力和精力投入到编码智能体上更为重要。

Oh, it would have been super successful. It was just that it was more important to put the compute and the energy into coding agents.

带我走进那样的会议。想象你在会议中说：我们在这件事上花了一年多，巨额资金，大量算力，很多人力和资源，势头很猛，人们用它，喜欢它，然后你决定要拔掉插头。你怎么做出这个决定？

Take me into one of those meetings. Imagine you're in that meeting where you were saying we've spent a year plus on this thing, huge amounts of money, huge amounts of compute, lots of people and resources, a lot of momentum into it, people use it, they love it, and then you decide we're going to pull the plug. How do you make that call?

这不是一次会议就能决定的。这是一个逐渐的意识过程：这些算力、这些人、这个产品方向有更重要的用途，我们将做出非常痛苦的决定去实现它。

It's not a one meeting thing. It's a somewhat gradual realization that there is a more important use of this compute, these people, this product direction, and we're going to make a very painful decision to get there.

当你不得不杀死他们的宝贝，让他们转向另一个宝贝时，如何重新调整团队的方向？

How do you realign people once you have to kill their baby to move them to another baby?

人们大体理解使命和其中的利害关系，以及重新调整方向的必要性。所以即使他们当下不开心，有时他们也非常高兴，有时他们会说：“是的，这对使命来说是正确的。”即使他们当下不开心，他们也会说：“我理解我们为什么要这样做。”这种持续聚焦于超级智能的重定向是件好事。

People kind of understand the mission and the stakes and the need to reorient to get there. So even if they're unhappy in the moment, sometimes they're very happy, sometimes they're like, 'Yeah, this is the right thing for the mission.' But even if they're unhappy in the moment, they're like, 'I get why we're doing this.' And the kind of continuing refocus towards superintelligence is a good thing.

你认为目前有项目会被砍掉吗？

Do you think that there are going to be projects currently that are going to get killed?

我想那会更难推进。

I assume it's harder to pursue.

John Collison 有句很棒的话：基本上世界上的一切，你环顾四周，做成任何事都如此困难，哪怕只是修一张公园长椅都极其困难。所以当你环顾四周时，可以把世界看作一系列热情项目。我认为如果你处于史蒂夫·乔布斯的位置，要创造数百万人每天花数小时与之互动的东西，你必须深入思考如何设计出人们会喜爱、最好不会感到不快的东西。你怎么看待设计？

There's this amazing line from John Collison where he said basically everything in the world, if you just look around, it's so difficult to make things happen even like getting a park bench built is incredibly difficult. And so when you look around, you can kind of think of the world as a universe of passion projects. I think that if you are in Steve Jobs's position, where you're going to create something that a billion people might interact with every single day for hours, you have to think deeply about how to design something that people are going to love and ideally not be unhappy about. How do you think about design?

我感到非常幸运能能与 Jony Ive 合作设计美丽的东西。从他身上我学到了很多：他真的会在尝试找到解决方案之前先研究问题，甚至在真正理解问题之前都不让自己过多思考解决方案。我认为这是我以前没有意识到的一个关键点。真正伟大的设计更多地在于理解问题，而不是灵光一现。如果你急于求成或者过于固守某个想法，你就不会做得那么好。我认为 iPhone 是人类迄今共同制造的最伟大的技术产品。它是一件不可思议的东西，但我已经不再喜欢自己与它的关系了。我关掉了通知，所以现在实际上更喜欢它了。

I feel super lucky to get to work with Jony Ive on designing beautiful things. And I have learned so much from him about how he really studies a problem before trying to get to the solution, and doesn't even let himself think too much about the solution until he really understands the problem. I think this is a key that I didn't appreciate before. Really great design is way more about understanding the problem than the flash of insight. And if you try to rush towards the thing or lock yourself too much into the thing, you will not do as good of a job. I think the iPhone is the greatest piece of technology humanity has collectively yet made. It is an incredible thing, but I don't love my relationship with it anymore. I turned off my notifications, so I actually like it much better now.

我一直开着勿扰模式。

I keep myself on do not disturb.

那还不够。我关掉了所有通知，只留了极少数。我从不开通知，因为想看的时候自然可以看。

That wasn't even enough. I just turned off notifications for everything except a very small number of things. I never have them on because I can just look at it if I want to.

连消息应用也是？

Even messaging apps?

连消息应用也是。那是生活的一大升级。我还删掉了 TikTok，因为它太强大了。

Even messaging apps. That was a big life upgrade. And I deleted TikTok because it was just too powerful.

你上瘾 TikTok 了？

You were addicted to TikTok?

你知道吗，有件疯狂的事发生了。

You know, here's a crazy thing that happened.

我们在构建 Sora 应用时，我让自己对 TikTok 上瘾，因为我想学习。我以前从未真正用过它。人们会给我发 TikTok 之类的，但我从未深陷其中。我想，我真的不想构建那种会消耗人们时间但对他们没有好处的东西。我喜欢 TikTok，真心觉得它很棒。然后我以为我能控制它。我想，它其实挺好玩的，但我只在睡前用 10 分钟放松，完全能控制。然后有一晚用了一小时，接着一个周六下午我在沙发上刷了 3 小时。我意识到这不是 iPhone 应有的样子。我当下享受它像毒品一样，但我知道它对我不好。我短暂地恢复了控制，降到每晚 5-10 分钟。然后我说够了。我觉得 iPhone 很神奇，但我没有足够的自控力保留那个应用，也不觉得消息应用的通知对我有净好处。我确信我们会制造出令人难以置信、美丽、有用且赋予力量的设备。我也确信人们会滥用它们，以我们无法想象的方式让生活变得更糟。我们会适应的。这就是强大技术的本质。

When we were building the Sora app, I made myself get addicted to TikTok because I wanted to learn. I had never really used it before. People would send me a TikTok or whatever, but I never got sucked into it. And I thought, I really don't want to build something that will absorb people's time but not be good for them. And I loved TikTok. I really thought it was great. Then I thought I could control it. I thought, it's actually kind of fun, but I only use it for like 10 minutes to wind down before bed, and I'm totally in control. Then it was an hour one night, and then some Saturday afternoon I was on the couch for 3 hours. I realized this is not what the iPhone was supposed to be about. I was enjoying it in the moment like a drug, but I could tell it's bad for me. I briefly got it back under control, down to 5-10 minutes a night. Then I just said enough. I think the iPhone is amazing, yet I didn't feel like I had enough self-control to keep that app, nor did I feel that notifications on messaging apps were a net good for me. I'm sure we will make incredible, beautiful, helpful, empowering devices. I'm also sure people will misuse them and it'll make lives worse in ways we can't imagine. And we'll adapt. Such is the nature of powerful technology.

当你发明一个新设备时，通过问题空间探索的过程是怎样的？比如如果你是 Johnny——我不代表他回答，但他以前谈过。这是他的过程。他会研究赛车运动的所有历史，汽车舱内文字使用的不同字体，材料及其原因，发动机随时间发出的不同声音以及为什么有些吸引人。他会写真正的书，记录他对极其细节的小部件的探索。

When you're inventing a new device, what is that process of exploration through the problem space? Like if you're Johnny — I don't answer on his behalf, but he's talked about this before. It's his process. He'll go study all the history of motorsport, the different typefaces people use for text in the car cabin, the materials and why, the different sounds engines have made over time and why some appealed. He'll write literal books of his explorations of incredibly detailed small components.

你们基本上是把那些书当作训练数据来制作一个 Johnny 模型吗？

Did you guys basically take those books and try to use them as training data to make a Johnny model?

没有，但那会是一个很好的尝试。他的过程中有一部分我确实不理解。我理解研究，我理解将全新概念打磨成伟大作品。但中间有一个步骤——从真正理解问题到非常新颖的想法之间的灵感——似乎一下子发生，我不明白。

No, but that would be a great thing to try. There is definitely a part of his process that I don't understand. I understand the studying, I understand the refinement of a very new concept into something great. But there is a middle step — the inspiration that comes from understanding the problem really well to the very novel idea — that seems to happen all at once, and I don't understand that.

当你考虑设计某样东西时，你脑子里想的是什么？

When you're thinking about designing something, what goes through your mind?

我完全不是设计师。我努力认清自己不擅长的东西，不会在这方面坚持强烈意见。

I'm not a designer in any way. I try to be pretty good about realizing what I'm not good at, and I wouldn't try to have a strong opinion there.

在商业构建方面，你最不擅长什么？

What are you the worst at on the business building side?

也许是产品。

Maybe product.

真的吗？好吧。如果你不擅长产品，你怎么找到优秀的产品人才？

Really? Okay. How do you find great product people if you're bad at product?

这是我从未完全认同的一点。有一种商业说法是，你只能在你深刻理解的领域招人。我认为这显然不对。我也不懂设计。但我知道 Johnny 很擅长设计。和他聊 30 分钟，我能指出很多证据——非常明显。

This is something I've never quite agreed with people on. There's a business meme that you can only hire people in things you deeply understand. I think it's clearly not true. I don't understand design either. I know Johnny's great at design. There are lots of things I can point to when you talk to him for 30 minutes — it's very obvious.

哪些事情你积极努力提升，哪些事情你只是外包，明白自己永远成不了最佳之一？

Which things do you try to actively get better at, and which things do you just try to outsource, understanding you'll never be one of the best?

我坚信你应该努力提升自己的优势。痴迷于改进你根本不擅长且永远也做不好的事情是一个巨大的陷阱。专注于把你的优势变得非常强。

I am a big believer that you should try to get better at your strengths. The obsession with getting better at things you're just not good at and never will be is a huge trap. Get super good at your strengths.

如果非要排序，你最大的优势是什么？

What are your biggest strengths, if you had to force rank?

我讨厌这个问题。不是假谦虚。我觉得很难对自己的优势说出什么有洞察力的东西。

I hate this question. Not from false modesty. I think it's very hard to say anything insightful about your own strengths.

你显然很善于把人们团结在某个目标周围。

You're clearly good at rallying people around some objective.

当然。但我不认为我能教会不擅长的人如何做到。那些对某人来说自然而然的东西很难解释。至少对我自己来说，我不完全理解为什么我擅长那个，也不知道该怎么办。正如我们之前讨论的，这类东西你可以学但教不了。我认为理解这一点非常重要。如果你想在某方面变好，让别人解释是行不通的。真正研究他们，和他们一起开会，尝试观察并自己学习——那是有效的。当我想提升某件事时，我会尽量和擅长的人在一起，深入研究。但我不认为他们能通过说话向我解释或教给我。

Sure. But I don't think I could teach someone who isn't good at that how to be good at it. The things that come supernaturally to someone are very hard to explain. At least in my own case, I don't deeply understand why I'm good at that or what to do about it. As we talked about earlier, this category of things you can learn but you can't teach. I think this is quite important to understand. If you want to get good at something like that, asking someone to explain it never works. Really studying them, sitting with them in meetings, trying to observe and learn it yourself — that does work. When there is something I do want to get better at, I try to be around someone who's great at it and deeply study it. But I don't think they could have explained it to me or taught it to me by talking.

完全正确。如果你看看最优秀的人学习电子游戏的方式，比如《反恐精英：全球攻势》，你先玩一些了解基本游戏机制，然后看职业选手如何在地图上互动。当你考虑组织节奏——雄心等等——你理想中希望组织不减速甚至更快。你如何融入这一点？

That's totally right. If you look at the way the best people learn video games, like CS:GO, you play some to get basic game dynamics, then watch a pro and see how they interact on the map. When you think about organizational pace — ambition, etc. — you ideally want your organization to not slow down, but even move faster. How do you bake that in?

我觉得 90% 取决于你放在领导岗位上的人。你可以做其他事情——运营节奏、管理技巧——但主要还是看人。对于担任那种角色的每个人，我肯定会考虑：他们是行动快的人还是行动慢的人？

I think it's 90% the people you put in leadership roles. There are other things you can do — operating rhythms, management techniques — but it mostly comes down to the people. Definitely something I think about with everyone in a role like that: are they a fast mover or a slow mover?

如果过去没有和他们共事过，你怎么衡量这一点？

How do you measure that without having worked with them in the past?

嗯，理想情况下你过去和他们共事过。我认为大多数时候公司的高管应该内部晋升，而不是外部招聘。

Well, ideally you have worked with them in the past. I think most of the time executives at a company should be promoted internally, not hired externally.

好的。

Okay.

当您确实需要从外部招聘某人时，您会花大量时间与他们交谈，花大量时间进行背景调查，尝试以某种非正式的方式一起工作。换个话题，过去 12 个月里最痛苦的事情是什么？

When you do need to hire someone externally, you spend a lot of time talking to them, spend a lot of time reference checking, you try to work together in some kind of casual way. Slightly different note. What was the most painful thing that happened in the last 12 months?

老实说，同时带孩子和拼命工作非常残酷。极其痛苦。我觉得自己相对而言是个很顾家的父亲。现在除了工作和陪家人，我什么都不做，但我仍然觉得自己错过了太多这种一去不复返的时光。非常痛苦。

Honestly, having kids and working really hard at the same time is brutal. It's just extremely painful. I think I'm a very present dad relatively. I don't do anything but work and hang out with my family at this point, but I still feel like I'm missing so much of this one-time thing. It's very painful.

当你想到像孙正义这样的人，他大脑的思维方式为何如此不同？我认为他会成为这个十年里最成功的狂人之一。

When you think of someone like Masa, what about his brain and the way he sees the world is so different? I think he will end up being one of the wild successes of this decade.

了不起的人。亲爱的朋友，令人难以置信的人，拥有巨大的信念和信仰，对庞大的数字和规模毫不畏惧，这很棒。世界上像孙正义这样的人不多，我们应该非常感激他们。

Incredible person. Dear friend, incredible person, massive conviction and belief, not afraid at all of big numbers and scale, and it's great. There are not many Masa-like people in the world and we should be very grateful for them.

是什么让他变成这样？我不知道你是否和他谈过，但...

What made him like that? I don't know if you've talked with him, but

我的意思是，据我所知，他一直都是这样。

I mean, as far as I can tell, he was always like this.

你也有这种对规模没有上限的概念吗？

Do you also have the same sort of no ceiling concept of scale?

不完全是。我认为孙正义是个独一无二的人物。

Not really. I think Masa is an N of one character.

你是否有点把 OpenAI 看作金鹅？

Do you kind of view OpenAI as a little bit like the golden goose?

从哪方面说？

In what way?

随着时间的推移，孵化出一堆非常有价值的东西，并建立一种提升世界智能的核心能力。

Spawning a bunch of very valuable things over time and building a core competency of increasing intelligence in the world.

我不太这么认为。我理解你的意思，但并不觉得我们只是在吐金蛋。感觉我们在构建一种复合增长的东西。

Not really how I think of it. I get what you're saying, but it doesn't feel like we're just spitting out golden eggs. It feels like we're building this kind of compounding thing.

我认为我从你那里学到的最喜欢的思维模型之一就是真实趋势与虚假趋势的概念。

I think one of my favorite mental models that I've taken from you is this idea of real trends versus fake trends.

哦对，这个问题很重要。比如每当人们谈论泡沫或其他东西时，我总是有点困惑。我研究了很多沃伦·巴菲特和查理·芒格，他们喜欢谈论泡沫和高估的东西。我试图弄清楚如何判断一件事是真实的还是虚假的，以及它是否会持续下去还是回归均值。

Oh yeah, this is an important thing. Like, whenever people talked about bubbles or anything else, I was always kind of confused. I studied Warren Buffett and Charlie Munger a bunch, and they love to talk about bubbles and things being overvalued. I was trying to decipher how do you figure out whether something is real or fake, and whether it's going to persist and continue or revert to the mean.

那你弄明白了吗？

And what did you figure out?

嗯，这就是你的思维模型。基本上，如果有一小群人……什么是真实趋势？什么是虚假趋势？虚假趋势就是 VR 的情况：炒作很多，但有人买了之后并不真正喜欢。他们并没有开始围绕使用体验来设计自己的生活，然后它就放在架子上落灰。

Well, it was your mental model. Basically, if there's a small group of people... what is a real trend versus a fake trend? A fake trend is, you know, the VR situation: there's a lot of hype, but then someone buys the thing, they don't really love it. They don't start to design their life around the experience of using it, and then it sits on the shelf.

而对我来说，ChatGPT 几乎每天都在用。有时我需要工作很多，所以会花 3 个小时，有时几乎不用，但它在我生活中是持续的。你现在如何判断公司内部或一般情况下的某个趋势是真实的还是虚假的？

Whereas for me with ChatGPT, it's pretty much every day. Sometimes I need to work a lot so I spend like 3 hours, some days it's almost nothing, but it's persistent in my life. How do you figure out now whether something is a real trend or a fake trend inside the company and generally?

同样的原则。是否存在真正深层持久的...

Same principle. Is there a real deep enduring...

你是如何形成这种哲学或思维框架的？

How did you come to that philosophy or mental framework?

嗯，我观察了很多初创公司。在 YC 工作有很多好处，但只是你获得的数据量，只要你愿意花时间去分析和理解，你真的能发现很多东西。

Well, I watched a lot of startups. There are many amazing things about working at YC, but just the amount of data you get, as long as you're willing to spend the time trying to analyze it and make sense of it, you can really figure out a lot of things.

有没有一些你过去拥有的大的思维框架，你认为在当今环境中是错误的，并且随着智能爆炸会继续变得更错误？

Are there any big mental frameworks that you used to have that you think are wrong in today's environment, and will continue to get more wrong over time as intelligence explodes?

我认为其中很多——在当时不一定错，尽管我肯定也有很多错的地方，但很多现在都错了。我认为今天的初创公司看起来仍然和 10 年前差不多，因为传统智慧告诉你应该这样做。当然，有些方面不同。人们会说，嗯，我要雇更少的人，花更多的钱在 Codex 之类的 token 上。但形式可能应该完全不同。我见过一些人尝试以完全不同的方式做初创公司。但大多数人只是用它来多用 Codex，这似乎还不够。

I think a huge amount of them — they weren't necessarily wrong at the time, although I'm sure I was wrong about a bunch of things too, but a huge amount of them are wrong now. I think a startup of today still looks mostly like a startup of 10 years ago, because that's what the received wisdom says you're supposed to do. Sure, it's different in some ways. People are like, well, I'm going to hire fewer people and spend more money on tokens for Codex or whatever. But it should probably look very different. And I have met a few people who are trying to do a startup in a completely different way. But most people just use it to mean use more Codex, and that doesn't seem like enough.