Sam Altman 和 Greg Brockman 谈 OpenAI 十年历程

我们要在 90 分钟内讲完所有那些个人 drama。哦，别担心，别担心。这部分都会播出来吗？请忍受我有点傻的开场，但会没事的。欢迎来到金融堡垒，资本之都。这不是那个播客。这是核心记忆。我是 Ashley Vance。

We're going to do all that personal drama in 90 minutes. Oh yeah, oh don't you worry. Don't you worry. Is this part all going to be in the podcast? Bear with me through my slightly silly intro, but it's going to be okay. Welcome to the fortress of finance, the capital of capital. This is not that podcast. This is core memory. I am Ashley Vance.

我是 Kylie Robison。我觉得这期节目会非常精彩。通常这时候我们会介绍嘉宾，但今天可能没必要了。我们有 Sam Altman 和 Greg Brockman，OpenAI 的联合创始人。你们可能听说过。谢谢你们来做客。

And I'm Kylie Robison. And I think we have a hell of an episode for you. Normally we do an introduction of people at this point, but that's perhaps unnecessary today. We have Sam Altman and Greg Brockman, the co-founders of OpenAI. You may have heard of it. Thank you guys for being here.

谢谢邀请。

Thank you for having us.

非常感谢。

Thank you so much.

我想这是你们第一次一起做播客。太棒了，但我觉得确实如此。至少是很久以来第一次。也许是第一次。我不会因为你们在我们的节目上做而争辩，但你们确实买了一个播客。是有什么……我们只是运气好。你们运气好。

I think this is the first time you guys have ever done a podcast together. That is amazing, but I think that's true. Certainly in a long time. Maybe the first one. And I will not fight you for doing it on our show, but you did buy a podcast. Is there... We just lucked out. You lucked out.

是啊，我接受。我接受。我很好奇你们为什么买了一个播客。不用深入，但有什么快速想法吗？

Yeah. I'll take it. I'll take it. I was curious why you guys bought a podcast. You don't have to go deep on it, but did you have quick thoughts on that?

我认为做 TBPN 的人非常了不起。他们是非常有创意的思考者，在这个我们正在构建对人们极其有用的 AI 系统的世界里，帮助人们理解为什么这对他们的个人生活和工作有价值。我觉得他们就是能传达这一信息的人。

I think that the people who do TBPN are incredible. I think they're just very creative thinkers, and I think that in this world that we're moving to of building these AI systems that are so useful for people and helping people understand why that's valuable for them in their personal lives and work lives. Like these are the kinds of people that I think could help tell that message.

我看过你上那个节目。你上过 TBPN 吗？上过。好吧。我不看所有集，但这是个有趣的播客。我想，既然你们很久没一起做节目了，我们可以在开头稍微怀旧一下。Kylie 和我这些年逐渐了解了你们。我们准备这期节目时也在反思。我们刚过了十周年纪念日。你们是剩下的联合创始人中的两位，我想 Wojciech 是第三位。所以你们是贯穿公司的一条恒定线。你们从弱者开始，最终成为强者。经历了这么多 drama 和起伏。我很好奇你们的关系在这过程中如何变化，你们如何相互配合，是否随着时间演变。

I've seen you on it. Have you been on TBPN? I have. Yeah, okay. Okay. I don't watch all the episodes, but it's a fun pod. Well I thought, you know since this is you guys haven't done this together at least in a while, we would... I was going to go down nostalgia lane just for a little bit at the beginning. Kylie and I have gotten to know both of you over the years. You know, I was just... we were reflecting as we were preparing for this. We're a little bit past the 10-year anniversary. You guys are two of the remaining co-founders. I think Wojciech is the third. So you know, you're this constant line that's been running through the company. You started as an underdog. You ended up as the top dog. It's all through a lot of drama, undulations. I was just genuinely curious of how your relationship through all this has changed and how you guys have played off each other and if it's morphed over time.

这非常美好。我们总是希望 drama 少一些，希望能专注于技术。但在一个充满混乱、drama、紧张、斗争和权力争夺的世界里，能和一个了解全部背景的人保持关系，真是难以置信的美好。我们拥有所有这些历史，在美好的时刻和非常艰难的时刻真正相互依赖。这是 OpenAI 最美好的事情之一。

It is extremely nice in... Look, we always wish there were less drama. We wish we just got to focus on the tech. But in a world of so much chaos and drama and tension and fighting and power struggles, it has been unbelievably nice to have a relationship with someone that's got the full context. We have all this history and to really depend on each other in amazing times, very tough times. It has been one of the nicest things about OpenAI.

你知道，在很多方面，OpenAI 的第一个时刻就是在 2015 年 7 月的那次晚餐之后，我和 Sam 一起开车回城，我们看着对方，说我们必须这么做。对吧？当时一直在讨论，现在成立一个追求 AGI 并产生积极影响的实验室是不是太晚了。

You know, in many ways the very first moment of OpenAI was right after this dinner that we did in July of 2015 and Sam and I were driving back to the city together and we looked at each other and we were like, we have to do this. Right? There'd been all this conversation of is it too late to start a lab that could go after AGI and have a positive impact.

现在回想起来，我们当时那么担心真是可笑。是啊，太晚了。你们错过了。你们错过了。但你们开始的时候就是这种感觉。我当时想，不，DeepMind 会在这上面遥遥领先。

So ridiculous now that we were so worried about that. Yeah, it was too late. You missed it. You missed it. Feeling like that when you guys started though. I was like, no. DeepMind's going to run away with this, you know, yeah.

是啊。晚餐的结论是，这并非明显不可能。我觉得我们俩都觉得，好吧，这太重要了，我们必须做。是的。我认为这种精神一直延续。早期我们的很多运作方式，比如我记得我当时失业，所以第二天就全职投入了。Sam 其实有份日常工作，但我们经常通电话，大概一天五次。没错。

Yeah. And you know, the conclusion of dinner was it wasn't obviously impossible. And I think both of us just felt like, okay, this is just so important. We just have to do it. Yeah. And I think that spirit continues. And a lot of how we operate in the early days, like I remember I was unemployed at the time, so I was full-time on it the next day. Sam actually had a day job, but we were constantly on the phone, like probably like five times a day. Yeah, exactly.

你们当时已经是亲密朋友了吗？

Were you guys already close friends at that point or not?

我们认识很久了。或者说我觉得很久……实际上我不确定。那是哪一年？2010 年？2011 年？大概是你开始在 Stripe 工作的时候。没错。所以我们是通过 Collison 兄弟认识的。

We had known each other for a super long time. Or I felt like a super... I don't actually... What year was that? 2010? 2011? Whenever you started at Stripe. That's right. Yes. So we met through the Collisons.

是的，所以我们算是普通社交朋友。

Yeah. And so we'd been kind of casual social friends.

没有我想的那么久。是啊，也许是 2010 年，现在是 2015 年，所以五年。时间压缩了。显然，在这种高压环境下做这项工作，我想你们只会随着时间的推移变得更亲密。

It wasn't as long as I thought. Yeah, maybe it was 2010 and this was now 2015. So five years. Time compresses. And obviously I mean the being in the pressure cooker of all this doing this work, I mean I would imagine it only you guys have only got closer over time.

是啊，人们用“创伤纽带”这个词，我讨厌这个词。我喜欢的是和你一起在战壕里的人。但努力工作的好处之一，尤其是在压力大的时候，就是你真的能建立起这些关系，至少我没见过其他方式能形成这样的关系。

Yeah, you know, people use the word trauma bonding. I hate that. I like the other things about like the people you're in the foxholes with, but the one of the nice things about hard work, no matter what, but certainly hard work in stressful times is you really forge these relationships that I at least have not seen get formed any other way.

是的。我确实认为 Sam 和我工作与相处的方式可能不同于典型的联合创始人关系。我觉得我们一直保持联系，那种一天五通电话，每通两到五分钟的精神依然存在。我们一直同步。我们并非在所有事情上都意见一致，对吧？我们看待世界的方式并不完全相同，但这也是我们如此强大的原因。我认为我们有非常互补的方法，Sam 会说‘有个想法’，我会想‘也许我们可以用另一种方式’，或者‘从这个角度处理怎么样？’或者‘这和我们正在想的另一件事有什么关系？’我非常欣赏 Sam 的一点是，他总能看出不同想法之间的联系，或者专注于我们需要达到的大局，然后我们一起弄清楚如何实际做到。我认为将宏大抱负与执行联系起来，正是 OpenAI 一直以来的特色。

Yeah. And I do think the way that Sam and I work and relate is maybe different from what you'd expect from a typical co-founder relationship. Like I think that we are just in constant contact, that five calls a day, two minutes, five minutes each, that kind of spirit remains. Like I think we're just in constant sync. And we don't always agree on everything, right? It's not like we come at the world from exactly the same point of view, but that's why we're so strong together, right? Is I think we have very complementary approaches, that Sam will say, 'Here's an idea.' I'll think about, 'Well, maybe we could do this other way.' Or what about if we approached it with this angle? Or how does this relate to this other thing that we're thinking about? And one thing I deeply appreciate about Sam is that I think he always sees these connections between different ideas, or just like keeps focused on here's the big picture that we need to get to and then together we figure out well how do we actually do it and I think connecting the grand ambition with the execution like that is what has always distinguished OpenAI.

在这十年里，有哪些时刻你们觉得分歧很重要？你们记得任何关键的时刻吗？

What are some of the points in these 10 years where you felt like it was really important that you guys diverged? Do you remember any key moments for you guys?

我认为 Greg 做得最好的一件事，也是我不太本能去做的，就是推动专注于最重要的事情，无论是他自己的工作还是公司要做的事情。所以有时候我想做更多事情，而 Greg 会说，这是最重要的事情吗？我们真的就做这个吧。让公司聚焦。我们在这一点上有分歧，但 Greg 的这种精神在整个公司都非常有帮助。

I think one of the things that Greg has done the best which is not my instinct is really just pushed to focus on the most important thing in his own work and also in what the company is going to do. So there have been times where I have wanted to do more things and Greg has just said, you know, is this the most important thing? Let's really just do this. Let's get the company focused. And we've diverged on that and that's been like a very helpful spirit of Greg's throughout the company.

是的，我还想补充一点，比如在算力方面，我们不断提高目标。有时候我觉得，好吧，我逻辑上知道我们正在迈向算力驱动的经济，需求总是会超过供给，但我们有这么多艰苦的工作要做，我们已经有了这些大型计算机，正在将它们投入运营，而且还有大量的物理基础设施要建设，我已经感到不堪重负了，但 Sam 却说，不，我们还需要更多。我认为这实际上非常重要，因为有时候很容易忽视更高层次的东西——这不仅仅对未来 6 个月至关重要，而是对未来 2 年、5 年、10 年都至关重要。你需要平衡，有时要深入细节，但不能被细节淹没。我认为这种平衡正是 OpenAI 之所以成为今天的样子，以及我们将走向何方的关键所在。你们之间有没有某个产品或策略是你们分歧最大的？

Yeah, and I would also add I think even for example thinking about compute and just constantly raising the ambition. And sometimes I feel like okay, I kind of logically know that yes, like we're moving to this compute powered economy and yes, that demand is always going to outstrip supply, but like we've got all this like hard work to do and we already have all these big computers and we're operationalizing them and you know, you still have all of this like just physical infrastructure to build and you feel already swamped in it and swamped in it and Sam is like no, we need even more. And I think that that actually has been a very important thing to really not like sometimes it's easy to lose sight of the higher order bit of just the fact of this is going to be so important for not just the next 6 months, but this is what's important for the next 2 years, the next 5 years, 10 years and I think that the keep like you need this balance of sometimes swimming in the details, but you can't be swamped in the details and I think that that balance is something that I think again is what really has contributed to what Open AI is and where we are going to go. Is there There must be one product or strategy you guys have What's the one that you've disagreed about the most vehemently?

我刚才在想 Greg 说的事。这不是一个产品，但我想到了这个。我们过去经常讨论如何谈论安全。我们从未在安全的极端重要性以及做对或做错意味着什么上有分歧，但这个领域在如何谈论安全、如何利用安全，以及这在多大程度上关乎权力而非真正保持安全方面，有着奇怪的关系。在我们早期，我更多地被卷入必须用特定框架来谈论这件事的思维中，而 Greg 非常坚持我们不会落入传统框架，我们不能那样谈论。即使在那时，我认为因为这件事太重要了，我们可能还是落入了用错误框架谈论过多的陷阱，但我认为 OpenAI 迄今为止最大的贡献之一是找到了一种不同的谈论安全的方式，不仅是在我们如何构建产品、如何谈论社会需要做什么上，还包括我们如何部署它们——迭代部署的整个理念，实际上走向一个我们学会如何在风险上升时部署越来越安全的产品的世界。Greg 在那条线上坚持住了，我认为这对公司非常重要，顶住了不这样做的巨大压力。我认为这对我们的整个战略都非常重要，不仅是我们如何谈论事情，还有我们如何发布和构建产品。

I was just thinking when Greg was talking about this. This is not a product, but it was this thing that came to mind is that I was going to say it before you asked that. We used to talk a lot about how to talk about safety. We never disagreed on the extreme importance of safety and what it will mean to get this right or get this wrong, but the field has had a strange relationship with how we've talked about safety, how we've used safety, and how much that becomes about power versus actually keeping things safe. And I think earlier in our history, I got swept up more in the we got to really talk about this in a particular frame, and Greg was very disciplined about we're not going to fall into the traditional frame. We can't talk about that way. Now, even then, I think we have because this is so important, we would ever have fallen into the trap of still talking more in the wrong frame than we should, but I think one of OpenAI's greatest contributions to date has been finding a different way to talk about safety, not just in how we build the products, and how we talk about society needs to do, but like what how we deploy them, the idea whole idea of iterative deployment, and actually getting to maybe not the actually getting to a world where we're figuring out how to deploy products that get increasingly safe as the stakes go up, and Greg really held a line there that I think has been quite important to the company against extreme pressure not to do that. And I think it's been like quite quite important to our whole strategy, not just how we talk about things, but how we ship and build products.

是的，如果你看看 OpenAI 基金会，它是管理 OpenAI 的非营利组织，持有很大一部分股权，其支柱之一是 AI 韧性。这实际上意味着思考如何让 AI 成为对世界有益的东西。答案不是单一的干预措施，对吧？不是你有思维链监控就完成了使命。它实际上是一整套社会围绕这项技术应如何定位的深度序列。我认为这种观点——你不能在一篇论文中解决 AGI 造福世界的问题——必须是一个全球性的努力，需要社会的贡献，来自许多不同的人，许多不同的方式来真正理解这项技术是什么，它将如何影响人们，如何影响世界。我认为这在 10 年前我们刚开始时完全没有被重视或理解，因为很容易只关注我们是技术专家，我们在构建技术，那是我们唯一需要解决的问题。我不是说有人明确那样说过，但我认为有时你会陷入那种思维陷阱。所以我们花了很多时间做第一性原理思考，真正思考如何以实际帮助人们日常生活的方式，将变革性技术交付给世界。你意识到的一件事是，如果你有一项非常强大的技术会改变一切，那么如果你之前已经用一项不那么强大的技术以积极的方式帮助改变了事情，情况可能会更好。所以如果你那样想，你就会开始被拉向思考韧性、思考迭代部署的道路。我认为这正是 OpenAI 内部我们两人之间的动态——我们总是在思考如何真正实现这个使命并让它变得更好。

Yeah, and if you look at for example the OpenAI Foundation which is the nonprofit that governs OpenAI and has a very large chunk of of equity one of its pillars is AI resilience. And what that really means is thinking about how do we make AI be something positive for the world? And the answer is not any one intervention, right? It's not you have chain of thought monitoring and now you've achieved the mission. It's really a whole deep sequence of different ways that society should orient around this technology. And I think that this perspective of you're not going to solve AGI going well for the world in a paper. It has to be a worldwide effort from contributions from society, from many different people, from many different ways of really understanding what this technology is, how it will affect people, how it will affect the world. And this this is something that was not, I think, at all appreciated or understood when we were starting out 10 years ago because it's very easy just to fix it on, you know, we're technologists. We're, you know, building technology. That's the only problem we need to solve. I'm not saying anyone explicitly said it that way, but I think sometimes you can fall into a mental trap of thinking about it that way. And so a lot of what I think we have spent time on is being first principles thinkers and really thinking about how do you operationally deliver transformative technology to the world in a way that is going to actually help people in their daily lives. And one thing you realize is, well if you have a very powerful piece of technology that will change things, probably it's going to go better if you've had a less powerful piece of technology you've already helped change things in a positive way. And so if you just think about it that way, you start to really be pulled down this road of thinking about resilience, thinking about iterative deployment. And again, I think this is a lot of the dynamic within OpenAI with between the two of us that we're always thinking about these questions of how do we actually achieve this mission and make it go better.

是的，感觉就像昨天我在 2022/2023 年的西南偏南大会上看到你，那时关于 AI 的讨论和今天听到的完全不同，感觉这 10 年来你们讨论安全和对齐的方式发生了变化。我想知道你现在如何反思这一点？你会改变那些小组讨论和新闻采访中的哪些内容？你从谈论安全中学到了什么？甚至在谈到安全之前，我认为我们作为技术极客陷入了这样的框架：我们要构建超级智能，等等，这对你们会很好。但我们没有充分填充那些“等等”。我们谈论我们正在构建这项惊人的技术，它会做所有这些美妙的事情，现在世界上有一种感觉：好吧，看来你们是对的，你们确实要构建这个东西。但为什么？我们为什么想要这个？它会为我们做什么？这个领域的很多人说，哦，它会治愈癌症，你会非常幸福，但这显然没有引起共鸣。很多人说，当然，治愈癌症会很棒。但我认为人们真正想要的是繁荣和自主权，他们希望继续有有意义的工作可做。前几天我看到一篇很棒的帖子，让我印象深刻，它提到了“逆境权”。人们实际上需要生活中的一些挑战。

Yeah, I feels like just yesterday I saw you at South by 2022 2023 and it was a completely different discussion about AI than what we hear today and it feels like that's something that's changed in 10 years of like how you discuss safety and alignment and I'm wondering how you reflect on that now like what would you have changed in those panels and those news hits and like what have you learned about talking about safety? Even before we get to safety, I think we have fallen in the frame as tech nerds of talking about we're going to build superintelligence and dot dot dot it's going to be great for you. And we've not filled in enough of the dot dot dot. Like we talk about we're building this amazing technology, it's going to do all these wonderful things and there is like a sense in the world now of okay looks like you were right, you are going to build this thing. Why like why do we want that? What's that going to do for us? And the thing that a lot of the field has said of oh it's going to you know cure cancer and you'll be so happy or like that's clearly not quite resonating. A lot of people are like sure cure cancer that'd be wonderful. I think what people really want is prosperity agency that they're going to continue to have meaningful work to do. There I saw an incredible post the other day that really stuck with me which was like a a right to adversity. People people actually you want some challenges in life.

你并不希望每一天都完美无缺、所有事都为你代劳。人们对 AI 有一种恐惧：假设你是对的，假设你造出来了，假设它赚了所有钱、干了所有活，那我做什么？我的孩子做什么？生活会变成什么样？成长从何而来？人们要为什么而奋斗？我认为，作为整个领域，作为 OpenAI，我和 Greg 谈了很多关于这项惊人技术及其能力，以及它的技术奇迹，但我们还没有足够地把未来的图景串联起来。当我与有学龄孩子的家长交谈时，最常见的问题是：我的孩子应该学什么？未来会是什么样？什么还有经济价值？我意识到这并非他们真正想问的问题。真正的问题是：我的孩子如何在这个新世界里过上充实的生活？你可以从经济角度回答，但那并不能真正让人安心。我认为这里有更深层的东西。

You don't want every day to be perfect and everything done for you. There's a fear with AI that, let's say you're right, let's say you build it, let's say it makes all this money and does all the work. What do I do? What's my kid going to do? What's life going to be like? Where's the growth going to come from? What are people going to strive for? I think we've talked as a field, as OpenAI, and Greg and I have talked a lot about the amazing technology and what it can do, the technological marvel, and we have not connected the dots enough on what the future is going to be like. When I talk to parents with school-age kids, the most common question is: what should my kids study? What's the future going to be like? What will still have economic value? And I realize that's not quite the question they're really trying to ask. It's: how is my kid going to have a fulfilling life in this new world? You can answer it economically if you want, but that doesn't actually settle people. I think there's a deeper thing here.

我正要说到，我认为我们实际上对此有很多见解。例如，在 ChatGPT 中，有很多人说他们的生命或亲人的生命是通过从 ChatGPT 获得的信息而得以挽救的。有一个人，他的孩子患有致盲性头痛，被拒绝做核磁共振。他们用 ChatGPT 研究症状，并以此为由争取到了核磁共振的保险。结果发现是脑瘤，他们得以干预并救了他的命。在应对这一经历时，那个家庭说：‘没有 ChatGPT，我们完全不知道该怎么办。’这只是一个故事，还有许许多多这样的故事。我认为人们正在真正理解这项技术不仅能抽象地帮助社会，还能切实帮助他们。它可以帮助他们赚钱。现在我们开始看到一波创业浪潮，我认为这将是今年的一大主题。我们将构建这些 AI，让计算机来适应你，而不是你扭曲自己去适应计算机。想想我们工作的方式，那并不自然，不是我们天生该做的。相反，计算机会为你工作。那么什么是工作？什么是好的？什么是你真正想做的事？那将是极其人性化的事情：这种能动性，这种赋权。所以我认为这项技术正在带来非常乐观——不是盲目乐观，而是非常积极的变化。但人们更容易注意到什么会消失——那些你认为稳固不变的东西即将改变。而更难看到的是即将到来的新事物，你得到的新东西，以及这场变革带来的好处。我认为我们越来越意识到，除了阐述其他方面，我们也必须把这一点讲清楚。

I was just going to say that I think we actually have a lot of perspective on this. For example, within ChatGPT, we have so many people who say that my life or the life of a loved one was saved through information I got from ChatGPT. There's someone who had a kid with blinding headaches. They were denied an MRI. They used ChatGPT to research the symptoms and used it to argue to get insurance for the MRI. It turned out there was a brain tumor. They were able to intervene and save his life. Navigating that experience, that family says, 'We have no idea how we would do this without ChatGPT.' That's just one story. There are so many of those. I think people are really understanding that this technology can help not just abstractly society, but can help them. It can help them make money. Right now we're starting to see this wave of entrepreneurship. I think that's going to be a huge theme throughout this year. We're going to build these AIs that are able to make the computer conform to you, rather than you contorting yourself to the computer. Think about the way we do work. It's not natural. It's not what we were designed to do. Instead, the computer is going to do work for you. The question of what is the work, what is good, what is the thing you actually want to be done—that's going to be something very deeply human: this agency, this empowerment. So I think there is this very optimistic, not blindly optimistic, but very positive change that this technology is bringing right now. But it's so much easier to notice what's going to go away—what you thought was solid and stable that's going to change. It's much harder to see what's coming, what's the new thing you get, what is the benefit that comes as part of this transformation. I think that's something we're increasingly realizing we also have to spell out, in addition to spelling out the other sides.

好了。我们在 Core Memory 做什么？我们报道创新、快速、前瞻的公司，这就是为什么 Core Memory 由 Brex 赞助，因为 Brex 是许多这类公司的智能金融平台。从初创公司到全球最大的企业，有 3 万家公司依赖 Brex 的技术进行财务管理。他们有智能公司卡、高收益商业银行账户和出色的费用自动化工具。我讨厌做报销，而 Brex 的 AI 和软件能直接处理这些费用，弄清楚我们的钱花在哪里，并为你处理很多事情，这样你就不必自己浪费时间了。访问 brex.com/corememory 了解更多，然后跟上节奏吧。让我们行动起来，摆脱那些过时的财务软件，迈向未来。Core Memory 和 Brex。

All right. What do we do at Core Memory? We cover innovative, fast-moving, forward-thinking companies, which is why Core Memory is sponsored by Brex, because Brex is the intelligent finance platform for many of these companies. 30,000 companies from startups to the world's largest corporations rely on Brex's technology for their finances. They've got smart corporate cards, high-yield business banking, and expense automation tools that are fantastic. I hate doing my expenses, and Brex's AIs and software run right through those expenses, figure out where we're spending money, and take care of so much stuff for you, so you don't have to waste your time on it yourself. Go to brex.com/corememory to learn more, and just get with the program. Let's get going. Let's get out of this archaic finance software and move toward the future. Core Memory and Brex.

基于你们两位所说的，我有几个问题。我刚刚和我儿子去了欧柏林学院，在一个大约有六七百人的房间里。这让我很着迷。那是在俄亥俄州，人们从全国各地赶来。校长在演讲，然后开放问答。很多问题都是关于 AI 的，但这让我更加意识到我们生活在一个多么大的泡沫里，因为那些问题在我看来——不是要贬低谁——都是相当基础的。我对那个房间的整体感觉是：哇，你们似乎不太了解正在发生以及即将席卷你们的事情。这令人警醒，也让我大开眼界。他们问的是如何在课堂上使用 AI，如何阻止孩子过度依赖 AI，但后来也稍微深入了一些。我觉得这个房间里相当聪明的人并没有真正了解正在发生的事情。所以如果问题是人们不知道这些可能做到的例子，而且我们甚至不知道这具体会是什么形态，我看不出我们如何解决这个问题并以某种方式让人们做好准备。

I have a couple questions based on what both of you were saying. I just went to Oberlin College with my son, and we were in a room with maybe 600-700 people. It was fascinating to me. It's in Ohio, and people had come from all over the country. The president was giving a talk and then opened it up to Q&A. So many of the questions were about AI, but it reinforced for me how much of a bubble we live in here, because they were what I would consider—not trying to put anyone down—pretty basic questions. My overwhelming sense of the room was: wow, you guys kind of don't know what's unfolding and about to come across all of you. It was alarming, eye-opening for me. They were asking about how you're going to use AI in class, how you're going to stop kids from just relying on it, but then they advanced a little beyond that. I felt like this room of pretty smart people were not really in touch with what's going on. So if this is a problem that people don't know these examples of what could be done, and we don't even know what shape this could take exactly, I don't see how we solve this and prepare people in some way.

嗯，我对此有一个积极的看法：我认为如果你只是阅读关于 AI 的内容，它会给你一种印象——感觉你是在努力理解这项新技术，它是什么。但当你使用它时，它非常直观。这在很多方面正是 AI 的目的。回想一下过去 70 年我们设计计算机的方式：实际上是机器不太理解你。你脑子里有目标，你必须把它们分解成机器的语言，无论是编写汇编代码还是更高级的语言。现在你可以和你的电脑对话了，但即使有了 ChatGPT，你仍然需要理解语言模型的概念。为什么你必须创建新的对话、新的标签页？为什么它不能只是一个你与之交谈并记住一切的东西？这些都是技术限制，但我们正在改进它们。所以我认为我们正在构建的是最直观的技术。我们正在构建能够使机器适应你的东西。

Well, I do have one positive take on this: I think that if you just read about AI, it gives you one impression—it feels like you're trying to wrap your mind around this new technology, what it is. But when you use it, it's so intuitive. That is the purpose of AI in many ways. Think back to how we designed computers for the past 70 years: it's really about the machine not quite getting you. You have these goals in your head, you have to break them down into the machine's language, whether that's writing assembly code or higher-level languages. Now you can kind of talk to your computer, but even with ChatGPT, you still have to understand concepts of language models. Why do you have to create new conversations, new tabs? Why can't it just be a thing you talk to that remembers everything? These are technological limitations, but we're improving them. So I think what we're building is the most intuitive technology. We are building something that can bend the machine to you.

当人们使用它时，他们会意识到：‘哇，这就是我能用它做的事情。’比如，我喜欢听的一些故事是，中西部有个人——我有个朋友，他妹妹在跟他讲一个她想要的应用程序，她希望有人能创建出来。她详细描述了它。他一边听一边‘嗯哼、嗯哼’，把她说的话一字不差地输入到 Codex 里。按下回车。几小时后，他把结果展示给她看：一个和她描述的一模一样的应用。她说：‘这是什么鬼？谁建的？这正是我想要的。’他说：‘你建的。’对吧？这就是我认为每个人都会经历的顿悟时刻。一旦你体验到 AI 能为你做什么，以及你被赋予了力量——就像你脑海中有一个你想在世界上看到的画面。看看我们即将推出的一些新图像模型，绝对不可思议。你可以以前所未有的方式进行创作。另一个我想到的故事：我祖母患有痴呆症、阿尔茨海默病，这对每个人来说都非常艰难。但有一件事让我印象深刻：她居然能用她的 Alexa 播放音乐，而且她还能记住歌词并跟着唱。这是她通过技术与自我保持联系的一种方式。这让我印象深刻。就像如果你能构建直观的界面，任何人都能以某种方式连接——现在感觉我们的技术是针对特定个体的，你必须培养所有那些与你真正想要的东西无关的技能，而只是某种特定于技术本身的东西。

And when people use it, they realize, 'Wow, this is what I can do with it.' Like, some of the stories that I love hearing are someone in the Midwest who—one of my friends, his sister was telling him about an app that she wanted, that she wished someone had created. She described it in a detailed way. While he was listening, he was like, 'Uh-huh. Uh-huh,' typing into Codex exactly what she was saying. Pushed enter. A few hours later, he shows her the result: an app that's exactly what she described. And she's like, 'What the hell is this? Who built this? This is exactly what I wanted.' And he said, 'You built it.' Right? And that is the kind of aha moment that I think everyone is going to go through. And once you experience what AI can do for you and the fact that you are now empowered—like you have an image in your head that you want to see in the world. If you look at some of the upcoming new image models we have, absolutely incredible. You can create in a way that was never possible before. Another story I think about: my grandmother had dementia, Alzheimer's, and it was extremely tough for everyone. But one thing that really stuck out to me was she was actually able to use her Alexa to play music, and she could still remember the lyrics and sing along to songs. It was a way that she was connected to who she was through technology. That really stuck out to me. It's like if you can build interfaces that are intuitive, that anyone can connect to in ways that—right now it feels like we have pieces of technology that are targeted to specific individuals, that you have to build all this skill that's not really related to deeply what you want, but it's just somehow specific to the tech itself.

Greg 说的让我想到了三件事，我全都同意。第一：在我们推出 ChatGPT 之前，我们曾试图告诉人们：‘AI 要来了。你们得注意。它会改变一切。这真的很重要。’但几乎没人关注。我们写了漂亮的博客文章。我们完成了各种令人难以置信的壮举。我们赢了电子游戏比赛。我们有一个能用一只手解魔方的机器人。我们做了所有这些了不起的事情。我们觉得自己很酷，还说：‘哦，《纽约时报》报道了这件事。我们一定做得很好。’但实际上根本没人关心。它没有产生任何实际影响。然后我们推出了 ChatGPT，这远不是我们做过的最令人印象深刻的技术——远非如此。但一旦人们能感受到它，他们就说：‘好吧，我明白了。’我认为那可能是世界第一次集体意识到也许 AI 是真的。因为人们可以使用它并从中获得价值。他们形成了自己的感受。这和听说它完全不同。正如 Greg 所说，这种情况在编程模型上再次发生了。但这两个确实是世界更新认知并说‘好吧，有件事正在发生’的重大时刻。未来还会有更多。但到目前为止，你可以问计算机任何问题并得到答案，或者让计算机用代码为你做任何事——这些都非常强大。我认为世界就是这样更新的。我们说超级智能即将到来，会改变一切。也许听这个播客的人会说：‘好吧，听起来合理。可能应该注意一下。’但它不会对世界产生那么大的影响。所以我认为，为了帮助那些受众思考这不仅意味着人们不能在课堂上作弊，而且对世界意味着什么，我们能做的最重要的事情就是推出优秀、令人愉悦的产品，这些产品能创造大量价值且易于使用。我们会继续这样做。

Three things came up for me during what Greg was saying, which I agree with all of. One: before we launched ChatGPT, we used to try to talk to people and say, 'This AI thing is coming. You've got to pay attention. It's going to change everything. It's really important.' And it got kind of no attention. We wrote these beautiful blog posts. We did all these incredibly impressive feats. We won video game competitions. We had a robot that could solve a Rubik's Cube with a hand. We did all this amazing stuff. And we kind of felt pretty cool and we said, 'Oh, the New York Times wrote about this. We must be doing great.' But truly no one cared. It got no actual impact. And then we launched ChatGPT, which was by far not the most impressive technological thing we had done—by far, by far. And as soon as people could feel it, they were like, 'Okay, I understand it.' I think that was the moment probably most that the world has said maybe this AI thing is real—collectively all at once. Because people can use it and they can get value out of it. They develop their own sense. It's very different than hearing about it. That's happened again, as Greg said, with coding models. But those have really been the two large significant moments where I think the world has updated and said, 'Okay, there's this thing happening.' There will be more in the future. But so far, the fact that you can ask a computer anything and get an answer, or have a computer do anything with code for you—those have been really powerful. And I think that's how the world is going to update. Us saying superintelligence is coming is going to change everything. Maybe the people who listen to this podcast will say, 'All right, sounds reasonable. Probably should pay attention to that.' But it won't have that big impact on the world. So I think the most important thing we can do to help that audience think not just about what this means for people cheating in class, but what this is going to mean for the world, is to ship great delightful products that create a lot of value and are easy for people to use. And we will continue to do that.

第二件事是，我们在历史上一次又一次地看到，当我们推出一个相当不错的东西时，人们会说：‘我真的无法想象如果它变得更好会怎样。它不可能好太多了。’Greg 提到的这个图像模型，我们很快会推出，对我来说就是一个真实的例子。我大致认为图像生成已经解决了。我想：‘是的，它真的很好。我不需要它变得更好。还有很多……’然后这个新东西真正提醒了我：‘哇，这还能走得更远，我能做的还有很多。’它做了什么？它生成了极其出色的图像。它解决了文本问题吗？因为我尝试制作核心记忆商品的模型——很快再试一次。团队在这方面做得非常好。但即使是 ChatGPT，当我们把 GPT-4 放进 ChatGPT 时，我记得很多人，包括我一些见多识广的朋友，都说：‘就是它了。这就是 AGI。如果模型变得更聪明，我不在乎。只要让它更便宜。这太棒了。’如果你回去用那个版本——大概是 2023 年 3 月的 GPT-4——你会觉得：‘这太糟糕了。’但当时人们说：‘解决了。解决了。它通过了图灵测试。完成了。它不可能更好了。’然后它变得越来越好，你可以不断提高期望和可能性，更不用说之后还要做推理模型和编程。我只是在说 ChatGPT 在那段时间里变得有多好。所以我认为我们一次又一次地看到这种情况：世界说：‘好吧，你做出了这个了不起的东西。它已经是最好的了。它已经满足我的需求了。’然后，月复一月，或者至少季度复季度，期望和能力大幅提升。

The second thing is we have seen again and again in our history that when we put something out that's pretty good, people say, 'I can't really imagine what happens if this gets better. It can't get much better.' This image model that Greg was mentioning, which we'll launch soon, was a real example of that for me. I kind of roughly thought image generation was solved. I was like, 'Yeah, it's really good. I don't need it to get any better. There's a lot of...' And then this new thing has been a real reminder of 'Wow, this can go so much further and there's so much more I can do.' What does it do? It makes ridiculously great images. And has it solved text? Because I tried to make core memory merch mock-ups—try again very soon. The team really did a great job on this one. But even with ChatGPT, when we put GPT-4 in ChatGPT, I remember a lot of people, like sophisticated friends of mine, saying, 'This is it. This is AGI. If the model got smarter, I don't care. Just make it cheaper. This is amazing.' And if you go back and use that whatever it was—I guess like March 2023 version of GPT-4—you'll be like, 'This was terrible.' But at the time people were like, 'It's solved. It's solved. It's beat the Turing test. It's done. It can't get any better.' And then it gets better and better, and you can just keep raising the expectations and possibility of what you can do, to say nothing of then you're going to do reasoning models and do code. I'm just talking about how much better ChatGPT got in that time period. So I think we see this again and again, too, where the world is like, 'Okay, you made this amazing thing. That's as good as it's going to get. That's as good as I need.' And then, month by month, or at least quarter by quarter, expectations and ability ratchet up in a huge way.

然后 Greg 说的第三件事是，这些模型相对于它们未来的样子仍然相当笨拙，但更重要的是，它们对你生活的了解非常有限。你仍然需要哄着它们、诱导它们，试图得到你想要的东西。我们离一个能了解你所有上下文的模型并不遥远。它了解你。它了解你的生活。它知道你在做什么。它知道你在乎什么。它了解你生活中的人。它可以访问你的电脑和浏览器，当然，如果你愿意，以你想要的方式，而且它可能随着时间的推移，越来越多地了解你周围真实世界中发生的事情。这将彻底改变使用电脑和使用 AI 的感觉。

And then the third thing that Greg was saying there is these models are still quite dumb relative to what they will be, but more than that, they have quite limited awareness of your life. You are still having to massage them and cajole them and try to get the thing that you want. We are not that far away from a model that just knows all of your context. It knows about you. It knows about your life. It knows what you're doing. It knows what you care about. It knows about the people in your life. It has access to your computer and your browser and, if you want, of course, in the ways you want, and it has access, maybe increasingly over time, to what's happening in the real world around you. That is going to be a complete change to what it feels like to use a computer and what it feels like to use AI.

我对此感到无比兴奋，但我觉得连我们自己都还没有很好的直觉，知道那到底会是什么感觉。

And I am tremendously excited about that, but I don't think even we have a good intuition yet for what that's really going to feel like.

是的。说到这一点，你想想现在你花多少时间向 ChatGPT 或你用的任何工具解释情况。想想那有多令人沮丧，就像你有个同事，你总得跟他们解释‘不，我大概想要这个。情况是这样的。’我有些项目回头再看时也是这种感觉。对吧？这根本不是你想要这些系统表现的方式。

Yeah. And to that point, you think about how much time you spend right now just explaining to ChatGPT or whatever tool you're using what's going on. And if you think about how frustrating that is, like you got a coworker you're constantly trying to explain to them, 'No, this is kind of what I want. This is what's going on.' I have projects I come back to and it's like that. Exactly, right? It's not really how you want these systems to behave.

再补充一点萨姆说的，我认为 OpenAI 面临的最大技术挑战之一是机会太多。AI 是一个充满无限机会的领域。无论你在哪个维度上扩展，都会有新的、前所未有的、令人惊叹的东西出现。所以重要的是要有愿景，知道我们真正该聚焦在哪里，从哪里获得最大的回报和收益，让多个不同的努力汇聚成整体。

And just to say one more thing on what Sam was saying, I think that one of the biggest technical challenges we have at OpenAI is too much opportunity. AI is this field of boundless opportunity. No matter what dimension you expand on, there's going to be something new and unprecedented and amazing. So the important thing is having a vision for where we can really focus, where we're going to get the most returns and the most benefit from this, having multiple different efforts that all add up to something.

我认为，现在这些编码系统，显然会扩展到所有计算机工作。有趣的是，你想想你做的所有工作——面对面交谈、在电脑上打字、试图把上下文放进去——这些是多么无缝地交织在一起。这将是一个非常重要的问题。所有这些都很重要，但在个人生活中你也很想要这个，我们开始把目标称为个人 AGI，对吧？这个 AI 真正了解你，拥有上下文，你可以信任它。你问它财务或健康相关的问题，它能给你可靠的信息。所有这些都很重要，它也需要那个上下文。所以你开始看到，在技术层面上，用于深度计算机工作的 AI 和用于个人生活的 AI 之间有一条模糊的界限，即使你想要不同的系统——一个只了解你的工作上下文，一个了解你的个人上下文。你可以并行推进它们，并建立在相同的技术基础上。我觉得我们正在构建的技术最神奇的地方在于，核心都是一个神经网络，对吧？仍然是深度学习。你在扩大规模。你在构建一个系统，并将其应用于所有这些不同的惊人应用。

I think that when it comes to these coding systems now, it's going to clearly expand to just all computer work. It's funny, by the way, that you think about the degree to which all the work you do is seamlessly interlinked between in-person conversation, typing on your computer, trying to figure out how to get the context in there. It's going to be such an important problem. All these things matter, but then also in your personal life you really want this, and we're starting to call where we want to go the personal AGI, right? This AI that really knows you, has the context, you can trust it. You ask it questions related to finances or health and it can give you trustworthy information. All these things are important and it needs that context as well. So you start to see this blurry line between an AI used for deep computer work and an AI used in your personal life, at a technological level, even if you want different systems because you want one that knows your work context and one that knows your personal context. You can pursue those in parallel and build on the same technological foundations. The thing I find most amazing about the technology we're building is it's all at the core one neural net, right? It's still deep learning. You're scaling it up. You're building one system and applying it to all these different amazing applications.

我能顺着这个问个问题吗？我向自己保证过不会这样对你，但直到最近我对 AI 都还有点怀疑。

Can I ask a quick question along those lines? I promised myself I wouldn't do this to you, but I was kind of skeptical until relatively recently on AI.

是什么让你改变了看法？

What updated you?

有几件事。一是我开始更多地使用智能体，真正让它们按我的意愿行事，感受到了你们俩说的那种感觉。我就想，‘哦，这真的在帮我省时间。它在执行我的指令，而且做得相当不错。’另一件事是我大量报道生物技术，看到一些成果出来，让我觉得编码和生物技术是目前最明显的应用领域。但后来，那天晚餐时和你聊天，读了你写的一些东西，我看到了一条基于 LLM 的清晰路径，通向你所描述的。但我内心仍然有一部分，可能是因为我经常和语言打交道，写作方面还是不太好。这纯粹是轶事和直觉。我感觉到的反馈仍然是，‘不，这不是超级智能。’在个性方面我们还没到那一步。

Well, a couple things. One was I started playing with agents a lot more and really shaping them to do what I wanted, kind of feeling what both of you were talking about. I was like, 'Oh, this is actually saving me a lot of time. It's doing my bidding. It's doing it quite well.' The other is I cover biotech so much, and just seeing some of the results that are coming out strikes me that coding and biotech seem like the most obvious places right now where it's really going. But then, I was talking to you the other day at dinner and reading some things you've written about, I see this clear path based off LLMs to going to what you describe. I still have this part of me though, I think because I deal in language so much, where the writing is still not great. It's nothing other than anecdotal and intuitive. It strikes me that the feeling I get back still is, 'No, this is not a superintelligence.' We are not there yet on personality.

但 LLM 会带我们到那里吗？没错，LLM 会带我们到那里。我只是觉得它非常参差不齐。例如，就在过去几天，我们的 AI 解决了一个长期存在的英国问题。一个数学谜题，长期以来备受关注。一位数学家花了大量时间研究这个问题，多年思考，发表了他对贡献的看法。陶哲轩也说，这看起来像是 AI 发现了数学不同领域之间的联系。我们开始看到这些机器中涌现出真正的美。当然，那是一个特定领域，对吧？数学和创意写作非常不同。这些 AI 能够产生有趣的见解并提供帮助，很多数学家现在都说，‘想想我们还能做更多什么。’所以我认为，在这些 AI 擅长的方面，我们有一个参差不齐的前沿。我们知道如何不断推进这个前沿。但我们真正在寻找的，如果你想想 AlphaGo 的第 37 步，那不仅是深刻的洞察，而且真正改变了人们对围棋的理解，现在下围棋的人更多了。所以它确实增加了人们所做事情的意义。我认为我们将要创造的东西，会让你抱怨更少，而且我认为你在写作方面能做到的事情，会比今天想象的要多得多。

But does LLM get us there? Right. LLMs will get us there. I was just thinking of it as very jagged. For example, just in the past couple days our AI solved this long-standing British problem. This mathematical mystery that's been of great interest for a long time. A mathematician spent a lot of time working on this, many years thinking about this problem, posted about his views on what the contribution was. Terence Tao also saying that this looks like there's a connection between different fields of mathematics that this AI has discovered. We're starting to see real beauty come out of these machines. Now, that's a particular domain, right? Mathematics is very different from creative writing. The fact of these AIs being able to have interesting insights and being able to help, there's so much that these mathematicians are now saying, 'Well, think about what more we can do.' So I think that we have a jagged frontier in terms of what these AIs are good at. We know how to keep pushing that frontier back. But the thing we're really looking for, if you think about AlphaGo move 37, that was something that not just had deep insight, but it really changed people's understanding of the game of Go, and more people play Go now. So it really increased the meaning of what people were doing. I think that what we're going to do is make something that you're going to have fewer complaints about, but I think you're also going to be way more able to do the thing that you want in writing than you can imagine today.

抱歉我老提这个，但我被告知 GPT-5 会在写作上非常出色。我变得不那么愤世嫉俗的转折点是推理模型，当 Claude 提供了非常好的研究，然后你们也有了研究功能，对我来说就像‘天哪，这太棒了。’帮我省了很多时间。但我之前被承诺现在会有好的写作。

I'm sorry to harp on this, but I was told GPT-5 was going to be really good at writing. And my update when I became less cynical was reasoning models, and when Claude delivered really good research, and then you guys had a research feature, that was for me like, 'Oh my god, this is amazing.' Saved me a lot of time. But I was promised before that I would have good writing by now.

写作方面还是，没有灵魂，你知道吧？少了点什么。

The writing still, there's no soul, you know? There's something missing.

好吧，我告诉你。从技术角度来看，我们拥有的技术是你可以用这种无监督的方式训练模型。所以我们查看所有公开可用的数据，它学会预测下一个是什么，这实际上是把模型放在新情境中，试图找出合理的行为。然后我们进行强化学习步骤，它实际尝试想法，并根据表现获得奖励或惩罚。

Well, I'll tell you. From a technical perspective, the technology we have is you can train a model in this unsupervised way. So we look at all the publicly available data, and it learns to predict what comes next, and it's really about putting it in a new situation, trying to figure out what's a reasonable thing to do. Then we do a reinforcement learning step, where it actually tries out ideas, and it gets rewards and punishments based on how well it did.

你知道，就是正信号、负信号。不是打败它。对，完全不是。就是这个信号。然后难点在于，你怎么判断？怎么决定某件事是点赞还是踩？所以在数学和科学领域，这比一些更开放的领域容易得多。但我们也有越来越聪明的 AI，它们能更好地提供这种奖励信号。所以我认为挑战的一部分是如何扩展可评分的任务集，这一直是重点。实际上有很多有趣的东西可以追溯到 OpenAI 成立之初。我们当时对如何发展有个蓝图。但我想说，我们正在接近目标。我们肯定还有很多进步要做。但希望你能继续给我们反馈，我们会为你改进。

You know, just positive signal, negative signal. Not beating it. Exactly. Not at all. Yeah, just this signal. And then the hard part is, well, how do you judge? How do you decide if something was thumbs up or thumbs down? So in math and science, it's much easier than in some of these more open-ended fields. But we also have AIs that are getting much smarter and much more able to provide that kind of reward signal. So I think part of the challenge has been how to expand the set of tasks that can be graded, and that's been a lot of the focus. There's actually very interesting stuff dating back to the very beginning of OpenAI. We kind of had a picture of how this would go. But I would just say I think we're getting there. I think we definitely have a lot more progress to make. But yeah, hopefully you can keep giving us feedback and we'll be able to improve it for you.

其中一个内在挑战是，你想要的写作风格和大多数人想要的很不一样。现在我们必须做一个大约十亿人使用的模型，大家都差不多。而我们想做的是让模型在个性化方面做得非常好，以至于你觉得它是个很棒的作家。而另一个生活和你截然不同、需求也不同的人，也觉得它是个很棒的作家，能满足他们的需求。这是非常不同的事情。

One embedded challenge in that is that the writing you want is very different than the writing that most people want. And right now we have to make a model that about a billion people use, all kind of like. And what we'd like to do is to get the model so good at personalization that you think it's a great writer. And some other person that has a very different kind of life and set of needs than you also thinks it's a great writer for what they want. Those are very different things.

我觉得这正是我想问的。比如对于个人 AGI，你希望它有那种闪光、灵魂、那种魔力，对吧？我只是，如果不是……我注意到，扬·勒昆一直反对这个观点很久了。我很快看到了丹尼斯的东西，他似乎和你们观点一致，但你知道，LLM 会带我们到那里，但他似乎在说我们可能还需要一些特别的东西……但你们似乎非常自信。

I think that's what I'm kind of asking though. Like for the personal AGI you would want it to have that sparkle, that soul, that sort of magic, right? And I just, if it's not... And I saw, you know, Jan LeCun's been arguing against this for a long time. I caught this Dennis thing very quick where he seemed to be kind of aligned with you guys but you know, LLMs will get us there but he seemed to be saying we still need a couple more ticks maybe of something special to... but you guys seem so confident.

我们并没有试图让它成为阿什利·万斯认为的伟大的作家。我们试图让它解决世界上最聪明的数学家都解决不了的开放数学问题。我不想对你的智力与这些数学家相比说什么，但我觉得……

We haven't tried to make it be, you know, what Ashley Vance will think is a great writer. We have tried to make it solve open math problems that the smartest mathematicians in the world can't solve. And I don't want to say anything about your intelligence relative to these mathematicians, but I think...

作家更聪明。

Writers are smarter.

但我想说，解数学也很难。所以如果我们能做到这一点，我很自信这种方法也能学会你认为的好写作，并为你做到。但顺便说一句，我们在这个维度上也有一些新模型即将推出。所以这个播客播出后……

But I was going to say, you know, solving math is also hard. So if we can do that, I'm pretty confident this approach can also learn what you think is great writing and figure out how to do it for you. But by the way, we also have some new models coming on this dimension as well. So after this podcast comes out...

等着评判吧。告诉我们。告诉我们它看起来是否更好。

Wait to judge it. Let us know. Let us know if it looks better.

写作，我认为个性，而且我们一直在改进。所有能力都在沿着锯齿状前沿上升。所以在这个领域，判断当前状态不如判断斜率重要，还要拟合指数曲线。所以想想写作和一年前相比如何？如果你对 2.5 版本失望，抱歉。别担心，我们非常有动力去真正交付。但我认为我们确实有清晰的路径来改进它，以满足人们想要的每一个应用。

Writing, I think personality, and I think we're always improving. All the capabilities are moving up the jagged frontier. So I think in this field it's always really important to judge not the current position but what the slope has been, and also fit it to an exponential. So if you think about how does the writing compare to a year ago? Sorry if you're disappointed with 2.5. Don't worry, we're very motivated to really deliver. But I think we really have a line of sight for how to improve it for every application that people want.

从一开始，核心记忆播客就得到了 E1 Ventures 的好人们的支持。他们是硅谷一家年轻而雄心勃勃的风险投资公司，投资于年轻而雄心勃勃的公司和人。非常感谢 E1 Ventures 的所有支持。在你们描绘的世界里，这项技术，好吧，假设一切顺利，这是一个相当乐观的情景：疾病被治愈，资源变得更丰富，问题被解决。所以整个人类都在被提升。我仍然看不到……我的意思是，似乎一些世界上最聪明的人正在开发这项技术。我觉得几乎在所有情景下，它都会不成比例地让他们受益，即使每个人可能都被提升一点。事情会变得更加极端，因为每次，你们谈论如何操纵这些工具、使用它们……我只是觉得事情会变得更加极端，富人会……

Since day one, the core memory podcast has been supported by the fine people at E1 Ventures. They are a young and ambitious VC firm in Silicon Valley investing in young and ambitious companies and people. Thank you so much to E1 Ventures for all your support. In the world that you guys have outlined, this technology, okay, say everything is working great and this is a quite optimistic scenario: diseases are being cured, resources are becoming more abundant, problems are being solved. So humanity as a whole is being lifted up. I still don't see... I mean, it seems like some of the world's smartest people are developing this technology. It strikes me it will disproportionately benefit them in almost all scenarios, even though everyone might get lifted up a bit. Things really get more extreme because every time, what you guys are talking about how you would manipulate these tools, use these... I just feel like things will go even more extreme and the haves will be...

永久的下层阶级，以及看到这些非常强大的工具时那种‘哦天哪，那我到底在这里干什么？’的感觉，感到非常被剥夺权利。

The permanent underclass and that sentiment of seeing these really powerful tools and feeling like, 'Oh my gosh, what am I even here for then?' and feeling very disenfranchised.

你有了所有时间去做有趣有创意的事情。我只是觉得其他人会以如此极端奇妙的方式操纵这些……

You get all this time to do fun creative things. I just feel like other people will be manipulating this in such extreme fantastic ways that...

这是 OpenAI 使命的核心。这真的是我们创办这个地方的原因。因为你看到这项强大的技术即将到来。它将成为有史以来最重要的技术。你如何确保它惠及每个人？真正惠及全人类。这就在我们的使命中：确保 AGI 惠及全人类。我们是认真的。如果你看看我们的公司结构，我们尝试了多次不同的迭代，试图在结构中编纂我们的一些价值观：如何确保这真的能惠及人们？你可以在我们做的产品选择中看到，对吧？我们决定推出 ChatGPT，因为我们真的相信这项技术需要交到人们手中。顺便说一句，这非常有争议。有另一种学派认为，正确的方式是秘密构建，不能让人们访问，需要以另一种方式来做。我认为当我们思考如何真正让社会变得有韧性，如何真正惠及人们时，一切都指向我们一直在走的方向。这里面有很多细微差别。但我认为我们前进的方向是，每个人的底线不会只提高一点点。我认为我们将走向这样一个世界：想想看，口袋里有一个比当今世界上任何人都能获得的最好的医疗团队还要好的医生，任何有智能手机的人都能获得。这即将到来，而且是免费的。对吧？这是一个疯狂的事实。这不是小事。它会好一点。这从根本上以巨大的方式提高了底线。现在我认为提高上限的问题也很公平，对吧？我认为问题在于分布到底如何。我们经常思考这个问题。看看 OpenAI 基金会。它拥有 OpenAI 大约 25%到 30%的股权，对吧？那价值超过 1500 亿美元。

This is at the core of the OpenAI mission. This is really why we started this place. Because you see this powerful technology coming. It's going to be the most important technology ever created. How do you ensure it benefits everyone? Like truly all of humanity. That is in our mission: ensure AGI benefits all of humanity. And we really mean it. And if you look at our corporate structure, we've tried multiple different iterations at trying to codify some of our values in the structure: how do we ensure that this is something that really does benefit people? You see it in the product choices we make, right? We decided to launch ChatGPT because we really believe this is technology that we need to be able to put into people's hands. And that was very controversial, by the way. There was a different school of thought saying that the way to do it is you have to build it in secret, that you can't get people access, that you need to do it this other way. And I think that when we look at how do we actually make society resilient, how do we actually benefit people, it all kind of points to the direction that we have been going. And there's a lot of nuance to it. But I think that the way we're headed is very much that I think the floor for everyone is not going to go up just a bit. Like I think we are going to head to this world where, I mean, you think about even having a doctor in your pocket that's better than the best medical team that anyone in the world could get today, accessible to anyone who has a smartphone. Like that is coming and that'll be free. Right? That is a wild fact. That's not a small thing. It's going to be a little bit better. That is fundamentally raising the floor in this massive way. Now I think the question of also raising the ceiling, I think that's fair too, right? I think that the question of exactly how the distribution goes. And we think about this a lot. You look at the OpenAI Foundation. That has, you know, somewhere between 25%, 30%, somewhere around there of OpenAI equity, right? That's like more than 150 billion dollars.

我能看到世界的三种未来。一种未来是底线大幅提升，每个人主观上都变得富裕 10 倍，物质丰裕和繁荣程度惊人。但那些学会使用智能体并拥有大量算力的人会成为万亿富翁，不平等加剧。另一种未来是底线提升没那么大，总繁荣度较低，但不平等程度下降。我认为真正的关键就在于这两种之间。

I can see three futures of the world. One where the floor comes way up, everybody gets subjectively 10 times richer, material abundance and prosperity are huge. But because people who learn to use agents and get a lot of compute become trillionaires, inequality gets worse. Another world where the floor doesn't come up as much, total prosperity is less, but inequality comes down. I think the real crux is between these two.

第三种更可怕吗？

Is the third much scarier?

不，但它会分散我想表达的重点。我认为很多人觉得我们应该偏好第一种是显而易见的。格雷格和我也觉得显而易见。但在情感上，很多人并不这么认为。

No, but it's a distraction from the point I want to make. I think a lot of people think it's obvious that we should prefer the first. Greg and I think it's obvious. But emotionally, it's not where a lot of people are.

我不想声称什么是显而易见的。就我个人而言，我看到了这项技术的巨大潜力。从社会层面看，甚至美国竞争力——看看机器人技术，我们并不领先。在软件方面，我们有领先优势和这个机会。所有这些因素都值得同时考虑。

I don't want to claim what is obvious or not. For me personally, I see so much potential in this technology. From a societal level, even American competitiveness—look at robotics, we are not ahead. On the software side, we have a lead and this opportunity. All these factors are worth considering at once.

在许多其他事情上意见不同的人，都同意美国需要在芯片、机器人、AI 等各方面保持竞争力。但关于如何组织社会和经济，有一个巨大的问题：我们是追求最大繁荣并接受不平等，还是因为担心擅长使用这项技术的人会掌握所有权力而加以限制？理智上，这对我来说很清楚。情感上，我理解为什么它不清楚。这个工具的复利性质是我们尚未理解的。无论如何，每个人都应该想要更多的算力和基础设施，以及尽可能便宜的 AI 接入，否则如果只有富人拥有它，不平等会加剧。

People who disagree on a lot of other things do agree that America needs to be competitive on chips, robots, AI, and everything else. But there's a huge question about how we organize society and the economy. Do we push for maximum prosperity and accept the inequality, or do we constrain it because of the fear that those good at using this will have all the power? Intellectually, it seems clear to me. Emotionally, I get why it's not clear. The compounding nature of this tool is something we don't yet understand. Regardless, everyone should want much more compute and infrastructure, and the cheapest possible access to AI, because otherwise inequality is exacerbated if only the rich have it.

我想在此基础上补充。列出这两个选项是很好的观点，但这是仅有的两个吗？山姆最后说的很关键：如果你有接入——如果你有算力，AI 对每个人都是机会。如果每个人都有算力，而且现在成长起来的一代在使用智能体方面会比我好 10 倍，那么这就是美国梦最极端的版本。任何有欲望和意志力的人都能表现出色并获得流动性。所以正确的答案不是两者之一——还有其他可能。

I want to build on that. Laying out these two options is a good point, but are those the only two? The last thing Sam said is an unlock: AI is opportunity for everyone if you have access—if you have compute. If everyone has access to compute, and the generation growing up now will be 10 times better at using agents than I am, then it's the most extreme version of the American dream. Anyone with desire and willpower can outperform and have mobility. So the right answer is neither exactly—something else is possible.

你知道，我对硬件的报道非常深入，我敢说我去过的美国工厂和硬件初创公司可能比任何人都多，同时我也关注中国。我脑子里一直想的是，是的，我们在软件上领先，在 AI 上领先。软件是美国过去四五十年来的优势故事。但是，当我想到这项技术如何在物理世界中体现时，我看不到美国有任何竞争力。不仅仅是机器人，还有这些硬件系统中的所有组件。我去埃尔塞贡多，看到 30 家初创公司，这很好，有各种想法和尝试，但跟我在中国看到的相比，简直是小巫见大巫。这让我觉得，所有这些最终都会在物理世界中体现，而美国似乎极其不利，难以获胜。

You know, I just cover hardware so deeply and I think I'm pretty, I would venture to say I've probably been to more factories and hardware startups in the US than just about anyone and keep an eye on China. I mean, the thing that always plays through my head is that yeah, we're ahead on software. We're ahead on AI. Software's been the story of the US for the last 40 or 50 years of our strength. But, you know, when I think about how this technology manifests itself in the physical world, I do not see a future where the US is competitive in any way, shape, or form. Not just in robotics, but you know, just all of the componentry that goes into these hardware systems. I go to El Segundo. I see 30 startups. It's great. There's like this flourishing of ideas and people trying things, but it's like such small potatoes compared to what I see in China. It just strikes me that there obviously will be a physical manifestation of all this in the world and the US just seems like extraordinarily disadvantaged to win that to me.

嗯，我想到那些正在努力改变这一点的人。我说山姆在机器人方面，嗯，我知道通过投资不同的硬件公司……不，不，我们正在想办法，机器人显然是其中的一部分，我们正在想办法在机器人领域非常成功。我认为，如果你能选一件事让美国在制造业和原子世界具有竞争力，那就是我们需要大量能制造更多机器人的机器人。但我们现在连一个执行器都造不出来。你知道，我们会搞定的。OpenAI 会搞定。通过你们的机器人？必须的。我们必须。但还有其他重要部分。想想世界需要多少吉瓦的电力来支持 AI 工作负载，我们需要制造多少芯片，还有那些无聊的事情，比如如何组装足够的机架、铺设足够的网线。美国在这方面极其落后。我认为机器人将是解决方案，看看这一切需要多快实现，美国目前擅长什么，我们需要建设多少基础设施，以及手工做需要多长时间。但你的诊断和关键性是对的，幸运的是，棋盘上有了新棋子。通过 OpenAI 还是美国？

Well, I think of people who are working incredibly hard to try to change that. I'd say Sam is like on robotics and well, I know through all the investments in different hardware companies or No, no, we're trying to figure out how Robots are an obvious part of this and we're trying to figure out how to be very successful at robotics. I think if you could pick one thing to make the US competitive at manufacturing and the world of atoms in general, you would say we need a lot of robots that can build a lot more robots. But like we can't even make an actuator. You know, like going to figure that out. We will Open AI. Okay. Like through your robotics? It would have to. We have to. The But there's other parts of this that really matter. If you think about like the number of gigawatts of power that the world is going to need just to support AI workloads, the number of chips we're going to have to fabricate, the boring stuff about like how we're going to just get enough racks assembled and enough network cables strung and made in the first place. The US is extremely behind here. I think robotics will be the solution if you just look at how fast this has to happen and what the US is currently good at and how much infrastructure we have to build up and how long that would take to do by hand. But you are right in the diagnosis and the criticality of it and I think luckily we have like a new piece on the chessboard. Through Open AI or the United States?

好吧，告诉我，你什么意思？我有不同的感觉。我觉得我们在假装做硬件，我们会彻底……

Okay, tell me I mean what do you mean? I have a different feeling. I feel like we're play acting hardware and we're going to get completely...

嗯，按照目前的轨迹，是的。不，我们完全同意。如果我们能制造真正的通用机器人，如果我们能有一个类似 Codex 的机器人能力，你可以说像这样配置工厂，给我造更多这种机器人，或者去搞一个矿并提炼这种东西，那么游戏规则就变了。

Well, on the current trajectory, yes. No, I we totally agree with that. If we can make true general purpose robots, if we can have a Codex equivalent power thing for robots that you can say like go configure a factory this way and make me more of this kind of robot or go figure out a mine and refine this kind of thing, then the playing field changes.

好吧，但你说的是美国，不是特指 OpenAI。但我觉得你在指某件事，你知道吗？嗯，我不认为美国除了这种 AI 加机器人之外有可靠的计划来快速追赶。我不知道那个棋子是什么。你最近看到了什么？

Okay, but you said the US and not Open AI specifically. But yeah, I feel like you're pointing at something that Did you know though? Well, I don't think the US has a credible plan other than this kind of AI plus robotics to catch up fast enough. I didn't know what the chess piece was. Did you see something recently?

哦，我只是说通用人工智能。

Oh, I just I mean I meant general purpose artificial intelligence.

我觉得机器人领域存在一个真正的鸡和蛋的问题：如果没有机器人硬件，就很难开发软件；如果没有好的大脑，就很难有动力开发硬件。我们确实看到了这一点。我们在 2018 年有一个机器人项目。你还记得那个机械手吗？是的，是的，超酷。关于那只手，我们通过强化学习训练它。实际上，和用来解决竞技游戏的算法完全一样，对吧？太神奇了，单一技术做两件事。唯一的区别是，这只手运行 20 小时后，它的肌腱——那些线——会断裂。然后你会停机，机械工程师进来修理，也许叫醒他们，让他们来办公室。你意识到你不能那样做机器学习。就是不能。所以，我们最终取消了那个项目。那个团队后来去开发了 GitHub Copilot，为那个项目做贡献。所以，你意识到软件世界发展得多快。但我认为我们现在已经到了一个点，我们有这些惊人的通用算法，可以以各种方式使用。我们看到，如果你开始将它们应用到物理世界，那么硬件开发和硬件调优的协同设计将会有完全不同的故事。所以，我认为有真正的变革潜力，但我们作为一个国家需要有意志力去真正做这件事。但我想我们同意你的观点，没有这样的东西，目前的轨迹看起来很糟糕。他确信情况真的很糟。相当糟。真的很糟。

I feel like there's a real chicken and egg that's existed in robotics where if you don't have the robotic hardware, it's kind of hard to develop the software. And if you don't have the great brain, it's kind of hard to be motivated to develop the hardware. And we really saw this. We had a robotics project back in 2018. Do you remember the robotic hand? Yeah, yeah. Super cool. The thing about that hand to know, so we trained it through reinforcement learning. Actually, the exact same algorithm we used to solve competitive video games, right? Wild. Single piece of technology doing both. The one difference is that this hand would run for 20 hours before it had these strings as its tendons, and they would snap. And then you'd have downtime, mechanical engineer would come in and fix it, you know, maybe be wake them up, have them come into the office. And you realize that you cannot do ML that way. You just cannot. So, actually, we ended up canceling that project. That team went on to go work on what became GitHub Copilot, you know, to contribute to that effort. And so, you just realize how much faster software world has been moving. But I think we're at a point now where we have these amazing general-purpose algorithms that can be used in all sorts of different ways. And we see how if you start to be able to apply those to physical world, then I think you are going to have a very different story for how the hardware development and how the hardware tuning to the co-design is going to happen. So, I think that there's like real potential for change, but we as a nation need to have the willpower to really do this. But I think we really agree with you that without something like that, the current trajectory looks terrible. We're so He's convinced that it's really bad. It's pretty bad. It's really bad.

退一步说，机器人、模型，这些都是你们感兴趣并真正关心的东西。但最近有报道说你们整合并聚焦了，我很好奇，我相信观众也想知道，现在桌上有什么？什么被砍掉了？你们关心什么？为什么做这些削减？为什么重要？所以，在我让格雷格回答之前，在他回答之前，格雷格接手了，真正弄清楚我们统一的产品组合和支持它的研究是什么。公司内部非常开心。所有东西出货需要更长时间。他上任才几周，大概这样。但格雷格在这里做的事情带来的能量、兴奋和热情是难以置信的。所以，你可以说说会是什么。

Taking a huge step back, you know, robotics, models, these are things that you guys are interested in building and really care about, but you guys have recently, it's been reported that you consolidated and focused, and I'm really curious, I'm sure like the audience wants to know, like, what's on the table now? What got cut? What do you guys care about? And why did you make those cuts? Why was it important? So, before I let Greg answer, before he does, Greg has taken over really figuring out what our cohesive product offering and the research to support that is going to be. And it's been amazingly joyful inside of the company. Like it's going to take a little bit longer for all the stuff to ship. He's only been in the role for a few weeks, maybe, something like that. But the energy and excitement and the sort of like enthusiasm about what Greg is doing here is unbelievable. So, you can say what it's going to be.

老兄，你能详细说说吗？比如几周前你来了，因为一些削减已经在进行，对吧？但你来了之后评估了一切。我一直幕后深度参与 OpenAI 的很多部分，所以在这个特定领域走到前台是相对近期的事。不过，有趣的是，我实际上构建了 API 的第一个版本。所以，从 OpenAI 有产品以来我就在做产品，并且一直非常关心。有很多事情我可以说。

Dude, like can you expand on so like a few weeks ago you came in and because some of these cuts were already taking place, right? I mean, so but you've come in and like assessed everything. So I've always been very involved behind the scenes with, you know, many parts of OpenAI and so I think that kind of taking a foreground role is relatively recent in this particular area. Although, fun fact is I actually built the very first version of the API. So, like I've been doing products since product existed at OpenAI and always deeply cared about it. And there's a bunch of things I could say there.

我认为我们的使命有多核心，是我们开始构建产品之前没有意识到的，之后才意识到它有多重要。这就是为什么我一直如此贴近它。我们现在所处的阶段是，我们显然正处在一个向智能体过渡的时刻。毫无疑问。软件工程领域的人在过去六个月里已经感受到了这一点。在 2025 年期间，我认为有一个转变：从‘它就像自动补全’到‘你在编辑器里有一个侧边栏’，再到现在，你真正想要像 Codex 这样的工具，它实际上是一个智能体管理平台。智能体将处理所有细节和琐碎的工作。可能还有 20%的部分，比如如何组合东西、如何组织代码以及一些更高层次的事情，人类仍然关心并想要管理，但代码本身的细节——那正是智能体要做的事。

I think that how core it is to our mission was something we didn't appreciate before we started building products, and afterwards you realize how important it is. So that's why I've always been so close to it. The place we're at now is that we are clearly at a moment of transition to agents. No question. People in software engineering have been feeling this for the past six months. Over the course of 2025, I think there was a transition from 'it's like autocomplete' to 'you have a sidebar in your editor' to now, where you want a tool like Codex that is really an agent management platform. The agents will take care of all the details and the nitty-gritty work. There's still probably 20% of how you put things together, the way you structure your code, and some higher-level things that the human still cares about and wants to manage, but the details of the code itself—that's what agents are for.

我们面临的问题是：我们如何真正迎接这个时刻？因为这不仅仅是软件。我们看到了每个垂直领域的可能性——法律、金融。那里有一些机械技能，比如创建电子表格和演示文稿。我们如何让我们的模型在这些方面变得极其出色？你与领域专家合作，制作评估，生成训练数据，让 AI 真正运用其强大的领域知识，并在这些垂直领域应用，积累经验，从而弄清楚什么才算做得好。我们有如何做到这一点的机械愿景，但我们需要确保构建正确的产品界面来解锁这一切。

The question we had is: how do we really rise to this moment? Because it's not just software. We see line of sight for every single vertical—for law, for finance. Some of the mechanical skills there, like creating spreadsheets and presentations. How do we make our models extremely good at those? You work with domain experts, you produce evaluations, you produce training data, you have the AI actually take its great domain knowledge and apply it in these verticals, get experience, and figure out what good looks like. We have the mechanical vision of exactly how to do this, but we need to make sure we're building the right product surface to unlock all of it.

我们发现的一件事是，模型已经从产品本身转变为产品的一部分。过去，模型上面只有很薄的软件层，你不需要太费心思考如何架构。但现在，这是一个非常厚的层。你有技能、连接器、如何接入计算机使用、如何管理上下文和记忆等等。这层深厚的软件几乎就像是 AI 本身。就好像我们有了模型形式的大脑，现在我们在构建身体。两者都很难，必须协同设计。

One thing we have found is that the models have shifted from being the product to being a part of the product. They used to have very thin layers of software on top of them, and you didn't have to think that hard about how it was architected. But now, it's a very fat layer. You have things like skills, connectors, how you hook up to computer use, how you manage context and memory, and all of these things. There's this deep layer of software, which is almost like the AI. It's like we have this brain in the form of the model, and now we're building the body. Both are hard. They have to be co-designed together.

我们关注的很多方面，第一是打造一个出色的智能体平台。这是我们交付的首要重点。我们有团队在这方面执行得非常好。我对未来几周将要发布的内容感到非常兴奋。第二是，你实际上想在哪里应用这些智能体？我们的优先方向是计算机工作。我特意用了这个词。人们喜欢谈论知识工作，但没有人认为自己是一个知识工作者。这个词几乎脱离了实际。但我喜欢计算机工作的一点是，就像‘我真的不想做计算机工作’，这听起来不像是我想要的东西。但你会意识到你花了多少时间在做这件事——被拴在办公桌前，打字，弯腰驼背，得腕管综合征。所以我们专注于这一点，让今天已经存在的 Codex 不仅适用于软件工程师，而是真正让 Codex 为所有人服务。这很快就会到来。截至本期播客录制时，我们今天就会有一些更新。这个方向上还有很多令人兴奋的事情正在筹备中。第三是真正思考个人 AGI，这关乎现在的 ChatGPT，有十亿用户在使用。地球上的每个人都想要一个代表他们、拥有他们的上下文、他们建立了信任的 AI，它不仅仅是你一对一交谈的东西，而是可以为你做事。例如，也许它知道你喜欢某个音乐家，那个音乐家正在城里，而且门票刚刚有售。它就直接去买下来。也许它知道已经和你建立了信任，所以它知道你允许这样做而无需批准，或者它意识到应该先确认一下。我们也在构建这个。如果你思考这些，它们都融合成一个 cohesive 的整体。最终，快进到我们要去的地方，你真正想要的只是一个 AGI。你不想要语言模型，不想要线程，不想要任何这些细节。你只想要一个能帮助你、代表你行事、能帮你解决问题、知道你的目标并在工作和个人环境中实现这些目标的东西。这就是我们优先考虑和构建的。

A lot of what we're focusing on is, number one, getting together an amazing agentic platform. That is the number one focus we are delivering on. We have teams executing extremely well on this. I'm super excited about what we'll be releasing over upcoming weeks. The second is, where do you actually want to apply these agents? Our priority there is really towards computer work. I use that term very deliberately. People like to talk about knowledge work, but no one thinks of themselves as a knowledge worker. It's a term that's almost removed from what the actual thing is. But the thing I like about computer work is it's like, I don't really want to do computer work. That doesn't sound like the thing I want. But you realize how much of your time you do spend doing it—chained behind your desk, typing away, hunching over, getting carpal tunnel. So we are focusing on that, bringing Codex that exists today, not just for software engineers, but really making Codex be for everyone. That's something coming very quickly. We'll have some updates even coming today as of this podcast filming. And there are a lot of exciting things still in the pipeline for that direction. The third thing is really thinking about personal AGI, which is about ChatGPT right now, used by a billion users. Every single person on the planet is going to want an AI that represents them, has their context, that they have built trust with, that it's not just something you talk to one-on-one, but it can be out there doing things for you. For example, maybe it knows you like a specific musician, and that musician's in town, and tickets have just become available. It just goes and buys them. Maybe it knows it's built trust with you, so it knows you're allowed to do this without getting approval, or maybe it realizes it should check. We're building that as well. If you think about these things, they all fit together into a cohesive whole. In the end, you fast forward to where we're going, you really just want an AGI. You don't want a language model, you don't want threads, you don't want any of those details. You just want something that is helping you, operating on your behalf, able to help you solve problems, that knows what your goals are and achieve those in work and personal context. This is what we are prioritizing and building.

对我来说，人们问的一个重要问题是：‘我们如何看待消费者？我们如何看待企业？’答案是，如果你按照这些词现有的定义，我们非常关心消费者，也非常关心企业。但我认为这些词的含义将会改变和模糊，因为我们要做的是释放这一波创业浪潮。我们已经看到了它的前沿。小公司可以获得以前不可能的大量收入。这已经是一个趋势，而且只会真正加速。那是企业吗？那是消费者吗？它有点两者都不是。所以我认为我们真正专注于在所有情境下解决问题。这就是我们看待事物的视角。这意味着我们需要降低其他本身也很棒的事情的优先级。

To me, one of the important questions people are asking is, 'How do we think about consumer? How do we think about enterprise?' The answer is, if you take the definitions of these words as they exist today, we care a lot about consumer, we care a lot about enterprise. But I think the meaning of these words will change and blur, because what we're doing is we are going to unlock this wave of entrepreneurship. We're seeing the leading edges of it. Small companies can get tons of revenue that was not possible before. That's been a trend for a while, and it's just going to really accelerate. Is that enterprise? Is that consumer? It's kind of neither. So I think we are really focused on solving goals across all context. That is the lens we look at things through. That has meant we need to deprioritize other things that are also amazing on their own.

那么，什么被砍掉了？

So, what got cut?

是的。嗯，Sora 是最明显的一个。

Yeah. Well, Sora is the most obvious one.

为什么呢？因为这是技术树上的一个不同分支，对吧？你看，真正驱动 Sora 的模型并没有和核心的 GPT 系列统一。其次，用例也不太统一，对吧？它不完全属于那个目标，比如创意表达。那里有很重要的东西。我认为那是一个不可思议的模型。团队做了了不起的工作。而且我认为这项技术会延续到其他应用中。但我们真正关注的是我们想要交付的产品套件。我们未来 3 个月、6 个月、12 个月想做什么？顺便说一句，我描述的只是第一步。因为我们也看到了更强大模型的路线。比如，看看我们现在在数学上做的事情。好吧，其实有点令人震惊。我刚才提到的解决新的 Erdős 问题的结果，看起来真的很重要。那只是有人用了 GPT-4 Pro。就像两年前，

And why? Well, because it's a different branch of the tech tree, right? So, if you look at the models that actually power Sora, they're not unified with the core GPT series. And secondly, the use case is not quite unified either, right? It doesn't fall as far under this goal like there's creative expression. There's something very important there. I think it was like an incredible model. The team does incredible work. And I think that technology will live on for other applications. But what we were really focusing on was this the product suite that we want to be delivering. What do we want to be doing over the next 3 months, 6 months, 12 months? And the thing I described, by the way, is just step one. Because we also see line of sight to much more powerful models. Like, you look at what we're doing right now in mathematics. It's okay, it's actually kind of mind-blowing. Where this result that I just mentioned of solving the new Erdős problem that seems actually really significant. That was just someone using GPT-4 Pro. Like, 2 years ago

是的，我们以前喜欢训练我们的模型。我们有一个 20 人的团队，试图训练我们的模型去解决计算机奥林匹克竞赛，我们拿到了一枚铜牌。20 人的团队花了大约两周时间，用了大量算力。而现在，这个我们随便训练的模型，有人就能直接把它指向问题，得到这样的结果。如果你把它指向药物发现呢？如果你真的把那 20 人的团队和所有算力都用来推动科学发现呢？而这是目前没有人定价的。所以，我认为这是智能体崛起的时刻，真正要确保我们的产品投资结构良好，思考所有这些部分如何组合在一起，我们有连接器工作得很好，每个部分都可以组合，并且真正构建一个生态系统，因为这不仅仅关乎我们构建了什么，对吧？我们想构建一些示例智能体。你可以把 Codex 几乎看作一个示例智能体，但应该是，如果你是一个开发者，如果你是一个有创意想法的人，你可以构建你自己的智能体，对吧？你可以为你的应用、你的目的构建它。你关心特定的数学问题，你应该能够将智能体应用于那个数学问题。所以，我们正在实现这一切。但是，我的意思是，好吧，请稍等一下，快速过几件事。我的意思是，所以，我们，我的意思是，部分原因是你可能不得不为了算力而削减 Sora，对吧？所以，那些算力。

Yeah, we used to like train our model. We had a team of 20 people to try to train our models to go solve a computing Olympiad, and we got a bronze medal. A team of 20 people for like 2 weeks and lots of compute. And now it's just this model that we trained very casually. Someone is able to just point it at problems and get this kind of result. What if you point that at drug discovery? What if you do take that team of 20 people and all that compute and really try to push it for scientific discovery. And that's something no one is pricing in right now. So, I think it's rising to the moment of agents really trying to make sure that the product investments we're making are sort of well-structured, that we think about how all these pieces fit together, that we have connectors that work really well, and each of these pieces can be composed, and really also build an ecosystem because it's not just about what we built, right? That we want to build some example agents. You think of Codex almost as an example agent, but it should be that if you're a developer, if you're someone with a creative idea, you can build your own agent, right? That you can build it for your application, for your purpose. You care about the specific math problem, you should be able to apply the agent to that math problem. And so, we're enabling all of that. But you I mean, okay, just bear with me for 1 second just to go quick through a couple things around this. I mean, so we I mean, part of this was like you had to probably cut Sora for compute to get compute, right? So, that compute.

据报道，它消耗了大量算力。我的意思是，这个领域里任何成功的东西都会消耗算力。我的意思是，特别是你们两个，还有 Ilya 和其他一些人，你知道，以早期全力投入算力方向而闻名，而现在你们必须做出这些非常艰难的决定。你们必须服务所有现有的客户。你们得赚点钱，因为你们花了很多钱，所以，你们两个天生就倾向于下最大的赌注，但现在在某种程度上受到业务的约束。所以，这看起来非常困难。我觉得这不符合你的本性。是吗？你看起来有些怀疑。

Reported that it was taking up a lot of compute. I mean, everything in this field that is successful is going to take a compute. I mean, the two of you in particular, and I think Ilya and some others, you know, kind of famous for going all in on where you were going to direct the compute in the early days, and now you have to make these very difficult decisions. You have to serve all these customers that you have. You got to make some money 'cause you're spending a lot of money, and you know, so you two are both wired to take the very biggest bet possible, but you're constrained now by the business to some degree. So, that just seems very difficult. I feel like it's not in your nature. Does it? You're looking skeptical.

这种说法很奇怪，因为我不觉得受到业务的约束。我觉得业务赋能了我们，因为正是业务让我们能够说我们可以扩展算力，对吧？我记得我们推出 ChatGPT 的时候，之后我们就在讨论要买多少算力。我当时就说，我们得全部买下。我们得这么做，因为很明显需求巨大。我认为这极大地解锁了我们为关键任务获取大量算力的能力。

It's a strange phrasing because I don't feel constrained by the business. I feel enabled by the business because the business is what has really allowed us to say we can scale compute, right? I remember when we launched ChatGPT, and I remember we were talking right afterwards, and we're trying to figure out how much compute to buy. And I was just like, we got to buy it all. We just got to do it 'cause it's just so clear there's so much demand. And I think that that has been a huge unlock in our ability to get lots of compute for mission critical.

这个不可思议的收入机器，我们无法说服任何人我们应该获得所有这些算力。

This incredible revenue machine, we would not be able to convince anyone that we should get all this compute.

听着。但是，当你看到关于 Stargate 之类的报道时，给人的印象显然是你们在某种程度上缩减了基础设施。我不知道这是从哪来的。比如，会有一些站点，我们说，好吧，这个特定站点可能只有风冷，所以对我们来说不如另一个站点有价值，这就是具体情况，而人们真的想写缩减的故事。但很快又会变成 OpenAI 太鲁莽了。他们怎么能花这么多钱？所以媒体无论如何都会大惊小怪，我想只是因为你们需要写点东西。但我们会继续尽可能多地建设算力。还有一点我认为也值得思考：算力对我们来说不是成本中心，而是利润中心，对吧？当你把它部署到产品中时。所以在很多方面，我们的业务极其简单，对吧？我们租用或购买算力，然后以一定利润率转售。只要我们有一些正利润率，它就是可扩展的，对吧？因为需求是无限的。

Get this. But then the impression obviously when you see the stories about Stargate and things like that is that you guys are somehow pulling back on infrastructure. I don't know where that's coming from. Like there will be like a site here and there we say okay, you know what this particular site maybe it only has air cooling and so this is not as valuable to us as this other site and that's like there's specifics and people really want to write the story of like pulling back. But very soon it will be again like OpenAI is so reckless. How can they be spending this crazy amount? So like the media will flip out either way just 'cause you all need something to write about I guess. But we will keep building out as much compute as we possibly can. One thing that I think is also worth thinking about is that compute for us is not a cost center. It's a profit center, right? When you deploy it in the product. And so in many ways our business is extremely simple, right? We rent or buy compute and then we resell it at a margin. And as long as we have some positive margin on it, then it's scalable, right? Because the demand is just unlimited.

那么，数据中心硬件仍然一切就绪？或者你是说我们自己的芯片？是的，自己的芯片，网络。是的，对我们的芯片非常非常兴奋。我们有一个不可思议的团队。

So okay, so data center hardware still all systems go? Or you mean like our own chip? Yeah, own chip, networking. Yeah, very very excited about our chip. We have an incredible team there.

一个 Titan 一个？呃，不，我们当然不讨论时间线，但我只想说，我花了很多时间和那个团队在一起，我觉得它非常……所以他们做得很好。所以那方面一切就绪。机器人听起来也一切就绪……

One Titan one? Uh no, we're not talking about timelines of course, but I'll just say that like I spend a lot of time with that team and I think it's so... So they're doing great. So that's our all systems go. Robotics sounds like it's all systems...

还需要一段时间才能有让你觉得‘啊，这是 ChatGPT 时刻’的东西，但你没有缩减那个项目。

To be a while till we have something to where you're like 'ah this is the ChatGPT moment', but you haven't pulled back on that program.

社交网络？机器人不是当前的热点，但在未来显然会非常重要。

Social net... social network? Robots are not the current thing, but are going to be so clearly important in the future.

好的。那社交网络呢？不做那个。

Okay. And social network? Not doing that.

现在不做那个。

Not doing that right now.

然后显然超级应用、浏览器这些东西还在进行。

And then clearly like the super app, the browser, all that stuff still going.

是的，关于超级应用有一点要意识到，因为我觉得这是一个很吸引人的词，但我以为你们想出了这个。我的意思是，你们在内部简短地用了这个词。没错。要意识到的是，有时候我们在和团队沟通，然后当然它最终变成了对世界的沟通，而这些本意并非相同。

Yeah, and one thing to realize about super app because I feel like it's one of these things where it's like a catchy word, but I thought you guys like came up with this. I mean but you guys used this like internal short. Exactly. This is the thing to realize is that like sometimes we're communicating to our team and then of course it ends up being communications the world and these are not intended to be the same thing.

超级应用在很多方面就像一座冰山，对吧？是的，我们会有一个应用，你会看到今天 Codex 应用的更新，让它成为面向所有人的 Codex。我认为最终它会变成什么样子，我们还有很多步骤要走，但这实际上是关于我们正在构建我描述的那个统一身份基础设施，这将是人们想做任何事情的巨大解锁。

Super app in many ways is an iceberg, right? It's like yes, we're going to have an app and you'll see updates to what the Codex app is today to make it so it's Codex for everyone. And I think that ultimately what that becomes, we have a lot of steps to get to where we want to be, but it's really about saying we're building this unified identity infrastructure that I described, and that is going to be a huge unlock for every single thing that people want to do.

好的。你和我之前聊过一点，你们经历了很多戏剧性事件，有时拖慢了你们的步伐。遗憾的是，你们现在还有戏剧性事件。有一场诉讼要来了。会很有趣的。是的，是的。好吧，我们一会儿再谈那个。你认为过去两年哪家公司执行得更好？OpenAI 还是 Anthropic？

Okay. You and I have talked a little bit about yeah, you guys have had lots of drama. It's slowed you down at times. You still have drama sadly. There's a lawsuit coming. It'll be fun. Yes, yes. Well, let's talk about that in a second. Which company do you think has actually executed better over the last two years? Open AI or Anthropic?

你看，我认为仅从当前时刻很难说，对吧？因为在我看来，你需要退一步，真正思考我们所有人在这里是要做什么。从 OpenAI 的角度来看，我们一直在谈论的事情在某些方面很清楚：是的，向企业销售，比如弄清楚如何真正交付这些编码工具，而且我认为竞争确实可以帮助提升你自己的思维，意识到我们需要专注于这一点。在编码方面的一个例子是，我认为我们在构建不仅在抽象编码上表现良好的模型方面起步较晚。比如我们在编程竞赛中一直有最好的成绩，但你还需要将它们应用于混乱的代码库、真实世界的数据等等。这是我们比 Anthropic 更晚意识到的事情。所以，我认为这归功于他们，但也帮助我们提升了自身的执行力。现在，正面比较 Codex 和 Claude，我认为我们得到了非常有利的结果。而且我认为我们做得非常出色，我们的团队在整个公司都做得非常出色，真正构建了一个不仅具有竞争力，而且在很多方面都领先的产品。但是，我认为核心从来不是新闻周期的起伏，而是朝着 AGI 的进步，朝着让所有人受益的方向。这是我们一直非常专注的事情。团队执行得非常好。所以，我只想说，从外部你并不总能看出来，但如果你想想我们讨论过的事情，这种专注是存在的，而且在多个时间尺度上，它们都汇聚到我们需要去的地方。

Look, I think that this is a hard thing to say from just the current moment, right? Because I think that the view in my mind is you got to step back and really think about what is it that we're all here to do, right? And I think that from the Open AI perspective, the things we've been talking about in some ways are about it's very clear that yes, selling to enterprises like figuring out how to really deliver these coding tools and it's been really, I think competition can really help elevate your own thinking and realize that hey, we need to focus on this. And one example in coding was that I think we got late to the game of not just building models that were good in the abstract of coding. Like we always had the best numbers on programming competitions, but you also need to apply them to messy repos, real-world data, those kinds of things. And that was something that I think we had appreciated later than Anthropic did. So, I think that's kudos to them, but also something that's helped elevate our own execution. And now, head-to-head, Codex versus Claude, I think that we get very favorable results. And I think we've done an incredible job, our teams have done an incredible job across the whole company to really build a product that's not just competitive, but actually ahead in many, many ways. But, I think that the core, it's never about the ups and downs of the news cycle. It's about progress towards AGI, towards it benefiting everyone. And that is something that I think we've been extremely focused on. The team's executing extremely well. And so, I just want to say that you can't always tell from the outside, but if you think about the things we talked about, that there is this focus, but on many timescales that all add up to where we need to go.

是的，在剩下的时间里我想问的问题有一百万个。但其中一个我们触及过，比如确保很多模型掌握在很多人手中，他们拥有那种强大的技术。但是，我们已经到了一个点，现在有些模型被说成太强大了，它们只对某些公司开放。Claude 和 Mythos 确实上了很多头条，我认为带来了更多的恐惧。我想知道你们如何看待这个新时刻。我们是否到了需要开始把这些强大模型藏在幕后而不是交到每个人手中的地步？

Yeah, there's a million questions I want to ask with the time we have left. But, one of them we touched on this, like making sure a bunch of models are in the hands of a bunch of people and they have that powerful technology. But, we've reached a point where now some of these models are too powerful, we're being told, and they're gated for only certain companies. Claude and Mythos have really made a lot of headlines and I think a lot more fear. And I'm wondering what you guys think of this new moment. Are we getting to a point where like we need to start keeping these powerful models behind the scenes rather than in the hands of everybody?

世界上有些人长期以来一直想把 AI 掌握在少数人手中。你可以用很多不同的方式证明这一点。其中一些是真实的，比如会有合理的安全担忧。但是，我认为如果你想要的是‘我们需要控制 AI，只有我们，因为我们是值得信赖的人’，那么基于恐惧的营销可能是证明这一点最有效的方式。这并不意味着在某些情况下它不合理。但是，你知道，说‘我们造了一颗炸弹。我们就要把它扔到你头上。我们会卖给你一个价值 1 亿美元的防空洞。你需要跑遍你所有的东西，但只有当我们选你为客户时才行’，这显然是令人难以置信的营销。我们看待平衡这些即将到来的新能力的方式，仍然相信世界需要获取、使用、理解并为这项技术提出新想法，这并不总是容易的。我们在准备框架中早就有了网络安全，我们一直在构建缓解措施，以弄清楚如何发布这些模型，如何将这些模型交到可信访问计划中，然后如何将更强大的模型交到每个人手中。但是，会有更多关于模型太危险而不能发布的言论。也会有非常危险的模型必须以不同的方式发布。但是，正如 Greg 所说的，这里的目标是让所有人受益。而且，我不想说以某种方式营销，而是让世界和我们一起踏上这段旅程，就像我们将给你更强大的技术。随之而来的是责任。我们将尽最大努力帮助世界为成功做好准备。但是，我们会尽可能避免基于恐惧的营销。

There are people in the world who for a long time have wanted to keep AI in the hands of a smaller group of people. You can justify that in a lot of different ways. And some of it's real, like there are going to be legitimate safety concerns. But, I expect if what you want is like, we need control of AI just us cuz we're the trustworthy people, I think the fear-based marketing is probably the most effective way to justify that. That doesn't mean it's not legitimate in some cases. But it is, you know, clearly incredible marketing to say, 'We have built a bomb. We are about to drop it on your head. We will sell you a bomb shelter for $100 million. You need to like run across all your stuff, but only if we like pick you as a customer.' And the way that we view balancing these new capabilities that are going to come with, still our belief that the world needs to get and use and understand and come up with new ideas for this technology is not always easy. We have had cybersecurity in our preparedness framework for a long time, and we have been building mitigations to figure out how we release this, how we put these models in the hands of a trusted access program, how we then put more capable models in the hands of everybody. But, there will be a lot more rhetoric about models that are too dangerous to release. There will also be very dangerous models that will have to be released in different ways. But, to the point Greg was making about the goal here is to benefit everybody. And also to, I don't want to say market this in a way, but get the world to come along on this journey with us where it's like we are going to give you more powerful technology. There's going to be responsibility that goes along with that. We are going to help set up the world for success as much as we can. But, we will try to avoid the fear-based marketing as much as we can.

所以，我直接问你。你认为 Mythos 只是很多营销，而不是……我确定它是一个很好的网络安全……

So, I'll just ask you directly. Do you think Mythos is just a lot of marketing and not... I'm sure it's a great cybersecurity like...

我们谈论这个已经很长时间了。这已经在我们的模型中，但有一种说法是，我们的版本是：这些模型在网络安全方面会变得更好。我们有这个准备框架类别。这是我们如何将其部署到世界上的计划。这是可信访问计划的样子。这是我们对模型施加的缓解措施。我想我是在说它并不是他们准备框架中的一个类别。所以，我确定 Mythos 是一个很好的网络安全模型。但是，我认为我们有一个我们感觉良好的计划，关于如何将这种能力推向世界。

We've been talking about this for a long time. Like this has been in our model, but there's a way of saying like our version of this is to say these models are going to get much better at cyber. We have this preparedness framework category. Here's our plan for how we deploy this into the world. Here's how this trusted access program looks like. Here's the mitigations we put on the models. I think I'm talking about how it is not how it is one of the categories in their preparedness framework. So, I'm sure Mythos is a great model for cybersecurity. But, I think we have a plan we feel good about for how we put this kind of capability out into the world.

当 Anthropic 的事情与战争部发生纠葛时，我的意思是，从我的角度看，似乎有 David Sacks 和与 Elon 有长期联系的人。我可以看到他们在施加压力促成这些事情，并让政府关注这一点。你们很快就跟进并发布了自己的公告，更有利一些。我的意思是，你知道，Elon 也对你们有攻击性。其中一些对我来说很有趣，我的意思是，有一个世界，你们和 Anthropic 实际上站在另一边，感觉 Elon 甚至 Zach 在某种程度上，据我所知，在施压。顺便问一下，是的，我的意思是，你觉得 Anthropic 在那件事中受到了公平对待吗？你觉得……

When the Anthropic stuff was going down with the Department of War, I mean, from my perspective, it looked like you had sort of David Sacks and people with long ties to Elon. I could see putting pressure to make some of these things happen and have the government focus it on that. You guys kind of came in pretty quick after and had an announcement of your own that was more favorable. I mean, you know, Elon's aggressive against you guys as well. And some of this is funny to me that I mean, there is a world where like you guys and Anthropic are actually on this other side and it feels like Elon and even Zach to some degree, my understanding, kind of pressuring things. As a side, yeah, I mean, did you feel like Anthropic was treated fairly in that? Do you feel...

不。

No.

我认为 Anthropic 没有得到公平对待。你能详细说说吗？哪里特别不对？

I don't think Anthropic was treated well. Can you explain more? What was particularly wrong?

随着后来发生的事情，比如这些模型达到了一个明显关乎国家安全的网络安全阈值，我觉得整个情况有了不同的味道。威胁使用《国防生产法》和实际运用供应链风险认定——这不是我们的政府和 AI 努力应该有的关系。我们非常关心支持美国政府，随着模型能力越来越强，这也会变得越来越重要。我不认为实验室说'我们有超级武器，不跟你们合作保卫国家'是个好立场。但我也觉得政府不应该在媒体上打口水仗，用那个应该很少用的大锤子。在 OpenAI，我们努力保持温和、中间和理性。这是我们试图与美国政府相处的方式。但我看不到任何好的未来，如果领先的 AI 努力不协助美国政府的话。如果我们关于模型走向的说法是对的，政府需要我们的帮助，我们很荣幸能提供。

With the things that have happened since, like these models reaching a cybersecurity threshold that is clearly in the national security interest, I think it all has a different flavor. Threats of using the DPA and actually using supply chain risk designation—this is not the relationship our government and AI efforts need to have. We really care about supporting the US government, and that will become increasingly important as models get more capable. I don't think it's a good stance for labs to say, 'We have the super weapon, and we're not going to work with you to help defend the country.' But I also don't think it's good for the government to fight this out in the press and use the big hammer that should be used very rarely. At OpenAI, we try to be moderate, centrist, and reasonable. That's what we've tried with the US government. But I see no good future where leading AI efforts don't assist the US government. If everything we say is right about where models are going, the government needs our help, and we are honored to provide it.

我刚开始写科技的时候，大战是微软对所有人。现在在 AI 领域，这太疯狂了——这么多事情似乎都跟个性有关：你、埃隆、扎克、达里奥、德米斯，带着各种敌意和世界观。你写过这种莎士比亚式的戏剧。我们怎么才能走出来？我看不到出路，而且还有一场诉讼要来。

When I was first writing about tech, the big war was Microsoft versus everybody else. Now in AI, it's insane—so much seems tied to personalities: you, Elon, Zuck, Dario, Demis, with animosities and world views. You wrote about the Shakespearean drama. How do we get past this? I don't see a path out, and there's a lawsuit coming.

有些参与者只相信自己能做好，因为他们认为赌注无限大，而且觉得别人做不好。这导致了有毒的行为。我们无法控制别人，但我们会继续倡导把这当作一个集体项目，人类必须一起做好，而不是关于某个人或某种意识形态的输赢。

Some people involved only trust themselves to get it right because they think the stakes are infinite and they don't think anybody else can do it right. That leads to toxic behavior. We can't control others, but we will continue to advocate for doing this as a collective project that humanity has to get right together, not something about one person or one ideology winning.

我们不能不提你上周的个人遭遇。有很多科幻小说写过那些不想看到技术进步派系。有种说法是开关正在翻转。那些想阻止的人……我读过理查德·克拉克的书，里面有个类似比尔·乔伊的人物认为 AI 不能发生，到处摧毁数据中心。感觉事情不能再紧张了。这一定很可怕。

We'd be remiss not to mention what happened to you personally last week. There are tons of sci-fi books about factions that don't want technological progress. There's an argument that the switch is flipping. With people who want to stop it... I've read Richard Clarke's book where a Bill Joy-like figure decides AI can't happen and destroys data centers. It feels like things couldn't be more heightened. It must be horrifying.

可怕。我没什么深刻的话要说。那样醒来太疯狂了。第一天我处于肾上腺素冲击状态，试图搞清楚后勤。第二天，我极度沮丧，经历了一个抑郁周期。这非常可怕。我觉得末日论调没有帮助。某些其他实验室谈论我们的方式也没有帮助。我真的不想附带声明，但 Anthropic 谈论 OpenAI 的方式没有帮助。我希望冷静的时刻会到来。

Horrifying. I don't have anything deep to say. That was a crazy way to wake up. The first day I was in adrenaline shock, trying to figure out logistics. The day after, I was incredibly disheartened and went through a depressive cycle. It's very scary. I think the doomerism talk hasn't helped. The way certain other labs talk about us hasn't helped. I actually don't want to make a side statement, but the way Anthropic talks about OpenAI doesn't help. I hope cooler times will prevail.

有一件事让我印象深刻，就是 Sam 在这段时间里一直坚持推进使命。这种韧性的程度非常极端，而且被低估了。我也看到了同样的事情。很难想到在过去三年里还有谁经历过更戏剧性的商业和个人事件。

One thing I've been impressed by is how Sam, throughout this time, kept pushing on the mission. The degree of resilience is extreme and underappreciated. I've seen the same thing. It's hard to think of another individual over the last three years who's been through more dramatic business and personal events.

叙事确实很戏剧化。人们可能会说'你很有钱，所以别求同情'。这没关系。事情发生的前一晚，我请了一些人吃晚饭，我们在讨论下一阶段。他们说我在媒体上经历了一段残酷的时期，但某种程度上也是好事。我说：'至少没人想杀我。'这让我看清了：人们可以说刻薄的话，但只要他们只是说说，就没那么严重。你继续前进。

The narrative has been quite dramatic. People might say, 'You're rich, so don't be sympathetic.' That's fine. The night before this happened, I had some people over for dinner, and we were talking about the next phase. They said it's been a brutal time for me in the press, but I guess it's good in some sense. I said, 'At least no one tried to kill me.' That put into perspective: people can say mean things, and as long as they only say mean things, it's not that big a deal. You keep going.

你一直告诉我你的梦想是有一天退休。那为什么不现在呢？

You always told me your dream is to retire out in that someday. So why not now?

我们还有很多工作要做。但很明显现在有人会实现 AGI。我们有大概五家公司有可能做到这一点。

We got a lot of work left. But it's clear somebody's going to get to AGI now. We've got like five companies that could reasonably do this.

你最希望哪家公司最先实现？归根结底是这样吗？

Which company would you most like to get there first? Is that what it comes down to?

听着，我认为缺失的一点是如何帮助人们真正理解这项技术能为他们做什么。每家公司都可以为此做出贡献。我们有这方面的见解，在这个播客里也谈了很多。从根本上说，所有试图创造这项技术的人都有责任展示它的好处，以及人们为什么应该想要它。为什么人们应该保护和捍卫创造它的能力？为什么美国需要在这方面保持领导地位？为什么这不仅对国家好，对你个人、你的未来、你孩子的未来也好？这是我们每天醒来都在思考的事情。我觉得这并不夸张。我们一直在谈论这件事。我们思考它，无论是公司的奇怪创意法律结构——我们做过很多次，因为我们试图解决这个使命。如果其他人想贡献，那就更好了。这是我们都应该做的事情，但它独特地驱动着我们。我们深信，如果我们能提供让所有人繁荣的技术，如果我们能赋予人们对未来更多的主动权，那么经过一些波折，这将带来一个更美好的世界。我不认为所有从事这个领域的人都相信这一点，但任何相信的人，我们都很乐意与之合作，这就是我们希望推动世界前进的方向。

Look, I think the missing thing is how to help people really understand what this technology can do for them. Every company can contribute to that. We have a perspective on it, and we've talked about it a lot on this podcast. Fundamentally, it's the responsibility of everyone trying to create this technology to also show its benefits and why people should want it. Why should people protect and defend the ability to create it? Why does America need to have leadership here? Why is this good for not just a country, but for you personally, your future, and your kids' future? That's something we wake up thinking about every day. I don't think that's an overstatement. We talk about it constantly. We think about it, whether it's weird creative legal structures for a company—we've done that many times because we're trying to solve for this mission. If other people want to contribute, more power to them. That's something we should all be doing, but it drives us uniquely. We so believe that if we can deliver technology that enables prosperity for everybody, if we can give people more agency over the future, that will, through some fits and starts, lead to a better world. I don't think all people working in the field believe that, but anyone who does, we are delighted to work with, and that's what we want to move the world towards.

你如何看待这场审判的生死攸关程度？

How existential do you view this trial?

我实际上认为这是一个真正的机会，让我们讲述自己的故事。因为回顾 OpenAI 的历程，当出现分歧时，我们总是让另一方来讲述故事。我们一直尽量不去说'嗯，那不是真的'。让我们谈谈真相。这次，我们终于别无选择，对吧？因为我们必须为自己辩护。我们必须说出真相。我们必须讲述发生了什么。我为此感到非常自豪。我花了很多时间回顾历史和各种信息。当然，总有一些事情你可以试图断章取义，比如'啊哈，你说过这个'。

I actually think it's a real opportunity for us to tell our story. Because if you look at the course of OpenAI, we have really let the other sides, when there are splits, tell the story. And we have done our best not to say, 'Well, that's not really what happened.' Let's talk about the truth. In this case, we finally have no choice, right? Because we have to defend ourselves. We have to tell the truth. We have to tell what happened. And I'm extremely proud. I've spent a lot of time looking back at the history and a bunch of different messages. Of course, there are always things you can try to cherry-pick, like 'Aha, you said this thing.'

你的日记很有名。

Your diaries are famous.

我知道，对吧？但问题是，它们不是——首先，这是极其私人的文件。如此私人的东西被夺走，然后被试图武器化，这非常痛苦。但那些特定的句子是对方能找到的最糟糕的东西。你会想，'真的吗？'问题是，它们全都被断章取义了。比如，'嘿，我们正在谈判中。我们都同意 OpenAI 唯一的出路是营利性。'Sam、Ilya、Greg、Elon——我们都同意这一点。我们都说过这是我们必须做的。这完全是为了使命。而现在你陷入了这场疯狂的谈判。Elon 说，'你需要多数股权。你需要当 CEO。你需要完全控制。'我们差点就同意了。就像，'好吧，行。我们不会成为平等合伙人。你需要这么多股权。如果你说你需要，我们希望你参与。我们可以做到。好吧，行。Sam 当 CEO。Elon 当 CEO。他需要这样，这样每个人都知道他负责。行。但绝对控制。对 OpenAI 的绝对控制。即使你说，'嗯，我会稀释股份。我将来会放弃。'你会想，'我们的使命是什么？我们真的相信我们的使命吗？我们真的在乎这幅图景吗？我们希望这项技术惠及所有人，不应该有一个人掌控整个未来。不管那个人是谁。'那就是破裂点。那就是导致我们说不的原因。所以，多年来我们从未讲过这个故事。但现在我们会讲。所以，我认为这是一个真正的机会，让人们理解真正驱动我们的是什么，我们真正代表什么。我认为他这样做很疯狂，但我现在的恐惧是他在审判前决定撤诉，我们就没机会做这些了。但我很乐意向世界解释这一切，把这一章翻过去。

I know, right? But the thing is, they're not—first of all, incredibly personal documents. Extremely painful to have something so personal taken from you and then attempted to be weaponized. But those particular sentences are kind of the worst thing the opposition could find. You're like, 'Really?' And the thing is, they're all taken out of context. The question of, 'Hey, we're in the middle of this negotiation. We've all agreed the only path forward for OpenAI is a for-profit.' Sam, Ilya, Greg, Elon—we all agreed on this. We all said this is what we have to do. This is literally the thing for the mission. And now you're in this crazy negotiation. Elon's like, 'You need majority equity. You need to be CEO. You need full control.' And we got so close. It's like, 'Okay, fine. We're not going to be equal partners. You need this massive amount of equity. If that's what you say you need, we would like you to be involved. We can get there. Okay, fine. Sam will be CEO. Elon will be CEO. He needs it so everyone knows he's in charge. Fine. But absolute control. Absolute control over OpenAI. Even if you say, 'Well, I'll dilute down. I'll give it up in the future.' And you're like, 'What is our mission? Do we really believe in our mission? Do we really care about this picture of we want this technology to benefit everyone, and there shouldn't be one person in charge of the whole future? Doesn't matter who that person is.' That was the breaking point. That was the thing that caused us to say no. So, we've never told that story for years. But now we will. So, I think it is a real opportunity for people to understand what truly motivates us, what we truly stand for. I think it's insane that he's doing this, but my fear at this point is he decides to drop the case right before the trial, and we don't get to do all this. But I am happy to explain all this to the world and have this chapter behind us.

我的问题回到这个播客的开头。我谈到作为 AI 记者，我成长过程中听到的叙事是'我们都会死，我们都会失业'。我觉得我一次又一次听到这个，现在你个人笔记本的部分内容公开了。我很好奇，如果事后诸葛亮，你会如何以不同的方式谈论这些？

My question rounds back to the beginning of this podcast. I was talking about this narrative I feel like I as an AI reporter grew up with hearing, 'We're all going to die, and we're all going to lose our jobs.' I feel like I heard that time and time again, and now portions of your personal notebook are public. I'm curious, how would you have talked about this differently with hindsight being 20/20?

嗯，我认为我们今天思考的方式就是我希望我们当时能谈论的方式。但我不知道以当时的认知我们能否做到。例如，甚至有些部分与技术本身有关。我们在 2017、2018 年有一个设想，构建 AGI 的方式是通过竞争性的多智能体模拟。想象一个有一千个智能体的岛屿，它们都在为生存和复制而战斗。你可以看到，如果你投入大量算力，它可能会构建出非常智能的东西。但那个智能的东西与人类价值观毫无关联。你必须有一个单独的步骤来弄清楚如何与这个东西对话。它完全不是在现实世界中成长的，没有任何语言概念，与我们的现实没有任何联系。它只是聪明而强大。那是一个非常可怕的系统。你从一开始就要思考如何可能对齐它。但相反，我们走了语言模型这条路，它植根于我们的价值观，植根于理解人类，而且我们有思维链，实际上可以监控，如果你在技术意义上处理得当，你就有路径让它变得可信。所以这真正代表了 AI 的驱动力是什么。

Well, I think that the way we think about it today is the way I wish we had talked about it in some sense. But I don't know if we could have with the knowledge we had at the time. For example, even some of it is related to the technology itself. We had a picture in 2017, 2018, the way we would build AGI was through a competitive multi-agent simulation. Imagine an island of a thousand agents that are all in a battle to survive and replicate. And you could see that if you put a ton of compute into it, maybe it would build something very smart. But that smart thing would not be connected to human values at all. You'd have to have a separate step of figuring out how to even talk to this thing. It didn't grow up at all in the real world, doesn't have any concept of language, doesn't have any connection to our reality. It's just smart and powerful. That's a very scary system. You start from a place of trying to think about how you could possibly align it. But instead we have this language model route, which is rooted in our values, which is rooted in understanding humans, and we have chain of thought that you can actually monitor, that if you approach things right in a technical sense, you have a path to make that be faithful. So it really represents what is truly motivating the AI.

你意识到我们有一条完全不同的技术路径来实现我们讨论的目标，而且这条路径要乐观得多。所以我认为，我们需要一些技术上的学习，才能真正理解这项技术会是什么样子，如何被创造出来，以及如何让它变得有用。这是我们过去无法理解，但现在明白了的事情。这是一场精彩的讨论。非常感谢你们两位一起出席。我们是一个谦逊的播客，很荣幸你们能花时间参与。我们显然涵盖了很多内容，但要点是：在相对较短的时间内，新模型真的很好。好了，非常感谢你们。

And you just realize we have a totally different technological path to get to the outcome we're talking about, and it's a much more optimistic one. And so I think that there was some technical learnings we had to have for what is this technology truly going to be, how will it be created, what are the ways in which you make it useful. And that is something I think we could not have appreciated, but we do today. It's been a fascinating discussion. It's really kind of you guys to both show up and do this together. We have been a humble podcast. We're honored for you guys to spend this time. And I guess I mean the obviously we covered a lot of ground, but the headline is in relatively short order, new models are really good. Well, thank you guys. Thank you guys so much.

谢谢你们让新模型对大家有用。谢谢你们。谢谢。谢谢。

Thank you for having new models useful for everyone. Thank you guys. Thank you. Thank you.

《核心记忆》播客由我 Ashley Vance 和 Kylie Robinson 主持，或者我们两人一起。节目由我和 David Nicholson 制作。主题曲由 James Mercer 和 John Sortland 创作，节目始终由 John Sortland 剪辑。非常感谢 Brex 和 Combinator Ventures 的支持，最感谢的是所有收听或观看的观众。我们爱你们。请给我们点赞、评论、订阅，所有这些了不起的事情。谢谢，我们下次再见。

The Core Memory podcast is hosted by me, Ashley Vance, and or Kylie Robinson, or both of us together. It is produced by me and David Nicholson. Our theme song is by James Mercer and John Sortland, and the show is edited, always, by the John Sortland. Thank you so much to Brex and Combinator Ventures for all your support, and thank you most of all to everybody for listening or watching. We love you. Please leave us a like, a review, a subscribe, all those tremendous things. Thank you and we'll see you again.