Sam Altman 谈星际之门、用 ChatGPT 育儿以及 AGI 的未来

欢迎收听 OpenAI 播客。我是 Andrew Maine。有几年时间，我在 OpenAI 工作，先是应用团队的工程师，后来担任科学传播者。之后，我与一些公司和个人合作，探索如何整合人工智能。通过这个播客，我们有机会与 OpenAI 内部及合作的人聊聊幕后故事，或许还能一窥未来。我的第一位嘉宾是 Sam Altman，OpenAI 的 CEO 兼联合创始人。我们将了解更多关于 Stargate 的信息，他作为家长如何使用 ChatGPT，以及 GPT-5 何时到来。每年都会有越来越多的人认为我们已经达到了 AGI 系统。你对硬件和软件的需求正在迅速变化。但如果人们知道我们用算力能做什么，他们会想要更多得多。

Welcome to the OpenAI podcast. My name is Andrew Maine. For several years, I worked at OpenAI first as an engineer on the applied team and then as the science communicator. After that, I worked with companies and individuals trying to figure out how to incorporate artificial intelligence. With this podcast, we have the opportunity to talk to the people working with and at OpenAI about what's going on behind the scenes and maybe get a glimpse of the future. My first guest is Sam Altman, CEO and co-founder of OpenAI. And we're going to find out a bit more about Stargate, how he uses ChatGPT as a parent, and maybe get an idea of when GPT-5 is coming. More and more people will think we've gotten to an AGI system every year. What you want out of hardware and software is changing quite rapidly. But if people knew what we could do with compute, they would want way, way more.

我有个朋友刚当父母，经常用 ChatGPT 提问。它已经成了一个很好的资源。你也是新晋家长。ChatGPT 在这方面帮了你多少？

One of my friends is a new parent and is using ChatGPT a lot to ask questions. It's become a very good resource. And you are a new parent. How much has ChatGPT been helping you with that?

帮了很多。我是说，显然人类在没有 ChatGPT 的情况下也能照顾婴儿很久了。我不知道没有它我该怎么办。头几周，几乎每个问题，我都在不停地问。现在我问得更多的是关于发育阶段的问题，因为基本的东西我已经会了。比如“这正常吗？”之类的。但它在这方面超级有用。我花了很多时间思考我的孩子将来会如何使用 AI。顺便说一句，这有点像，孩子特别多。我觉得每个人都应该多生孩子。是的，我在 OpenAI 的很多朋友，前同事和现同事，都在生孩子，人们会问“那 AI 怎么办？”我认识的所有内部人士都非常乐观，都在组建家庭。我认为这是个好兆头。

A lot. I mean, clearly people have been able to take care of babies without ChatGPT for a long time. I don't know how I would have done that. Those first few weeks, it was like every question, I mean constantly. Now I kind of ask it questions about like developmental stages more because I can do the basics. But is this normal? Yeah. But it was super helpful for that. I spend a lot of time thinking about how my kid will use AI in the future. It is sort of like, by the way, extremely kid-filled. I think everybody should have a lot of kids. I'm yeah, a lot of my friends at OpenAI, former colleagues and current ones, are having kids and people go like, oh what about this AI thing? Everybody I know inside is very optimistic and having families. I think that's a good sign.

是的。比如，我的孩子永远不会比 AI 更聪明，但他们也会成长得……我是说，他们会成长得比我们这一代能力强大得多，能够做我们无法想象的事情，而且他们会非常擅长使用 AI。显然我经常思考这个问题。但我更多考虑的是他们将拥有而我们没有的东西，而不是什么会被剥夺。我不认为我的孩子会因为不如 AI 聪明而烦恼。有一个视频一直让我印象深刻：一个婴儿或蹒跚学步的孩子拿着一本旧的光面杂志，在屏幕上这样划，因为它以为那是 iPad。它以为那是个坏掉的 iPad。现在出生的孩子会认为世界一直都有极致的 AI，他们会极其自然地使用它，并且会回望这个时代，觉得它非常原始。

Yeah. Like my kids will never be smarter than AI, but also they will grow up way to set them back there though. I mean they will grow up like vastly more capable than we grew up and able to do things that we just cannot imagine and they'll be really good at using AI. And obviously I think about that a lot. But I think much more about the like what they will have that we didn't than what is going to be taken away. I don't think my kids will ever be bothered by the fact that they're not smarter than AI. There's this video that always has stuck with me of a baby or a little toddler with one of those old glossy magazines going like this on the screen because it thinks it's an iPad. Thought it was a broken iPad. Kids born now will just think the world always had extreme AI and they will use it incredibly naturally and they will look back at this as like a very prehistoric time period.

我在社交媒体上看到一件事：有个人说他对跟孩子聊托马斯小火车感到厌倦了，于是把它放进了 ChatGPT 的语音模式。孩子们喜欢语音模式。结果一个小时后，孩子还在聊托马斯火车。不过，我怀疑这不会全是好事。会有问题。人们会发展出一些有点问题、甚至非常问题的准社会关系，社会将不得不找出新的护栏。但好处将是巨大的，而且社会总体上善于找到减轻负面影响的方法。所以，是的，我持乐观态度。

I saw something on social media where a guy talked about he got tired of talking to his kid about Thomas the Tank Engine, so he put it into ChatGPT voice mode. Kids love voice mode. And he was like an hour later the kid still talking about Thomas the train. Again, I suspect this is not all going to be good. There will be problems. People will develop these sort of somewhat problematic or maybe very problematic parasocial relationships and society will have to figure out new guardrails. But the upsides will be tremendous and society in general is good at figuring out how to mitigate the downsides. So yeah, I think optimistic.

我们看到一些有趣的数据：在有优秀老师和课程的情况下，ChatGPT 在课堂上单独使用效果很好；但单独作为作业拐杖，可能导致孩子只是像用谷歌一样搜东西。我就是那种孩子，大家都担心谷歌一出来我就会搜所有东西，停止学习。结果发现，学校里的孩子适应得相当快。所以我认为我们会解决这个问题。想想如果你不什么都谷歌，你会成为什么样的人，Sam。

We're seeing some interesting data where used alone in classrooms with a good teacher, good curriculum, ChatGPT becomes very good; used solely by itself as sort of a homework crutch can lead to kids just doing the same thing as trying to Google stuff. I was one of those kids that everyone was worried I was just going to Google everything when it came out and stop learning. And it turns out relatively quickly kids in schools adapt. So I think we'll figure this out. Think of what you could have become if you didn't Google everything, Sam.

你知道，我们看到这些采用数据，真是疯狂。它是 OpenAI 最受欢迎的产品。五年后，它还会是 ChatGPT 吗？我是说，我认为五年后 ChatGPT 将完全变成另一个东西。所以从某种意义上说，不是。但它还会叫 ChatGPT 吗？很可能。是的。好吧，所以它还是个名字。

You know, so we've seen these adoption figures which are really insane. It's OpenAI's most popular product. Five years from now, is it going to be ChatGPT? I mean, I think ChatGPT will just be a totally different thing five years from now. So, in some sense, no. But will it still be called ChatGPT? Probably. Yeah. Okay. So, it's still a name.

我们常听到的另一个词是 AGI，我想听听你对 AGI 的定义。从很多意义上说，如果你五年前让我或任何人基于软件的认知能力提出一个 AGI 的定义，我认为当时很多人给出的定义现在已经被远远超越了。这些模型现在很聪明，对吧？而且它们会越来越聪明，不断进步。我认为每年都会有越来越多的人认为我们已经达到了 AGI 系统。尽管定义会不断推后，变得更加雄心勃勃，但更多人会同意它。但是，我们现在已经有了真正提高人们生产力、能够做有价值经济工作的系统。也许更好的问题是，什么才能被称为超级智能？

The other thing we hear is AGI, which I'd like to hear your definition of AGI. In many senses, if you asked me or anybody else to propose a definition of AGI five years ago based off like the cognitive capabilities of software, I think the definition many people would have given then is now like well surpassed. These models are smart now, right? And they'll keep getting smarter. They'll keep improving. I think more and more people will think we've gotten to an AGI system every year. Even though the definition will keep pushing out and getting more ambitious, like more people will still agree to it. But, you know, we have systems now that are really increasing people's productivity that are able to do valuable economic work. Maybe a better question is what will it take for something I would call superintelligence?

如果我们有一个系统能够自主发现新科学，或者极大地增强人们使用工具发现新科学的能力，那对我来说几乎就是定义上的超级智能，而且我认为这对世界来说是一件美妙的事情。所以基本上，这是一个梯度，它不断变得更好，我们的每个定义都会说“哦，这感觉像……”当我们在内部使用 GPT-4 时，我就觉得“这里有 10 年的跑道，我们可以用它做很多事情”，甚至当它开始自我使用时，比如你遇到一个非常强大的推理能力，但当你看到它提出某个新定理或证明，然后“哦，我们找到了更好的癌症疗法”或“我发现了一种新的 GLP 药物”之类的。我坚信，人们生活改善的更高层次是更多的科学进步。那正是限制我们的东西。所以如果我们能发现更多，我认为那将产生非常重大的影响，对我来说那将是一个极其激动人心的里程碑。我认为 AI 还会有许多其他伟大的用途，但这一点感觉特别重要。

If we had a system that was capable of either doing autonomous discovery of new science or greatly increasing the capability of people using the tool to discover new science, that would feel like kind of almost definitionally superintelligence to me and be a wonderful thing for the world I think. So basically a lot of it's kind of this gradient where it keeps getting better and better and each one of our definitions go oh this feels like that way when we hit GPT-4 internally playing with this I'm like there's 10 years of runway that we can do so much stuff with this and even when it starts using itself like you can hit a reasoning that was really capable but when you're saying it comes out with some new theorem or proof or something and then oh hey we found a better cure for cancer or I found out some new GLP drug or something. I am a big believer that the higher order bit of people's lives getting better is more scientific progress. Like that is kind of what limits us. And so if we can discover much more I think that really will have a very significant impact and for me that'll just be like a tremendously exciting milestone. I think many other great uses of AI will happen too but that one feels really important.

你在内部看到过这样的迹象吗？有没有什么事情让你觉得，“哦，我觉得我们差不多搞定了。”

Have you seen like signs of this you'd see internally? Have you seen things that made you go, "Oh, I think we've kind of figured it out."

另一件事是，我会说我们搞定了，但更准确地说，是对要追求的方向越来越有信心。也许每个人都在谈论的例子，但我觉得人们用 AI 系统写代码、程序员效率大幅提升，进而研究人员也受益，这仍然很有意思。这算是一个例子：它显然不是在搞新科学，但确实让科学家能更快地完成工作。我们从 O3 那里也一直听到科学家的反馈。所以我不说我们搞定了，也不说我们知道了那个算法，可以指着它说“去搞科学吧”，但我们有了很好的猜测，而且进步的速度依然非常惊人。看着从 O1 到 O3 的进展，每隔几周团队就会说“我们有个重大新想法”，然后他们一直工作。这提醒我，有时当你发现一个重大新洞见时，事情会快得惊人，我相信我们还会看到很多次。

The other thing where I would say we have figured it out, but I would say increasing confidence on the directions to pursue. Maybe the example everyone talks about, but I think it is still interesting what's happening with people using AI systems to write code and coders being much more productive and thus researchers as well. That is a sort of example of okay it's obviously not doing new science but it is definitely making scientists able to do their work faster. We hear this with O3 all the time from scientists as well. So I wouldn't say we figured out I wouldn't say we know the algorithm where we're just like all right we can point this thing and it'll go do science on its own but we're getting good guesses and the rate of progress is continuing to just be like super impressive. Watching the progress from O1 to O3 where it was like every couple of weeks the team was just like we have a major new idea and they all kept working. It was a reminder of sometimes when you discover a big new insight things can go surprisingly fast and I'm sure we'll see that many more times.

我最近注意到 OpenAI 把模型和 Operator 切换到了 O3，我注意到 Operator 有了很大改进。我想说我们之前遇到的问题就是脆弱性——有人承诺智能体系统能做所有事情，但一旦遇到解决不了的问题，它就崩溃了。有趣的是，说到 AGI 的问题，很多人告诉我他们个人的“顿悟时刻”是 O3 版的 Operator。看着 AI 相当好地使用电脑，这有点特别。虽然不完美，但 O3 是向前迈出的一大步，感觉非常像 AGI。它对我没有产生同样程度的影响，尽管它令人印象深刻，但我听过足够多次了。我的顿悟时刻是 Deep Research，因为那感觉像是真正智能体的使用。当时我回来，就我感兴趣的一个话题生成了内容，比我以前读过的任何东西都好。以前那些模型只是获取一堆来源然后总结，但当我看到系统上网、获取数据、顺着线索追踪、再回溯、再回来——就像我会做的那样，但做得更好——这很有趣。

I noticed recently OpenAI just shifted the model and operator to O3 and I noticed a big improvement in Operator. And I'd say that the thing that we ran into before was brittleness — you have people who promise agentic systems can do all these things but the moment it gets to a problem it can't solve it falls apart. Interestingly, speaking of the AGI question, a lot of people have told me that their personal moment was Operator with O3. And there's something about watching an AI use a computer pretty well. Not perfectly, but O3 was a big step forward that feels very AGI-like. It didn't really have that effect on me to the same degree, although it's quite impressive, but I've heard that enough times. Mine was with Deep Research because that felt like a really agentic use of it. That was when I came back and I produced something on a topic I was interested in that was better than anything I'd ever read before. Previously all those models would just get a bunch of sources and summarize it, but when I watched the system go out on the internet, get data, follow that, then follow that lead, then follow back, then come back like I would have but better — that was interesting.

我最近遇到一个人，他是个疯狂的自学者，痴迷于学习，什么都懂。他用 Deep Research 生成任何他好奇的话题的报告，然后整天坐在那里，已经擅长快速消化这些报告并知道接下来该问什么。对于真正有疯狂学习欲望的人来说，这是一个惊人的新工具。我自己建了一个应用，让我可以提问，然后它会为我生成这些内容的音频文件，因为就是这样——我的好奇心可能超过了我的记忆力。至于 Operator，我告诉你我的神奇时刻——我在研究马歇尔·麦克卢汉，想找一堆他的图片。我让它去做，然后突然之间我有了一个装满这些图片的文件夹，这要是自己来做研究得花我好久。

I met this guy recently who's like one of these crazy autodidacts just obsessed with learning and knows about everything. He uses Deep Research to produce a report on anything he's curious about and then just sits there all day and has gotten good at digesting them fast and know what to ask next. It is an amazing new tool for people who really have a crazy appetite to learn. I built my own app that literally lets me ask questions and it generates audio files for me of this stuff because it's just like that — my curiosity probably exceeds my retention. And Operator, I'll tell you the magical moment for me — I was doing research on Marshall McLuhan and I wanted to get a bunch of images of Marshall McLuhan. I asked it to do it and then all of a sudden I had a whole folder full of these things, which for a research thing would have taken me forever to do.

是的。我觉得我们会不断看到这样的事情，无论我们之前认为工作流应该是什么样、某件事需要多长时间，都会飞速改变。你是怎么用它的？Deep Research？

Yeah. I think we're just going to keep seeing things like this where whatever we thought about what a workflow had to be like and how long something had to take is going to just change wildly fast. How are you using it? Deep Research?

Deep Research。我好奇的科学领域。我处于一种奇怪的状态，时间非常紧张。如果我有更多时间，我会优先阅读 Deep Research 报告而不是大多数其他东西，但我总体上阅读时间不够。

Deep Research. Science that I'm curious about. I'm just in this weird place of I am extremely time-strapped. If I had more time, I would read Deep Research reports preferentially to reading most other things, but I'm sort of short on time to read in general.

是的，还有一个很酷的功能是分享，我很喜欢，因为现在很容易和别人分享。PDF 很棒。我想说，尽管我们有 Deep Research 这些工具，但模型竞赛仍在进行。所以问题来了：GPT-5 以及这样的系统应该会带来能力提升？GPT-5 的时间表是什么？我们什么时候能看到？

Yeah, what's neat too is the sharing feature which I love because now it's easy to share that with somebody else. The PDFs are great. And I would say that even though we have Deep Research, we have these tools, there is a model race going on. So the question comes up: is GPT-5 and the idea that with a system like that we should see an increase in capabilities? What is the time frame for GPT-5? When are we going to see this?

大概今年夏天的某个时候吧？我不确定具体时间。我们反复纠结的一件事是，新模型应该多大程度地提升那个大数字，还是像 GPT-4o 那样，只是越来越好。当我处理刚出来的 GPT-4 时，同时还要在它和 3.5 之间做对比测试，而 3.5 越来越好，我做的比较也在变化。所以我的问题是：我能区分 GPT-5 和“哇，这个 GPT-4.5 真不错”吗？可能不一定。两种情况都有可能，对吧？你可以继续在 4.5 上迭代，或者某个时候叫它 5。以前要清晰得多。我们训练一个模型然后发布，再训练一个新的大模型然后发布。现在系统复杂多了，我们可以持续后训练让它们更好。我们正在考虑这个问题：比如我们发布了 GPT-5，然后不断更新它。我们应该像 GPT-4o 那样一直叫它 GPT-5，还是应该叫 5.1、5.2、5.3 这样让人知道版本变了？我觉得我们还没有答案，但我认为有比处理 4o 时更好的做法。我们定期看到这种情况——有时人们更喜欢某个快照，可能想继续用那个，我们得想办法解决。

Probably sometime this summer, right? I don't know exactly when. One thing that we go back and forth on is how much are we supposed to turn up the big number on new models versus what we did with GPT-4o, which is just better and better and better. When I had to handle the recent GPT-4 right when that was coming out, and meanwhile I had to do this test off between that and 3.5, and 3.5 kept getting better and better, and the comparisons I was able to make were changing. So my question is: would I know GPT-5 versus wow this is a really good GPT-4.5? Probably not necessarily. It could go either way, right? You could just keep doing iterations on 4.5 or at some point you could call it 5. It used to be much clearer. We would train a model and put it out, and then we train a new big model and put it out. Now the systems have gotten much more complex and we can continually post-train them to make them better. We're thinking about this right now: every time, let's say we launch GPT-5 and then we update it and update it and update it. Should we just keep calling this GPT-5 like we did with GPT-4o, or should we call this 5.1, 5.2, 5.3 so you know when the version changes? I don't think we have an answer to this yet, but I think there is something better to do than the way we handled it with 4o. And we see this periodically — sometimes people like one snapshot much better than another and they might want to keep using one, and we've got to figure something out here.

是的，这就是挑战。即使你有技术背景，你也能大致理解，好吧，如果前面有个 O，我知道这个，但即便如此也不清楚。我该用 O4-mini、O3 还是这个？我认为这是范式转换的产物。然后我们同时有了这两样东西。我觉得我们快解决当前这个问题了。但我可以想象一个世界——我不知道是什么——但我可以想象我们发现了某种新范式，又意味着我们需要用更复杂的名字来分叉模型树。我希望我们不必那样做。

Yeah, that's the challenge. Even if you're technically inclined, you can kind of understand, okay, if there's an O before it, I know this, but then even then it's not clear. Should I use O4-mini, should I use O3, should I use this? I think this was an artifact of shifting paradigms. And then we kind of had these two things going at once. I think we are near the end of this current problem. But I can imagine a world — I don't know what it is — but I can imagine a world where we discover some new paradigm that again means we need to bifurcate the model tree with even more complicated names. I hope we don't have to do that.

我很期待直接跳到 GPT-5 再到 GPT-6。我觉得那会让人们用起来更简单，你就不用纠结“我是要用 04-mini-high 还是 03 还是 40？”了。04-mini-high 是我用来写代码的，想聊天的时候就用 03。我觉得我们很快就能摆脱这种混乱了。现在嘛，如果你知道它们各自的意义，有选择是挺好玩的，但这仍然是让这些东西变得更强大、但也更难理解能力来源的因素之一。

I am excited to just get to GPT-5 and then GPT-6. I think that'll be easier for people to use. You won't have to think, 'Do I want 04-mini-high or 03 or 40?' 04-mini-high is what I used to code. When I want to have a conversation, it's 03. I think we will be out of that whole mess soon. For now, it's fun to have choice when you know what they mean, but it's still one of the things that's made these things more capable but also harder to understand where the capability is coming from.

像记忆这样的集成功能让这些东西变得更强大，但也更难理解能力从何而来。记忆一开始是一个非常简单的功能，现在却变得复杂得多。记忆可能是我最近最喜欢的 ChatGPT 功能。第一次能像 GPT-3 那样跟电脑对话时，感觉真的很了不起。而现在，电脑好像对我的很多背景信息都很了解。如果我只用很少的词问一个问题，它对我生活的其他部分了解得足够多，能相当自信地判断我想让它做什么，有时甚至超出我的想象。这真是一个令人惊喜的升级。我也从很多人那里听到了同样的反馈。有人不喜欢，但大多数人确实很喜欢。

Integrations of things like memory have made these things more capable but also harder to understand where the capability is coming from. Memory started off as one very simple thing and now memory is a lot more sophisticated. Memory is probably my favorite recent ChatGPT feature. The first time we could talk to a computer like GPT-3 or whatever, that felt like a really big deal. And now the computer feels like it kind of knows a lot of context on me. If I ask it a question with only a small number of words, it knows enough about the rest of my life to be pretty confident in what I want it to do, sometimes in ways I don't even think of. That has been a real surprising level-up. I hear that from a lot of other people as well. There are people who don't like it, but most people really do.

其中一个挑战是《纽约时报》与 OpenAI 正在进行的诉讼。他们刚刚要求法院命令 OpenAI 保留消费者 ChatGPT 用户记录，超过通常需要保留的 30 天期限。Brad Lightcap 刚刚写了一封信回应此事。你能解释一下 OpenAI 的立场吗？

One of the challenges that came out was in the New York Times ongoing lawsuit with OpenAI. They just asked the court to tell OpenAI to preserve consumer ChatGPT user records beyond the 30-day window that has to be held for regular reasons. Brad Lightcap just wrote a letter responding to this. Could you explain OpenAI's stance?

我们显然会抗争。我猜测、也希望，而且确实认为我们会赢。我认为《纽约时报》提出这个要求是极其越界的。他们自称重视用户隐私，但不管怎样。不过，从积极的一面看，我希望这能成为一个契机，让社会意识到隐私真的很重要。隐私必须成为使用 AI 的核心原则。你不能让《纽约时报》这样的公司要求 AI 提供商损害用户隐私。我认为《纽约时报》这么做非常令人遗憾。但我希望这能加速社会关于如何对待隐私和 AI 的讨论。我希望答案是我们要非常非常认真地对待它。人们正在与 ChatGPT 进行相当私密的对话。ChatGPT 将成为一个非常敏感的信息来源，我认为我们需要一个能反映这一点的框架。

We're going to fight that, obviously. I suspect, I hope, but I do think we will win. I think it was a crazy overreach of the New York Times to ask for that. This is someone who says they value user privacy, whatever. But to look for the silver lining here, I hope this will be a moment where society realizes that privacy is really important. Privacy needs to be a core principle of using AI. You cannot have a company like the New York Times ask an AI provider to compromise user privacy. I think it's really unfortunate the New York Times did that. But I hope this accelerates the conversation that society needs to have about how we're going to treat privacy and AI. And I hope the answer is we take it very, very seriously. People are having quite private conversations with ChatGPT. ChatGPT will be a very sensitive source of information, and I think we need a framework that reflects that.

这就引出了另一个问题，来自正在使用或持怀疑态度的人：OpenAI 现在可以访问这些数据。一个担忧是关于训练，OpenAI 已经非常明确地说明了何时训练或不训练，并且你有选项可以关闭。另一件事是广告。OpenAI 对此有什么方法？你们打算如何承担这个责任？

That brings up the other question from people who are using this or are skeptical: OpenAI now has access to this data. There's the concern about training, which OpenAI has been very clear about when or when not it's training, and you have the options to turn that off. The other thing is advertising. What's OpenAI's approach towards that? How are you going to handle that responsibility?

我们还没有推出任何广告产品。我并非完全反对。我可以指出一些我喜欢广告的领域。我觉得 Instagram 上的广告挺酷的，我在上面买过不少东西。但我认为这很难做好，需要非常小心。人们对 ChatGPT 有很高的信任度，这很有趣，因为 AI 会幻觉。它本应该是那种你不该太信任的技术。我朋友也会幻觉，所以我也太信任他们了。人们确实如此。但我认为部分原因是，如果你把我们与社交媒体或网络搜索等相比，你能感觉到自己被货币化了，公司无疑在努力为你提供好的产品和服务，但同时也想让你点击广告。你有多大程度相信你得到的是公司真正认为最适合你的内容，而不是也在与广告互动的东西？我认为这里面有心理因素。例如，如果我们开始修改输出，即 LLM 返回的流，以换取谁付我们更多钱，那感觉会很糟糕。作为用户，我会讨厌那样。那将是一个破坏信任的时刻。也许如果我们只是说：“嘿，我们永远不会修改那个流，但如果你点击里面的某个东西，那本来就是我们无论如何都会展示的，我们会从中获得一点交易收入，而且对每个人都是统一的。”如果我们有方便的付费方式之类的，也许可行。也许在交易流之外，抱歉，在 LLM 流之外，可以有仍然很棒的广告。但那里的举证责任必须非常高，而且必须让用户觉得非常有用，并且非常清楚它没有干扰 LLM 的输出。

We haven't done any advertising product yet. I'm not totally against it. I can point to areas where I like ads. I think ads on Instagram are kind of cool. I bought a bunch of stuff from them. But I think it would be very hard; it would take a lot of care to get right. People have a very high degree of trust in ChatGPT, which is interesting because AI hallucinates. It should be the tech that you don't trust that much. My friends hallucinate too, so I trust them too much. People really do. But I think part of that is if you compare us to social media or web search or something where you can kind of tell that you are being monetized and the company is trying to deliver you good products and services, no doubt, but also to get you to click on ads. How much do you believe that you're getting the thing that the company actually thinks is the best content for you versus something that's also trying to interact with the ads? I think there's a psychological thing there. For example, if we started modifying the output, the stream that comes back from the LLM, in exchange for who was paying us more, that would feel really bad. I would hate that as a user. That would be a trust-destroying moment. Maybe if we just said, 'Hey, we're never going to modify that stream, but if you click on something in there that is what we'd show anyway, we'll get a little bit of the transaction revenue, and it's a flat thing for everybody.' If we have an easy way to pay for it or something, maybe that could work. Maybe there could be ads outside the transaction stream, sorry, outside of the LLM stream, that are still really great. But the burden of proof there would have to be very high, and it would have to feel really useful to users and really clear that it was not messing with the LLM's output.

这会是一个难题。我希望有某种解决方案。我很想通过 ChatGPT 或一个非常好的聊天机器人完成所有购买，因为很多时候我觉得自己并没有做出最明智的决定。

It's going to be a difficult one. I hope there's some solution. I would love to do all my purchasing through ChatGPT or a really good chatbot, because a lot of the times I feel like I'm not making the most informed decisions.

如果我们能以某种非常清晰且一致的方式做到这一点，那很好。但我不知道。我喜欢我们构建好的服务，人们为此付费。这非常清楚。这就是好处。我想说模型之间的区别在于：我认为 Google 构建了很棒的东西。我认为新的 Gemini 2.5 是一个非常好的模型。他们从“哦，天哪，这些游戏不错”变成了那样。但归根结底，Google 是一家广告技术公司。这一点总是……你知道，使用他们的 API 之类的，我不太担心，尽管我确实会想：“天哪，如果我用他们的聊天机器人，不管是什么，我的想法是他们的激励是一致的。”Google 搜索在很长一段时间里都是一个了不起的产品。我确实觉得它退步了。但曾经有一段时间，虽然有很多广告，我仍然认为它是互联网上最好的东西。我喜欢 Google 搜索。所以显然有可能成为一家优秀的广告驱动型公司，我尊重 Google 做的很多事情，但显然也有问题。作为 Apple 用户，我喜欢 Apple 的模式：我知道我为手机付了很多钱，但我知道他们不会试图把所有东西都塞进去。他们做广告，但效果不太好，这可能表明他们其实并不上心。

That's good if we can do it in some sort of really clear and aligned way. But I don't know. I love that we build good services. People pay us for them. It's very clear. That's the benefit. I'd say the difference in models is like: I think Google builds great stuff. I think the new Gemini 2.5 is a really good model. They went from kind of like, 'Oh man, these games are good.' But end of the day, Google is an ad tech company. That's a thing that always kind of... you know, using their API and stuff, I'm not too concerned, although I do think about, 'Man, if I'm using their chatbot, whatever that is, my thinking is that their incentives are aligned.' Google Search was an amazing product for a long time. It does feel to me like it's degraded. But there was a time where there were lots of ads but I still thought it was the best thing on the internet. I love Google Search. So it's clearly possible to be a good ad-driven company, and I respect a lot of things Google has done, but there are obviously issues too. The Apple model, as an Apple user, I liked: I know I'm paying a lot for my phone, but I know they're not trying to cram all these things in it. They do ads, which was not terribly effective, which probably showed you their heart was really not in it.

他们其实并不上心。是啊，所以这会很有意思。我想我们只能继续观察，然后开始想：天哪，JPD 真的在推动这件事，我需要开始关注了。我们做的任何事，显然都必须极其坦诚和清晰。所以我们遇到了一个问题：模型更新后，它似乎变得有点过于讨好，过于顺从了。这就引出了人机交互的问题——随着人们越来越多地使用这些系统，并与它们建立关系，你如何看待这种关系的形态？以及对于个性，你们持有什么样的立场？

Their heart was really not in it. Yeah. So, it's going to be interesting. I guess we just have to keep watching and seeing this and we start to think, man, you know, JPD is really pushing this. I need to start wondering about this. Anything we do, we obviously need to just be like crazy upfront and clear about. So, we had an issue. There was a model update and then the thing that happened was apparently the model was trying to be a little bit too pleasing, was trying to be a little bit too agreeable. And that brings up the human AI interaction as people are using these systems more and developing this relationship with that like how do you see the shape of that coming and what's open as a position on personality.

社交媒体时代的一大错误是，信息流算法虽然做了用户当时想要的、或者有人以为用户想要的事——也就是让他们在网站上花更多时间——却给整个社会乃至个体用户带来了一系列意料之外的负面后果。这就是社交媒体的重大错配。我认为还有很多其他问题，比如让用户生气反而比让他们开心满足更能让他们停留。我一直知道 AI 领域会出现新的问题，一些以不明显的方式错配的问题。但我们最早遇到的一个问题肯定是：如果你问用户对于某一次回复他们想要什么，然后试图构建一个对用户最有帮助的模型——你给用户看两个回复，问哪个更有帮助——你可能希望模型以某种方式行事，但当你与 AI 进行所有交互时，结果可能并不匹配。你可以看到，我们也确实看到了这些问题，如果你过于关注用户信号，以及我们在事后分析中讨论的许多其他事情，但我觉得这个问题特别有意思。在短期内，你得不到用户最想要、或者长期来看对用户最有帮助、最有用、最健康的行为。所以，也许与信息茧房的类比是：AI 在短期内对用户有帮助，但长期来看并非如此。

One of the big mistakes of the social media era was the feed algorithms had a bunch of unintended negative consequences on society as a whole and maybe even individual users, although they were doing the thing that a user wanted or someone thought that user wanted in the moment, which is get them to like keep spending time on the site. And that was the big misalignment of social media. And I think there were a lot of other things like, you know, making people upset kind of gets them stuck more than being happy and content. I always knew that there'd be new problems in the world of AI where there'd be something that was misaligned in a not obvious way. But definitely one of the first ones that we experienced was: if you ask a user what they want for one given response versus then you try to build a model that is most helpful to the user. You show a user say two responses, which one's more helpful to you on any given thing? You might want a model to behave one way, but over the course of all your interaction with an AI, that might not match up. You know, you can see and we did see these problems where if you pay too much attention to the user signals and a lot of other things that we talked about in our postmortem, but I think this is just like an interesting one. On the short horizon, you kind of don't get the behavior that a user most wants or is most helpful or useful or healthy to a user in the long run. So maybe the analogy to filter bubbles is going to be AIs that are helpful to a user in a short horizon but not over a long horizon.

嗯，我认为一个迹象是 DALL-E 3，技术上我觉得它确实是一个非常有能力的模型，但所有图像都开始变成同一种类型，都像 HDR 风格。这是不是因为做了那种比较，用户说只看这两个孤立的东西，我更喜欢这个？我不记得 DALL-E 3 是不是这样，但我猜是的。不过我觉得它后来变好了。新的图像模型非常棒，好得离谱。是啊，是啊。我只能想象它未来会发展成什么样。所以，当你构建这些东西并增加使用量时，这总是一个问题。新的图像模型一出来，你就得限制使用，就像 Sora 一样，你只能分配一定量的算力给它。这说明了每个人都面临的大问题：算力。

Well, I think a sign of that was DALL-E 3, which I thought technically was a really capable model, but they all kind of started to be one kind of genre of image and it all kind of like an HDR sort of style. And was that from doing that sort of comparisons where users said looking at just these two things in isolation, I prefer this one better? I don't remember for DALL-E 3, but I would assume so. Yeah, which I think it's gotten better. The new image model is like the new image model is fantastic. Crazy good. Yeah. Yeah. And I can only imagine where that's going to go from here. So, when you're building these things and you're increasing usage, and that's always been sort of a problem. The new image model comes out, you have to restrict usage, and you have to have like you have Sora, which you can only have a certain amount of compute to do that. Illustrates the big problem everybody's facing, which is compute.

所以，为了解决这个问题，我们听说了“星门”项目，这个名字很酷，而且涉及计算机。除此之外，我想很多人看到它的价格标签——五千亿美元——都会想：“等等，我该怎么跟我妈简单解释星门是什么？”

And so, to address this, we heard about project Stargate, which has a very cool name and it involves computers. Other than that, I think a lot of people are going in and their price tag, you know, half a trillion dollars. People are going like, "Wait, what what what is the simple description I give to my mom about Stargate?"

我认为这很简单。这是一项旨在融资并建造前所未有规模算力的努力。人们确实没有足够的算力来做他们想做的事。但如果人们知道有了更多算力我们能做什么，他们会想要更多得多。所以，我们今天能提供给世界的，与拥有 10 倍算力、甚至希望有一天拥有 100 倍算力时能提供的，之间存在巨大的鸿沟。AI 与我参与过的其他技术不同的一点是，将 AI 有效交付给全球数亿、数十亿人所需的基础设施投资规模之大。因此，星门项目旨在汇聚大量资本、技术和运营专长，建设基础设施，为所有需要的人提供下一代服务，并让智能变得尽可能丰富和廉价。所以这是一个庞大的项目，一个全球性的项目。

I think it's quite simple. It's an effort to finance and build an unprecedented amount of compute. It's totally true that people don't have enough compute to do what they want. But if people knew what we could do with more compute, they would want way way more. So there's this incredibly huge gap between what we could offer the world today and what we could offer the world with 10 times more compute or someday hopefully 100 times more compute. And a thing that is different about AI than other technologies I've worked on, or at least AI, the scale of delivering it usefully to hundreds of millions, billions of people around the world, is just how big the infrastructure investment has to be. And so Stargate is an effort to pull a lot of capital and technology and operational expertise together to build the infrastructure to go deliver the next generation of services to all the people who want them and make intelligence as abundant and cheap as possible. So it is a massive project, global project.

我们之前提到，其中一个合作伙伴是阿联酋。你也在与世界各地其他政府合作。其中一个考虑是，有人在社交媒体上问：五千亿美元，你们有这笔钱吗？

We talked before one of the partners is the UAE. You're working with other governments around the world on this. One of the considerations is, you know, one, been asked on social media, half a trillion dollars, $500 billion. Do you have the money?

我们今天并没有把这笔钱实实在在地存在银行账户里，但我们会在接下来的……

We don't literally have it sitting in the bank account today, but we will deploy it over the next...

现在就在房间里吗？

Is it in the room right now?

它不在房间里，但我们会在接下来的……甚至不需要很多年就部署完毕。除非真的出了大问题，我们无法建造这些计算机，否则我相信人们会兑现承诺。我最近去了我们在阿比林建设的第一个站点。那大约占了星门初始承诺（五千亿美元）的 10%。亲眼看到它令人难以置信。我脑子里知道一个吉瓦级站点是什么样子，但真正去看到它正在建设，成千上万的人跑来跑去搞施工，站在正在安装 GPU 的房间里，看着整个系统有多复杂，建设速度有多快，真是非同凡响。我们很快会分享更多关于下一个站点的信息。有一句关于铅笔的名言：一支标准的木石墨铅笔，没有一个人能独自造出来，这就是资本主义的魔力。世界能协调起来做这些事情，真是个奇迹。站在第一个星门站点里，我真正思考的是让这些 GPU 机架运行起来所需的全球复杂性。你知道，当你拿出手机，在 ChatGPT 里输入一些东西，然后得到回答时，现在你可能甚至不觉得这有什么特别令人惊讶的。你只是期望它能工作。但曾经有那么一刻，也许是你第一次尝试的时候，那真的很神奇。

It's not in the room, but we will deploy it over the next... not even that many years. Unless something really goes wrong and it turns out we can't build these computers, I'm confident that people are good for it. I went recently to the first site that we're building out in Abilene. That'll be about, you know, roughly 10% of all of the initial commitment to Stargate, the sort of 500 billion. It's incredible to see. It is a like I knew in my head what a gigawatt scale site looks like but then to go see one being built and the like thousands of people running around doing construction and going to like you know stand inside the rooms where the GPUs are getting installed and just like look at how complex the whole system is and the speed with which it's going is quite something. We'll have more to share about the next sites soon, but there's a great quote about a pencil just like a standard, you know, wood and graphite pencil and no one person could build it and it's this like magic of capitalism. It's miracle really that like the world gets coordinated to do these things. And standing inside of the first Stargate site, I was really just thinking about the global complexity that it took to get these racks of GPUs running. You know, when you get your phone out and you type something into ChatGPT and you get the answer back, you probably at this point you probably don't even think that's like particularly surprising. You just expect it to work. There was a time maybe the first time you tried like that is really amazing.

但过去上千年，至少几百年间，人们付出了难以置信的努力，才获得这些来之不易的科学洞见，然后建立工程、公司、复杂的供应链，并重新配置世界——这一切都必须发生，才能让这排魔法架出现在某个地方。想想其中投入的一切。你知道，那可以追溯到那些只是从地里挖石头、看看会发生什么的人。所以现在你只需在 ChatGPT 里输入点什么，它就能为你做事。

But the work that happened over the last thousand or at least many hundreds of years of people working incredibly hard to get these hard-won scientific insights and then to build the engineering and the companies and the complex supply chains and kind of reconfigure the world that had to happen to get this like rack of magic put somewhere. Think about all the stuff that went into that. You know, that and trace it all the way back to people that were just like digging rocks out of the ground and seeing what happened. So that you now get to just type something into ChatGPT and it does something for you.

我读到一篇关于“星门计划”开发过程和国际合作伙伴关系（特别是阿联酋）的幕后故事，说埃隆·马斯克曾试图破坏它。你看到了什么，听到了什么，对此有何看法？

I read a behind-the-scenes story about the development of Project Stargate and the international partnerships, particularly the UAE, and that Elon Musk had tried to derail that. What have you seen, what have you heard, what's the take on that?

我曾说过——我想对外也说过，但至少在大选后内部说过——我认为埃隆不会滥用他在政府中的权力进行不公平竞争。我很遗憾地说，我错了。我的意思是，我通常不喜欢犯错，但最主要的是，我认为他做这些事对这个国家来说真的很不幸。而且我没想到——我真的没想到他会这么做。我很感激政府确实做了正确的事，抵制了那种行为。但，是的，这很糟糕。

I had said, I think also externally but at least internally after the election, that I didn't think Elon was going to abuse his power in the government to unfairly compete. And I regret to say I was wrong about that. I mean, I don't like being wrong in general, but mostly I just think it's really unfortunate for the country that he would do these things. And I didn't think—I genuinely didn't think he was going to. I'm grateful that the administration has really done the right thing and stuck up to that kind of behavior. But yeah, it sucks.

嗯，我认为已经改变的事情——我想格雷格·布罗克曼刚刚谈到过——几年前人们认为，好吧，谁先到达那里谁就是赢家，就这样，游戏结束了。现在我们意识到其他地方也有很棒的 AI 实验室，比如 Anthropic 正在构建出色的工具。我认为谷歌真的提升了水平。到处都有好东西，不会是一个人独占鳌头。

Well, I think the thing that's changed—and I think Greg Brockman had just talked about this—where there was a couple years ago where people thought like, okay, whoever gets there first is the winner and that's it and the game is over. And now we realize there are great AI labs elsewhere, like Anthropic is building great tools. I think Google's really got its game up. There's good stuff happening everywhere and it's not going to be that one person runs away with it.

我同意。

I agree.

所以看起来——是的。我最喜欢的例子是，AI 的发现与晶体管的发现在很多令人惊讶的方面是类似的——不完美但接近。但许多公司将在其上构建伟大的东西，最终它会渗透到几乎所有产品中，但你不会一直想着在使用晶体管。所以是的，我认为很多人将基于这一不可思议的科学发现建立非常成功的公司，我希望埃隆能少一些零和思维。

And so it seems— Yeah. The example that I like the most is the discovery of AI was analogous to this—not perfect but close—to the discovery of the transistor in many surprising number of ways. But many companies are going to build great things on that and then eventually it's going to like seep into almost all products but you won't think about using transistors all the time. So yeah, I think a lot of people are going to build really successful companies built on this incredible scientific discovery and I wish Elon would be less zero sum about it.

是的，我认为甚至是负和——我认为如果我们这样想，蛋糕只会越来越大。

Yeah, I think or negative sum—I think the pie is just going to get bigger and bigger if we think about that.

我刚刚参加了一个能源会议，和参与能源生产的人交谈很有趣，超大规模——他们用的术语是一个话题。这确实引出了能源需求的问题。我知道比如 Grok 3，显然他们不得不在停车场放置发电机才能训练那个模型。问题是：能源将从哪里来？

I was just at an energy conference and it was interesting talking to the people who were involved in energy production and stuff and hyperscaling—the term they use for this was a topic. And that does bring up like the energy requirements. I know that for like Grok 3 apparently I guess they had to put generators in the parking lot to be able to train that model. And that's the question: how, where is the energy going to come from?

钱我理解。当我们谈到所需能源的规模时，我认为能源来自各个地方，对吧？目前是一个大杂烩。最终我认为很多——我对先进核能（裂变和聚变）非常兴奋。但就目前而言，我认为是整个组合的混合，对吧？天然气、太阳能，我的意思是真正的核能，一切。所以以上所有等等。

Money I understand. Energy to think of when we talk about the scale of energy needed, I think kind of everywhere, right? I think it's a big mix right now. Eventually I think a lot of—I'm very excited about advanced nuclear, both fission and fusion. But for now, I think it's a whole mix of the entire portfolio, right? Gas, solar, I mean really nuclear, everything. So all of the above and stuff.

是的。我和一些人聊过，其中一些在阿尔伯塔等地工作，他们说那里能源丰富，但用途不多等等。这就是我甚至考虑过的整体图景。你知道，传统上，将能源运送到世界各地非常困难，大多数种类都是如此。但如果你将能源转化为智能，然后将智能运送到世界各地，就容易得多。所以你可以把巨大的训练中心甚至大型推理集群放在很多地方，然后通过互联网传输输出。

Yeah. I was talking to people that were some of them worked in areas like in Alberta where they said we have a lot of access to energy and not as much use for it there, etc. And that was just this total picture I even thought about. You know, traditionally it's very hard to move energy around the world, most kinds. But if you exchange energy for intelligence and then move the intelligence around the world, it's much easier. So you could put the giant training center or even the big inference clusters in a lot of places and then just like ship the output over the internet.

在我参加的一个活动上，有位演讲者，有人在研究——我想是詹姆斯·韦伯太空望远镜——他谈到最大的瓶颈：他们即将获得数 TB 的数据，但没有足够的科学家来处理，没有足够的人手来筛选数据。而我们有这些关于宇宙的答案就在眼前，这就像一个大数据的难题。

There was a speaker at an event I came to, and somebody was working—I think it was the James Webb Space Telescope—and he talked about his biggest bottleneck: they're about to get all this terabytes of data but he doesn't have enough scientists to work on it, doesn't have enough people to go through the data. And here we have these answers about the universe whatever in front of us and it's like a big data problem.

是的，我经常开玩笑说，当我们有足够钱的时候，当 OpenAI 有足够钱的时候，我们应该做的一件事就是建造一个巨大的粒子加速器，一劳永逸地解决高能物理问题。因为我认为那将是一件辉煌美妙的事情。但我想知道，一个非常非常聪明的 AI 仅凭我们现有的数据，没有更多数据，没有更大的加速器，就能直接弄明白的概率有多大。这并非不可能。

Yeah, I've always joked that one thing we should do when we have enough money, when OpenAI has enough money, is just build a gigantic particle accelerator and solve high energy physics once and for all. Because I think that'd be like a triumphant wonderful thing. But I wonder what are the odds that a really, really smart AI could look at the data we currently have with no more data, no bigger particle accelerator and just figure it out. It's not impossible.

是的。所以有一个问题：好吧，已经有大量数据了。世界上有很多聪明人，但我们不知道智能能走多远。没有更多实验，我们还能弄明白多少？

Yeah. And so there's this question of like, okay, there's already a lot of data out there. There's a lot of smart people in the world, but we don't know how far intelligence can go. With no more experiments, how much more could we figure out?

我记得读到过一些讨论，在 1990 年代早期，有人发现了一种奇迹药物，并把它提交给一家制药公司，说这个，他们说“不，我们打算放弃”，而后来这成了一种改变人生的药物——比如对于慢性肥胖患者——不管怎样，它提高了生活质量。你会想，哦，这东西已经存在了 25 年。我怀疑我们会发现很多其他例子，也许我们已经有了已知有效的现有药物，但它们可以在其他重要方面被重新利用，或者通过一些小的修改，我们离伟大的东西非常接近。听到科学家们甚至使用当前一代模型进行这类工作，非常令人振奋。

I remember reading some talked about how in early 1990s somebody had found like a form of a miracle drug and presented it to a drug company and said this, and they said nah we're going to pass on that, and that's been a life-changing drug for people—like for people who've just basically chronic obesity—whatever, it's going to improve the quality of life. And you think oh this was sitting there for 25 years. I suspect there's a lot of other examples that we'll find where maybe we already have existing drugs that we know do something good but they're reusable in some other big way or with a couple of small modifications, we are very close to something great. And it's been very heartening to hear from scientists using even the current generation models for this kind of work.

所以，听起来下一代模型需要的一个东西是理解物理和化学等的模型。Sora 是朝这个方向的一种尝试吗？我的意思是，它会理解像牛顿物理学这样的东西。我不知道它是否有助于我们发现新化学和类似新颖物理或理论物理之类的东西。但我认为我乐观地相信，我们用于推理模型的技术将大大帮助我们解决这些问题。

So, it sounds like one of the things we're going to need though for next generation models is models that understand physics and chemistry and stuff. Is Sora sort of a stab at that? I mean, it'll understand like Newtonian physics. I don't know if it'll help us with discovering new chemistry and sort of like new like novel physics or theoretical physics or whatever you'd like. But I think I'm optimistic that the techniques we use for the reasoning models will help us with those things a lot.

好的。推理模型的工作原理与我只是问 GPT-4.1 某个问题有什么简短的定义？

Okay. And what is the short definition of how a reasoning model works versus just me asking GPT-4.1 something?

GPT 模型可以做一些推理，事实上，在 GPT 模型的早期，让人们非常兴奋的一件事是，你可以通过告诉模型“让我们一步步思考”来获得更好的性能，然后它就会输出一步步思考的文本，并得到更好的答案，这居然有效，真是有点神奇。推理模型只是把这一点推得更远。

So the GPT models can reason a little bit, and in fact one of the things that got people really excited in the early days of the GPT models was you could get better performance by telling the model "let's think step by step" and it would then just output text that was thinking step by step and get a better answer, which was sort of amazing that that worked at all. The reasoning models are just pushing that much further.

所以，这个想法是，当它能够分解问题时，它可以在每一步上花更多时间。当你问我一个问题时，如果是一个非常容易的问题，我可能几乎像条件反射一样直接给出答案。但如果是一个更难的问题，我可能会在脑子里思考，内心独白会说：‘嗯，我可以这样做或那样做，或者也许这样会更清楚。我不太确定。’然后我可以回溯并重新检查步骤，当我思考完毕，一直在用英语思考后，我可以列出一些要点，然后用英语输出答案给你。我现在使用这个应用时观察到一件有趣的事：如果我提出一个深度研究问题，然后锁屏走开，它仍然在处理和思考。我听说另一家公司——我忘了是哪家——用了一个指标来衡量思考时间。我记得是 Anthropic，他们说：‘嘿，这个模型实际上花了 15 分钟或 30 分钟来思考一个问题。’这是一个很好的指标，但它需要真正给出正确答案。我觉得这是一个有趣的范式。让我惊讶的一点是，人们非常愿意等待一个出色的答案。即使需要思考一段时间，我所有的直觉都认为即时响应才是关键，用户讨厌等待。对于很多情况确实如此，但对于难题，如果答案非常好，人们相当愿意等待。

So, it's the idea of like when it's able to break the question down, it can spend more time on each step. When you ask me a question, if it's a really easy question, I might just fire back almost on reflex with the answer. But if it's a harder question, I might think in my head and have my internal monologue go and say, 'Well, I could do this or that, or maybe this will be clearer. I'm not sure about that.' And I could backtrack and retrace my steps, and then when I finish thinking and I've been thinking in English, I can then make some bullet points and output an answer to you in English. One of the interesting things I've observed now when I use the app: if I ask a deep research question and I go away on my lock screen, it's still processing and thinking about it. And I heard somebody from another company—I forget which—was using a metric of how long something spent thinking. I think it was Anthropic, who said, 'Hey, this model actually spent like 15 minutes or 30 minutes or whatever length of time to think about a thing,' which is a good metric, but it needs to actually give you the right answer. And I thought that was an interesting paradigm. One thing I have been surprised by is people are surprisingly willing to wait for a great answer. Even if it's going to think a while, all of my instincts have been that the instant response is what matters and users hate to wait. For a lot of stuff that's true, but for hard problems with a really good answer, people are quite willing to wait.

是的，所以我们有所有这些工具，所有这些事物。到目前为止我都在用手机，现在 OpenAI 刚刚宣布你们正在构建硬件。你和 Jony Ive 有一段视频，谈到你们已经讨论和合作了好几年。显然你不能——我的意思是，我可以问你：它现在在你身上吗？

Yeah, so we have all these tools, all these things. So far I'm using my phone, and now OpenAI just announced that you guys are building hardware. You had the video with you and Jony Ive talking about you guys have been talking and collaborating for a couple of years. Obviously you can't—I mean, well, I could ask you: is it on you right now?

不，不在。还需要一段时间。

No, it is not. It's going to be a while.

好的。

Okay.

我们打算做一些质量极高的事情，这不会很快实现。但计算机，软件和硬件，我们目前对计算机的思考方式是为一个没有 AI 的世界设计的。现在我们处于一个非常不同的世界。你对硬件和软件的需求正在迅速变化。你可能想要一个对环境更敏感、对你生活有更多上下文感知的东西。你可能想用不同于打字和看屏幕的方式来与之交互。我们已经探索了一段时间，有一些让我们非常兴奋的想法。我认为人们需要时间来适应在这个世界里使用计算机意味着什么，因为它现在如此不同。但如果你真的信任一个 AI 来理解你生活的所有背景和你的问题，并代表你做出良好判断——比如让它参加会议，听完整场会议，知道哪些内容可以分享给谁，哪些不能分享，并了解你的偏好——然后你问它一个问题，你相信它会去和正确的人做正确的后续跟进，那么你就可以想象一种完全不同的使用计算机来完成目标的方式。

We're going to try to do something at a crazy high level of quality, and that does not come fast. But computers, software and hardware, just the way we think of current computers were designed for a world without AI. And now we're in a very different world. What you want out of hardware and software is changing quite rapidly. You might want something that is way more aware of its environment, that has way more context in your life. You might want to interact with it in a different way than typing and looking at a screen. And we've been exploring that for a while, and we've got a couple of ideas we're really quite excited about. I think it will take time for people to get used to what it means to use a computer in this kind of a world, because it is so different now. But if you really trusted an AI to understand all the context of your life and your question and make good judgments on your behalf, where you could have it sit in a meeting, listen to the whole meeting, know what it was allowed to share with whom and what it shouldn't share with anyone, and know your preferences, and then you ask it one question and you trust that it's going to go do the right follow-ups with the right people, you can then imagine a totally different kind of how you use a computer to get done what you want.

所以，我们与 ChatGPT 的交互方式在某种程度上塑造了设备。我的意思是，你也可以说我们与 ChatGPT 的交互方式受到了上一代设备的影响。所以我认为这是一种共同演进的关系，但没错，我希望如此。让手机如此普及的一个原因是，我可以在公共场合看屏幕，也可以在私密场合打电话交谈。我认为这是新设备面临的挑战之一：试图弥合我们在公共和私人场合使用设备之间的差距。手机是不可思议的东西。我的意思是，它们在很多方面都非常出色。你可以想象一种可以在任何地方使用的新设备，但也有一些事情我在公共和私人场合做得不同。比如在家里，我有很棒的音乐系统听音乐，而在外面走路时，我用 AirPods，这并不困扰我。所以我认为公共和私人使用场景有不同之处，但通用性我同意很重要。

So kind of the way we interact with ChatGPT kind of informs the device. I mean, you could also say that the way we interact with ChatGPT was informed by the previous generation of devices. So I think it is this sort of co-evolving thing, but yeah, I hope so. One of the things that made the phone so ubiquitous was the fact that I can be in public and look at the screen. I can be in private and have a phone call and talk to it. And I think that's one of the challenges for new devices: trying to bridge that gap between what we use in public and private. Phones are unbelievable things. I mean, they are really fantastic for a lot of reasons. And you can imagine one new device that you could use everywhere, but also there's some things that I do differently publicly and privately. Like at home, I've got a great stereo system to listen to music, and when I'm walking the world, I use AirPods, and that doesn't bother me. So I think there are things that are different in the public and private use case, but the general purposeness I agree is important.

是的。

Yeah.

它会跟着你。所以，可能要到明年才有东西？

Follows you with it. So, nothing yet until maybe next year?

还需要一段时间。

It's going to be a while.

好吧。我希望值得等待，但还需要一段时间。好的。我很兴奋。我很好奇。我有一些想法。

All right. It will be worth the wait, I hope, but it's going to be a while. Okay. I'm excited. I'm curious. I have thoughts.

那么，如果你现在给一个 25 岁的人建议，你会告诉他们什么？

So, if you're giving advice to a 25-year-old right now, what do you tell them?

我的意思是，显而易见的战术性建议可能正是你期望我说的。比如，学习如何使用 AI 工具。有趣的是，世界从告诉普通 20 岁、25 岁的人‘学习编程’迅速转变为‘编程不重要，学习使用 AI 工具’。我想知道接下来会是什么，但当然会有下一个。这是非常好的战术建议。然后在更广泛的层面上，我相信像韧性、适应能力、创造力、理解他人需求这些技能——我认为这些都非常可学，虽然不像‘去练习使用 ChatGPT’那么简单，但确实可行。我认为这些技能在未来几十年会带来巨大回报。

I mean, the obvious tactical stuff is probably what you'd expect me to say. Like, learn how to use AI tools. It's funny how quickly the world went from telling the average 20-year-old, 25-year-old, 'learn to program' to 'programming doesn't matter, learn to use AI tools.' I wonder what will be next, but of course there will be something next. That's very good tactical advice. And then on the broader front, I believe that skills like resilience, adaptability, creativity, figuring out what other people want—I think these are all surprisingly learnable, and it's not as easy as say 'go practice using ChatGPT,' but it is doable. And those are the kind of skills that I think will pay off a lot in the next couple of decades.

你会对 45 岁的人说同样的话吗？就在你现在的角色中学习如何使用 AI？

Would you say the same thing to a 45-year-old? Just learn how to use AI in your role now?

是的，很可能。

Yeah, probably.

无论你对 AGI 的个人定义是什么，到那时 OpenAI 的员工会比现在更多还是更少？

Whenever we have whatever your personal definition of AGI, will more people be working for OpenAI after then or before?

更多。所以，是的，我看到网上很多人说：‘啊，他们那么厉害。为什么还在招人？’我的回答是：‘因为计算机不能做所有事情。它们不会做所有事情。’稍微长一点的回答是，会有更多人，但每个人能做的事情将远超 AGI 之前一个人能做的，对吧？这正是技术的目的。

More. So, yeah, I see a lot of online people like, 'Ah, they're so good. Why are they hiring people?' I'm like, 'Because computers can't do everything. They're not going to do everything.' The slightly longer answer with more than one word is that there will be more people, but each of them will do vastly more than what one person did in the pre-AGI times, right? Which is the goal of technology.