OpenAI 的 ChatGPT 负责人：我们（又一次）进入了 AI 的新时代

Tibo SottiauxOpenAI

我经常看到外面大家在做的东西，就觉得：你们还没真正想明白。如果你逼自己一把，真正去想象一年后这一切都会比今天好上 10 倍左右，你做产品的方式就会完全不一样。

Often, when I look at what people are building out there, it's just like you're not quite getting it. If you were just pushing yourself and just really imagining all of this being roughly 10 times better than it is today in a year, you would build in a different way.

Tibo SottiauxOpenAI

互联网上大部分操作将由智能体来完成。如果你希望自己的产品在智能体面前也能成功，就必须按照某种量级的规模去构建。

The majority of actions on the internet will be taken by agents. If you want your product to be successful for agents, you have to build for a certain level of scale.

Tibo SottiauxOpenAI

随着我不断往前沿推进，我发现自己搭建的智能体团队越来越大；然后等模型迎来下一次突破，突然就变成：「哦，其实一个更强的智能体就能把这些全包了。」于是我又把团队缩小，然后就这样不断经历扩张、收缩、再扩张。

As I'm pushing the frontier, I find myself building larger and larger teams of agents, and then, when we have the next breakthrough with models, suddenly, it's just kind of like, "Oh. Well, a bigger agent can just do all of it." And so I kind of shrink the team again, and then it just goes through this expansion and shrinking expansion.

Tibo SottiauxOpenAI

要自己去搭建、去折腾你的循环，也许让一些人很兴奋，但我不认为未来会是这样运作的。随着时间推移，你想要的只是一个会学习的系统。你不一定想去琢磨：「哦，我刚在想，我得把循环精确地搭成这样才能出结果。」

Having to set up and fiddle with your loops is something that maybe people get excited about, but I don't think this is the way that it's going to work. Over time, you just want a system that learns. You don't want to necessarily think about, "Oh. I was just thinking I've got to loop it exactly this way in order to get results."

Tibo SottiauxOpenAI

它还会继续发生相当剧烈的变化。哪怕是今天，用起来还是有点笨拙，而且我觉得它会一直笨拙下去，直到某一天突然不再笨拙。

It's going to continue to change quite radically. Even today, it still feels a bit clunky, and I think it will feel clunky until it isn't.

Tibo SottiauxOpenAI

嗯，谢谢你邀请我。

Yeah. Thanks for having me.

Tibo SottiauxOpenAI

这房间不错。

Nice room.

Tibo SottiauxOpenAI

我挺好的。我很幸运有一支非常棒的团队。所以我睡觉的时候，他们几乎通宵都醒着。我们办公室有个很棒的房间叫「图书馆」，我们把它整个改成了一个超大的作战室，里面的氛围简直无可挑剔。大家在那儿一直待到很晚很晚。

I'm doing great. I'm fortunate to have a very good team. So while I sleep, they were awake most of the night. We have this amazing room at the office called the library, which we completely repurpose as a mega war room, and the vibes in there are just immaculate. People were there until very, very late.

Tibo SottiauxOpenAI

什么是写代码？

What is coding?

Tibo SottiauxOpenAI

有的有的，时不时会合并一些代码。严格来说，有大量代码是为我而写的，用来做各种分析、了解趋势、了解业务、思考下一个功能，以及看我们之前的发布效果怎么样。这些代码很多都是 Codex 写的。我当然已经不再手写了。偶尔周末的时候，我会给自己来一段疗愈时光，手动刷几道 LeetCode，但这也是我唯一真正手写代码的时候。

Yeah, yeah. Merging some code every once in a while. Technically, a lot of code is getting written for me to do all sorts of kinds of analysis, understand trends, understand the business, understand the next feature, how well our previous launches are doing. A lot of code is written by Codex for that. I'm obviously not writing it by hand anymore. Occasionally, during the weekends, I will have a little therapeutic moment where I do a little lead code just by hand, but that's the only actual hand coding that I do.

Tibo SottiauxOpenAI

我以前并行跑的要多得多，后来我们很幸运地在 Ultrafast 上取得了突破。所以现在我觉得自己又能重新进入心流状态了。光是有一个更快的智能体，就对我帮助很大。我很兴奋，我们很快就会把这个已经做了半年的 Ultrafast 推出去。另外，嗯，我觉得这个数量其实时多时少。

I used to run a lot more in parallel, and then we were fortunate enough to get a breakthrough at Ultrafast. And so now, I feel like I'm able to be in the flow again. And so just having a faster agent helps me a lot. I'm excited that we're getting six-months-old Ultrafast out there soon. Then, yeah, I think it sort of varies.

Tibo SottiauxOpenAI

随着我不断往前沿推进，我发现自己搭建的智能体团队越来越大。然后等模型迎来下一次突破，突然就变成：「哦，其实一个更强的智能体就能把这些全包了，还能把所有东西记在记忆里，还能学习。」于是我又把团队缩小，然后就这样不断经历扩张、收缩、再扩张。

As I'm pushing the frontier, I find myself building larger and larger teams of agents. Then, when we have the next breakthrough with models, suddenly, it's just like, "Oh. Well, a bigger agent can just do all of it, and keep everything in memory, and learn." And so I kind of shrink the team again, and then it just goes through this expansion and shrinking expansion.

Tibo SottiauxOpenAI

我觉得，要自己去搭建、去折腾你的循环、把它摸索清楚，也许曾让一些人很兴奋，但我不认为未来会是这样运作的。我认为未来的运作方式是……就是我们在 Dots 上的定位和正在发布的东西：你拥有一个极其聪明的智能体，全天候不间断地工作，理解你的目标，理解你的偏好，能从反馈中学习。我们这次发布的东西并不完美，把它开放给 Pro 用户之后，我会从中学到很多。但长远来看，你想要的只是一个能从你想达成的目标中学习的系统。你不一定想去琢磨：「哦，我刚在想，我得把循环精确地搭成这样才能出结果。」

I think having to set up and fiddle with your loops and figuring that out is something that maybe people got excited about, but I don't think this is the way that it's going to work. I think the way that it's going to work is that ... how we're positioning and what we're shipping with dots, where you have an incredibly smart agent that works 24/7, understands your goals, understands your preferences, learns from feedback. What we launched is not perfect. I will learn very much from making this available to our pro users. But over time, you just want the system that learns from what you want to achieve. You don't want to necessarily think about, "Oh. I was just thinking I've going to loop it exactly this way in order to get results."

Tibo SottiauxOpenAI

对。我觉得从根本上说，如果你退一步看，不管是不是 Dots，关键都在于摆脱技术本身的束缚，拥有一种常驻的、主动的智能：它知道自己需要做的一切，并且可以通过任何客户端、任何屏幕找到它。你可以呼叫它。比如你走进一间会议室，它就出现在会议里，帮你做笔记。之后你在邮件里接着跟它聊，需要的时候再给它发条消息。

Yeah. I think fundamentally, if you take a step back, and whether it's dots or not, it's all about breaking free from the technology and having this sort of permanent active intelligence that knows everything it needs to do and is available through any client, any screen. You can call it. Maybe you walk into a meeting room. It shows up in the meeting. It takes notes. You pick it back up on email, and then you text it if you need it.

Tibo SottiauxOpenAI

你不需要一直粘在笔记本电脑前，也不用一直盯着手机。它就是在你需要的时候随时都在，不需要的时候也不会碍事。我非常期待这一天成为现实，因为我觉得走到哪儿都得扛着笔记本这块「砖头」，意味着你被技术绑住了，而不是让技术为你所用。

It's not that you need to be glued to a laptop or you're always on your phone. It's just like it's available when you need it, and it also gets out of the way when you don't need it. I can't wait for that to come to fruition, because I think just carrying your laptop as this brick around everywhere is just you're tied to the technology instead of the technology working for you.

Tibo SottiauxOpenAI

哦，我觉得它还会继续发生相当剧烈的变化。哪怕是今天，技术看起来终于在语音、多模态输入和输出上汇集到了一起，但用起来还是有点笨拙。我觉得它会一直笨拙下去，直到某一天突然不再笨拙，然后你就会觉得：「哇，我可以像我们现在这样跟它聊天。它能把事情记得一清二楚。如果我想构思、头脑风暴，它可以随手画个草图，我有一个可以协作的界面。」

Oh. I think it's going to continue to change quite radically. Even today, it does seem like the technology is finally coming together with voice, multimodal inputs, outputs, but it still feels a bit clunky. I think it will feel clunky until it isn't, and then, you're just like, "Wow. I can just talk to this thing the same way that we're talking right now. It just remembers things perfectly. If I want to ideate and brainstorm, it can just squiggle a thing, I have this collaborative surface."

Tibo SottiauxOpenAI

这正是我们用 ChatGPT Space 发布的东西，可以说是这一切的起点。你可以想象我们有一块共享白板，人和人之间、人和智能体之间都能在上面协作。我觉得它会超越对话本身，超越你今天用的很多客户端。这些东西今天都还没真正跑通。所以我觉得我们会经历一场新的大变革。

This is what we launched with ChatGPT Space. It's like the very start of that. You can imagine we have a shared whiteboard. We can collaborate between humans, between agents, and I think it's going to transcend conversations, transcend a lot of the clients that you have today. None of this is just really working today. So I think we're going to go through this big new transformation.

Tibo SottiauxOpenAI

是的，至于 Codex 和 ChatGPT，我们会把……我们有 Chat，也有 Work 开关。我们清清楚楚地听到了大家的反馈。大家很喜欢 Work 里的能力，但有时候也还是想用 Chat，因为它更快一点，用起来也更舒服。所以我们正在把两者合并，降低这种复杂度。最终，我们也会把 Dots 的全部能力直接搬进 ChatGPT，抬高我们 12 亿用户的体验底线，当然到那时我们会有多少用户，就看情况了。

Yes, and for Codex and ChatGPT, we're going to merge the ... We have Chat. We have the Work toggle. We heard the feedback loud and clear. People love the capability in Work, but then, also, sometimes you want to use Chat. Because it's a little bit faster and more pleasant to use. And so we're merging that. We're reducing that complexity. Eventually, we will ship all of the capabilities of dots straight into ChatGPT, as well, and lift the floor for our 1.2 billion users, which, at that time, we'll see how many we have.

Tibo SottiauxOpenAI

我们希望它真的、真的非常顺滑。Dots 有一点让我特别兴奋：它没有模型选择器，完全不需要配置，你直接跟它说话就行。你唯一需要配置的，就是你想通过哪些渠道跟它交流。

We want it to be really, really seamless. One thing that I'm very excited with dots is they don't have a model picker. There's just no configuration. You just talk to it. The only thing that you have to configure is which channels you want to talk to.

Tibo SottiauxOpenAI

我只看过一遍。

I watched it once.

Tibo SottiauxOpenAI

我觉得去想这个挺有意思的，也许更多是……我想得更多的是 30、40 年前的科幻作品，以及它们在当时有多么超前。

I think it's interesting to think about maybe more ... What I think about more is sci-fi from 30, 40 years ago and how visionary it was at the time.

Tibo SottiauxOpenAI

像《神经漫游者》这样的书，还有最早的《星际迷航》。我经常想起它们。《星际迷航》有一点很特别，我觉得它里面的技术是非常非常实用的。显然，你可以直接跟电脑说话，它就替你把事情办了。你可以跟你的飞船说话，它就开动了，对吧？这些正在变成现实。我们正在进入那个时代。

Books like Neuromancer and then the original Star Trek. I think about that a lot, and there's something about Star Trek, which I think it was this technology that was very, very useful. Obviously, you're able to just talk to a computer, and it does things for you. You can talk to your spaceship, and it just moves. Right? It's like this is getting real. We're entering that era.

Tibo SottiauxOpenAI

我觉得成为社区的一部分、和大家一起去构建，是非常重要的，因为从某种意义上说，我们是在一起探索这项技术，探索我们能共同做成什么。所以每次和大家交流，我都会被提醒：你在做一件我根本没预料到的事；它以某种方式影响着真实的生活，也影响了另一个人的人生。所有这些加在一起，既让我非常谦卑，也让我深受鼓舞，而且我觉得，要打造真正对人类有用的东西，这是非常必要的。没有社区的话，我都不知道我们该怎么做下去。

I think it's very important to be part of a community and to build with people, because in a way, we're discovering this technology and what we can do together. And so every time I talk to people, I'm reminded of you're doing this thing which I hadn't even anticipated. It's impacting real life in a certain way, and it's impacted this other person's life. Then it's just like all together, it's both very humbling, very inspiring, and I think very necessary in order to build something that is truly useful for humans. I don't know how we would do it without the community.

Tibo SottiauxOpenAI

对。总的来说，就连 Twitter 社区也一直挺友善的。嗯。

Yeah. Even the Twitter community has been very nice in general. Yeah.

Tibo SottiauxOpenAI

我每天大概花半个小时……

I spend maybe half an hour per day-

Tibo SottiauxOpenAI

……就刷一刷，等有灵感、有点有意思的话想说的时候，就发点东西。很多时候都是一时兴起。

... just scrolling and then doing something when I feel inspired and I have something interesting to say. Often it just comes in the moment.

Tibo SottiauxOpenAI

团队 Dots 我们还没发布，但今天我们发布了主 Dot，你可以对它进行配置。你可以先熟悉一个 Dot 能做什么，把它连到你的各种应用上，之后我们会开放创建多个 Dot 的能力。比如我就有一个专门负责 Twitter 的。

We haven't released the team dots yet, but today, we're releasing the primary dot, which you can configure. You can get used to what a dot can do, connect it to your apps, but we'll release the ability to create more than one. And so I have one that's responsible for Twitter.

Tibo SottiauxOpenAI

对。

Yeah.

Tibo SottiauxOpenAI

没错。

That's right.

Tibo SottiauxOpenAI

对。我们想先从一个开始，也就是主 Dot，你大概也会给它发消息，它对你偏好的了解也最深。然后看大家拿它做什么，看我们需要怎么调整这个系统，在和社区的协作中学习。接下来很快，你就能加第二个、第三个、第四个，需要多少就加多少，组建你的虚拟团队。然后你可以给它们分配特定的角色，对吧？

Yeah. So we wanted to start with one, the primary dot, which is the one that you'll probably also text, that learns your preferences most deeply, as well, and see what people do with it, see how we need to tune the system, learn by working with the community. Then, very quickly, soon, you'll be able to add a second one, a third one, a fourth one, as many as you need to create your virtual team. Then, you can give them specific roles. Right?

Tibo SottiauxOpenAI

我倒不觉得这是绝对必要的，但有时候我手上有件事需要相当多的工作量，比如帮我盯着 Twitter。那就成了一件独立的事。就好像：「好吧，这活儿多到够一个 Dot 全职干了。」然后你可以有好几个这样的 Dot。

I don't find it absolutely necessary, but sometimes, I have a thing that requires quite a bit of work, such as monitoring Twitter for me. Then, that's a thing. It's just like, "Okay. This will be so much work that is the work of an entire dot," and then you can have a couple of those.

Tibo SottiauxOpenAI

对，对，对。我觉得真正的黑马是生态，也就是把一切都开放出来，以及我们对开放一切的承诺，非常坚定的承诺。「用 ChatGPT 登录」现在有 16 个合作伙伴，我对此非常非常自豪。这件事其实是去年某个时刻临时起意开始的。当时我在和 Pi 和 OpenCode 的作者聊，就觉得：「当然了，你们就应该能直接用 Codex 登录，然后用上它的额度。」

Yeah. Yeah, yeah, yeah. I think the sleeper hit is a ecosystem, so opening it all up and the commitment, the deep commitment we have towards opening it all up. So we have, signing with ChatGPT, 16 partners. Very, very proud of that. We started that really, it was sort of in the moment last year. I was talking to creators of Pi and OpenCode, and it was like, "Of course you should just be able to use the Codex sign-on and then use the usage."

Tibo SottiauxOpenAI

然后我们就隔空握了个手，意思就是：你们用这个认证吧，我们相信你们不会做什么见不得人的事。后来这件事发展起来了，显然变得相当受欢迎。现在我们把它变成了一项正式支持的功能，和很多很多合作伙伴一起，我很期待能快速把它做大。另一方面，我们也在开放所有基础设施，以及我们为 ChatGPT 构建东西的方式，包括插件扩展，还有插件发现，让每个人都能把东西推给大约 12 亿用户，享受这种分发能力。

Then, just, we shook hands virtually. We're just like use this auth, and then trust you to not do anything shady. That grew. Obviously, that became quite popular. Then, now, we made it an actual thing that we support with many, many partners, and I'm excited to grow that very quickly. The other side, as well, is opening up all of the infrastructure and how we build for ChatGPT with plugin extensions, also plugin discovery, just allowing everyone to ship something through maybe 1.2 billion users and benefiting from that distribution.

Tibo SottiauxOpenAI

我们还有收益分成，这点其实在主题演讲里没讲：对于那些受欢迎、使用量很大的插件，我们会付费给它们，它们也会拿到一部分收入分成。所以我觉得这种对开放生态的承诺会非常令人兴奋。

We also have shared economics that we didn't actually talk in the keynote, but we will pay our plugins that are popular and are seeing a lot of usage. They're going to get part of the revenue share, as well. And so I think this commitment towards an open ecosystem is going to be very exciting.

Tibo SottiauxOpenAI

是的。我们有大量订阅 ChatGPT 的用户，他们有一定的使用额度。当用户把这部分额度用在插件上，或者用在别的产品里，也就是通过「用 ChatGPT 登录」来使用时，我们会和这些合作方分享收益，他们会得到报酬。

Yes. So we have all these subscribers that use ChatGPT. They get a certain amount of usage. Then, when the users use that usage with the plugin or within the other product, when they use sign-in with ChatGPT, we will have some shared economics with them where they will get remunerated.

Tibo SottiauxOpenAI

做一个好插件。对。它的运作方式是这样的，当然我们会持续调整这套系统，但我们会看留存数据，看插件有多成功、质量如何，然后据此开始在对话里向用户推荐。所以你的插件可能会被推荐给相当大一部分用户。当然，如果你的插件不够好，它就不会再被推荐了。

Build a good plugin. Yeah. The way that it works, and we'll obviously tune the system, but we look at retention numbers. We look at how successful the plugin, the quality, and then, that's what we then start to recommend to users in conversations. And so your plugin can just get recommended to a significant slice of users. Obviously, if your plugin is not quite good, it will stop being recommended.

Tibo SottiauxOpenAI

没错。它是不是真的带来了价值？它有没有让 ChatGPT 能做到别的事？

That's right. Is it actually adding utility? Is it enabling ChatGPT to do something else?

Tibo SottiauxOpenAI

没错。

That's right.

Tibo SottiauxOpenAI

这次才是对的那个。

This one is the right one.

Tibo SottiauxOpenAI

你还记得 Codex Cloud 吗？

Do you remember Codex Cloud?

Tibo SottiauxOpenAI

那是一年前了，对吧？如果你回头看 Codex Cloud，看看我们当时用的那个小动画，我觉得那就是 Grok Bot 的灵感来源。

That was a year ago. Right? If you look back at Codex Cloud, and you look at the little animation that we used, I think that was the inspiration for Grok Bot.

Tibo SottiauxOpenAI

不过说正经的，在研究层面，长时间、持续的长程任务，这是我们做了两年多的东西。能保持连贯的记忆系统，也差不多做了这么久。其中很多我们直接上线到了 ChatGPT 里。我觉得大家真的很喜欢 ChatGPT 能了解到关于你的这么多相关信息。我聊过的很多人都有类似的小故事，比如：「我病了，结果 ChatGPT 记得我之前参加了一个烧烤，我拿着青柠在调龙舌兰。它说我手上这块灼伤很可能就是青柠引起的，再加上晒伤。」他们就觉得：「它怎么会知道？」其实就是 ChatGPT 里的记忆功能做得非常非常好，我们在这上面投入了大量研究。

But seriously, on the research, long, persistent, long-horizon tasks, this is something that we've worked on for more than two years. Memory systems that stay coherent, roughly the same amount of time. We launched a lot of that straight into ChatGPT. I think people just really enjoy that ChatGPT gets to know so much relevant information about you. So many people I talk to have an anecdote of like, "Hey. I fell ill, and ChatGPT remembered, actually, I had this barbecue. I had lime, and I was making tequila. Just, probably, this burn that I have on my hand is actually this lime, it's like, and sunburn." They're just like, "How did it know that?" Then it's like, it's just memory works very, very well in ChatGPT, and so we've put a lot of research in that.

Tibo SottiauxOpenAI

所有这些，再加上 Codex 运行框架，以及能 24/7 不间断运行并保持高效的能力，汇聚到一起，就是我们这次发布的 Dots。所以这是很多很多个月的工作成果。另外一件很重要的事是把安全和防护做对。这也是为什么我们发布时搭配的是 Astra。它是我们最安全、对齐程度最高的模型，我们也投入了多得多的精力，确保 Dots 真正安全可靠。

All of this coming together on top of the Codex harness and also the ability to just run 24/7 with good productivity, that's what we launched with dots. And so it's the work of many, many months. Then, one of the important things was also getting safety and security right. This is also why we launched with Astra. It's our safest, most aligned model, and a lot more effort went into making dots just really safe and secure.

Tibo SottiauxOpenAI

是的。我觉得这很大程度上取决于我们如何构建这项技术。我们在 OpenAI 设计一切的方式，就是把人放在中心，把它打造成人的延伸，延伸你的意志和你的品味，让它真正成为一种极其赋能的东西。我觉得只要我们继续这样做，它就能让你把任何你想做的事，把你的创造力和品味，以一种让人感觉很棒的方式放大。在当下，它其实就是一种艺术工具，对吧？

Yes. I think this is very much a function of how we build the technology and the way that we design everything at OpenAI, is to put humans at the center of it and build it as extensions of humans, of your will and your taste, and so just really be this super empowering thing. I think as long as we continue to do that and it enables you to just take whatever you wanted to do, and your creativity, and your taste and expand that in a way that feels awesome. In the moment, it's just really this artistic tool. Right?

Tibo SottiauxOpenAI

我们以后可能不再有程序员了，但构建者会比以往任何时候都多。我觉得这里面有某种东西会始终深深属于人。人想要学习，想看看别人在造什么，我对这个要感兴趣得多。我在这儿，我不是在跟 Dot 说话，我是在跟你说话，对吧？我觉得这一点在非常非常长的时间里都会成立。

We may not have coders anymore, but we have more builders than ever. I think there's something that's going to remain deeply human about that. Humans want to learn and see what other humans are building, and I'm much more interested. I'm here. I'm not talking to dot. I'm talking to you. Right? I think that will remain true for a very, very long time.

Tibo SottiauxOpenAI

有，一直在想。减少配置疲劳是其中之一。还有就是改变「跟智能体对话是一场独自冒险」这件事：你只是和你的一个智能体对话，然后你有了很多个，把很多事情都委派出去。很多东西会汇集起来，让这件事变得愉快得多。这是我花了很多时间、团队也花了很多时间在思考的事。

Yes. All the time. Reducing configuration fatigue is one. Reducing the fact that talking to an agent is a solo adventure, and you just have this conversation just with your one agent, and then you have many and then you're delegating so many things, a lot of things are going to come together to make it much more delightful. It's something that I spend a lot of time, the team spend a lot of time thinking about.

Tibo SottiauxOpenAI

所以对我来说，把事情做成的理想方式，是让它就存在于你的物理空间里。我们就像现在这样聊天，它能观察到：「这些是我们提出的想法。」我们可以在一张白纸上随手记点东西。它也许能拿着这个开始在后台构建，然后你说：「其实我还有另一个想法。」它就开始构建另一个东西。

So I think for me, the ideal way to get things done would be to just have it in your physical space. We're just having conversation like we have now. It's able to observe, "Here are the ideas that we have." We can jot something down on a white paper on a paper. I can maybe take that and start building in the background, and then you're like, "Actually, I have another idea." It's just like it starts building another thing.

Tibo SottiauxOpenAI

你就把它投到屏幕上，然后直接跟它说话。它就参与到和其他人的对话中。你不会有那种必须盯着屏幕、满脑子想着怎么写提示词的疲劳。一切都变得非常自然。这是我们正在走向的方向。我觉得今天还没完全做到，但未来我们会做到。

You just project it on a screen, and then you just talk to it. It's just involved in a conversation with other humans. You don't have this fatigue of having to be on this screen, and you're just thinking about prompting. It just becomes super natural. That's what we're trending towards. I think we didn't fully get there today, but we'll get there in the future.

Tibo SottiauxOpenAI

我觉得我们也听到 Sam 讲了这件事的愿景，就是减少噪音，让你把注意力放在你想放的地方。现在有各种各样的事情，也许重要也许不重要，但都在争夺你的注意力。就是要把这些调低一点，让你专注于真正重要的事。

I think we heard Sam, as well, about the promise of this, as well, which is reduce the noise, and allow you to spend attention where you want to spend attention, and having all these things that are maybe important, maybe not important, but vying for your attention. It's just sort of like tone that down a bit and just get you to focus on the things.

Tibo SottiauxOpenAI

我每次去度假，花一周时间彻底断网，都会觉得很神奇。我开始用不同的、更有创造力的方式思考。我很想看看，我们能不能把这种状态带回来？能不能让它成为你的日常？也许你需要更少的会议。也许你不需要做那么多事，而如果你休息得更好，其实会更有生产力。所以我觉得整个行业得把这个问题想清楚，但我认为这才是 AI 的承诺所在，而不只是每秒多发一条提示词。

I find it remarkable every time I go on a holiday, and I take that one week to just fully disconnect. I start to think in different and more creative ways, and I'm very eager to see, can we bring that? Can we just make this your day-to-day? Maybe you need fewer meetings. Maybe you don't need to do as much, and you would actually be more productive if you're better rested. And so we'll have to, I think, as an industry, figure that out, but I think that's the promise of AI. It's not just one more prompt per second.

Tibo SottiauxOpenAI

噢。有一件挺酷的事，因为刚才现场演示当着所有人的面失败了，对吧。当时我想：「好吧，这可不太妙。」但其实，我的 Dot……它意识到我在 DevDay 现场。当时 ChatGPT 的生产环境挂了，所以它在现场演示前五分钟提醒了我，这其实让人挺紧张的。因为它说：「嘿，生产环境挂了。」然后又说：「噢，要不要我试着修一下？」

Oh. There was something cool, because the live demo just failed, right, in front of everyone. So it was like, "Okay. Well, that's not cool." But actually, my dot was ... It's like realized I was at DevDay. We have ChatGPT production went down, and so it pinged me five minutes before the live demo, which was actually quite stressful. Because it was like, "Hey. Production is down." It's like, "Oh. Well, do you want me to try and fix it?"

Tibo SottiauxOpenAI

我说：「小 Dot，我觉得你还没到那个水平，但谢谢你的尝试。」然后我联系了工程团队，我们开始排查到底发生了什么。现在已经修好了。但有这么一个东西能理解：「好，现在有件挺重要的事在发生，叫 DevDay。有这样一个生产系统。有一场现场演示，它很可能用的就是这个生产系统，所以这两件事是相关的。这件事五分钟后就要发生了。所以我应该提醒他，因为他大概会想知道。」我觉得这相当了不起。

I'm like, "I don't think you're there yet, little dot, but thank you for trying." Then, I got in touch with the engineering teams, and we started to look at what was going on. It's fixed by now. But having this thing that just understands, "Okay. There's a pretty important thing happening. It's called DevDay. There's this production system. There's a live demo. It's probably using this production system, so these two things are connected. This is happening in five minutes. So I should ping him, because he probably wants to know about it." I think that's quite remarkable.

Tibo SottiauxOpenAI

我当时也没跟它说「去吧，让我刮目相看」，不过要是那样，故事会更精彩。

I don't think it was like, "Please go impress me," but that would've been a better story.

Tibo SottiauxOpenAI

它能访问一部分生产系统，但有防护措施。我们聊到过专用 Dot，专用 Dot 就是在额外的防护措施、额外的监控下运行，而且跑在它们自己的硬件上。我们其实有一些跑在 Mac mini 上。我们构建 Dots 的方式有一个根本性的不同：运行框架并不跑在那台机器上。这一点大家可能还没意识到：它可以连接任意多台设备。它有自己的电脑，你也可以把它连到你自己的笔记本上。慢慢地，你可能会把它连到 10 台不同的设备上，它能把这些全部控制起来，有点像一只章鱼。

It has access to some production systems with guardrails. We talked about specialized dots, but specialized dots is just operate with additional guardrails, additional monitoring, and then on their own hardware. So we actually run some on Mac minis. The fundamental difference in the way that we've built dots is the harness does not run on the machine. Then, this is maybe people have not realized, is like it can connect to as many devices as you want. So it has its own computer. You can actually connect it to your own laptop, as well. Over time, you might connect it to 10 different devices, and it can control it all, a little bit like an octopus.

Tibo SottiauxOpenAI

对。可能是 VM，也可能不是 VM。它就是一个东西，能连接很多设备。

Yeah. Might be a VM, might not be a VM. It's just like it's a thing, and it can connect to many devices.

Tibo SottiauxOpenAI

嗯。我觉得在走下坡路的技能是打字快。哦，这个现在真的没那么有用了。越来越重要的，就是出色的品味，站在用户的角度思考，和你要服务的人群建立连接。很多前创业者现在都做得特别成功，或者说连续创业者。我们在 OpenAI 招人时也发现了这一点：我们有很多创始人。我想 OpenAI 现在有超过 120 位前 YC 创始人，所以它就像一家超大型创业公司。我觉得那种热情、那种想做出真正有意义的东西的渴望，再加上知道什么是「好」，比以往任何时候都更重要。

Yeah. I would say skill that is trending down is typing fast. Oh. That is not that useful anymore. Skills that are trending up is just great taste, thinking about the user, connecting to the audience that you're building for. A lot of ex-founders are incredibly successful right now or just repeat founders. We're also finding this when we're hiring at OpenAI, is we have a lot of founders. I think we have more than 120 ex-YC founders currently at OpenAI, so it's like this mega startup. I think just this passion and this envy to build something that matters and then know what good looks like is more important than ever.

Tibo SottiauxOpenAI

我同意。

I agree with that.

Tibo SottiauxOpenAI

而且角色之间的边界也在变模糊。你以前可能只做设计，或者只做工程，总觉得有点施展不开。但现在，正是你大放异彩的时候。

Also, roles are blurring, so you might have been only into design or only an engineer and felt always a little bit uncomfortable. But now, it's just like it's your time to shine.

Tibo SottiauxOpenAI

我怀念过一阵子，但现在有了非常快的速度，我又找回了那种感觉。所以我非常希望能尽可能让更多人拥有它。要让它真正无处不在、覆盖超过十亿用户，还需要一段时间。但现在有了 Ultrafast 的速度，我们已经能做到了，而这才过去半年左右。它的成本和 Astra 大致相当，所以这方面的进展相当惊人。

I missed that for a while, and then, now, with really fast speeds, I'm finding that back. And so I'm really excited to democratize that as much as possible. It will take a while for it to be absolutely ubiquitous and available to more than a billion users. But now, with Ultrafast speeds, we're able to achieve it for six months old. It costs roughly the same as Astra. And so the progress there is quite astonishing.

Tibo SottiauxOpenAI

就连我们内部也觉得「哇，原来我们也能做这些事」。因为我们在用 Astra，发现它能被推到非常远的地步。有了这样的速度，尤其是能用语音来控制它的时候，你就重新回到了一种非常有创造力的状态，回到了心流。它和过去的心流不一样，但我并不怀念过去。

Even for us internally, we're like, "Wow. We can do these things," also. Because we are using Astra, and we're finding we can push it super far. When you have those speeds, and especially when you can be able to voice control it, you kind of get back into this really creative state of mind and this flow. It is different from the flow of the past, but I don't miss it.

Tibo SottiauxOpenAI

没错，是这样。

That's right. Yeah.

Tibo SottiauxOpenAI

嗯。我觉得下一个 Tibo 已经在 OpenAI 了，他叫 Ahmed Ibrahim，我很喜欢和他一起工作。他加入 OpenAI 的时候还是应届生，现在负责 Applied 部门整个 OpenAI 的算力集群，这是一份非常非常重大的责任，Codex 的运行框架很多也是他搭的。让 Ahmed 脱颖而出、取得成功的，就是他……他人特别好，特别善于协作，总是去解决真正重要的问题，而且从不把自己放在第一位。

Yeah. I think the next Tibo is already at OpenAI. His name is Ahmed Ibrahim. I love working with him. He was a new grad when he was hired at OpenAI. He's now responsible for all of OpenAI compute fleet in Applied, which is a massive, massive responsibility, and built a lot of the Codex harness. The thing that has made Ahmed really stand out and successful is just his ... He's incredibly kind. He's incredibly collaborative and always tries to solve the problem that is really important, but in a way that he's not putting himself first.

Tibo SottiauxOpenAI

他就像一块海绵。他学得特别快，什么都学，还会用各种最新的技术来学习，那种速度我以前从没见过。所以在短短几个月到一年的时间里，他在公司的成长简直让人惊叹。现在，在 OpenAI 最关键的那些发布上，他是我非常信任的人。

He is just a sponge. He's learning so fast, and learning about everything, and using all of the latest technology to learn at speeds I haven't seen before. And so he, just in the span of a couple of months and a year, it's like his progress at the company was just absolutely fascinating to see. He's now someone I trust a ton in doing the most direly launches at OpenAI.

Tibo SottiauxOpenAI

你应该请，真的应该。

You should. You should.

Tibo SottiauxOpenAI

我不知道这算不算意外，但就像我说的，有很多前创业者，在自下而上地推动很多令人兴奋的想法。比如 Decisions API 就是很快攒起来的。我们意识到 Luna 是个非常好的模型，只需要做一些约束采样，再把它以不同于 Responses API 的形态发布出来就行。这样做之后，它更快了，在某些场景下体验相当惊艳。如果你去看看幕后，就会发现，哦，原来是有这么一个 Slack 频道自发组织起来的。

I don't know if that's surprising, but it is, as I said, a lot of ex-founders doing a lot of bottoms-up, very exciting ideas. For example, Decisions API came together very quickly. We realized we have a really good model with Luna. We can just do some constraint sampling and then have it shipped as a different shape to the Responses API. Then, when we do that, it's faster, and it's quite delightful for some cases. It's just, if you were to feel under the hood, you realize, okay, there's this Slack channel that came together.

Tibo SottiauxOpenAI

一开始只是四个人周末在上面捣鼓，然后开放给公司内部试用。接着大家兴奋起来，开始拿它做东西，然后说「哦，我们可以支持视觉输入，这比市面上任何东西都好」。然后大家越来越兴奋，开始以一种完全没有组织的方式出手帮忙，最后一路推进，把东西发布出去。

Initially, it was four people hacking on it on a weekend, and then it was made accessible as company food. Then, people got excited, started to build things, and they're like, "Oh. We can support visual inputs. This is even better than whatever is out there." Then, people get more and more excited. People start pitching in in a way that feels completely unstructured, and then take it all the way and ship something.

Tibo SottiauxOpenAI

然后我们会努力守住很高的质量标准，在东西发出去之前叫停，或者它可能还需要再打磨一下。今天 DevDay 上原本可能还要发布更多东西，后来我们觉得「好吧，这已经够多了，先压一压，把节奏拉开」。所以下周和下下周我们还会有更多发布。总之就是这种惊人的自下而上的能量，再由一些人担起角色，把这股能量引导到有成效的方向上。

Then, we try to hold a high bar for the quality and then stop things before they go out, or maybe they need a little bit more baking. There's many more things that we were maybe going to ship today at DevDay. Then, we were like, "Okay. I think this is a lot already. Let's hold it back a little bit and space it out." So we'll have more launches next week and the week after. But it's just this incredible bottoms-up energy, and then people assuming roles to channel that in productive ways.

Tibo SottiauxOpenAI

你拥有很大的自主权，然后就得为此负责。所以我们就是信任大家会做出好的选择，出了问题就赶紧修，或者从错误中吸取教训。到目前为止都很好，它确实营造了一个非常赋能的环境。确实，我可以在必要的时候、觉得合适的时候按下重置按钮，这是一种特权。这也让我能真正贴近社区，我觉得换个环境是不可能做到这种程度的，对吧？我不需要层层审批。

You get a lot of autonomy, and then you have to just own up to it. So we just trust people to make great choices, and then, when things go wrong, fix it quickly or learn from the mistake. So far, so good. It's just really worked as a very empowering environment. It is true that I can press the reset button whenever is necessary and whenever it feels right, and that's a privilege. Also, it allows me to just really be close to the community in ways that I think would not be otherwise possible. Right? I don't have to run it through an echelon of approvals.

Tibo SottiauxOpenAI

有。我觉得有些时候，我本可以更好地激励团队去做不那么复杂的东西，而不是一直往复杂的方向走。Codex 早期我们有过几次宕机是我造成的。我刚到 OpenAI 的时候，第三天就把生产环境搞挂了。我当时就，嗯。不过那之后我还是保住了工作，也吸取了教训。

Yes. I think at times, I could have made teams more inspired to build things that are less complex and just really continue to strive towards complexity. I think we, early days in Codex, we had a couple of outages that I caused. Very, very earliest for me at OpenAI, I took production down day three. I was just, yeah. But I was still employed after that, and I learned my lessons.

Tibo SottiauxOpenAI

取决于我们把东西搞坏几次。

Depends how many times we break things.

Tibo SottiauxOpenAI

对。搞坏一些东西，也庆祝一些东西。

Yeah. Break some things, celebrate things.

Tibo SottiauxOpenAI

我觉得有太多事情还没有被充分预期到。我认为互联网上的大多数操作，将由智能体来完成。模型会以相当惊人的速度变得更便宜、更快。我们终于能以非常无缝的方式把所有模态整合到一起。就这三点，我常常觉得，当我看大家在做的东西时，会想：不对，你们还没完全想明白。你们已经很接近了，但如果再逼自己一把，真正去想象一年后这些都比今天好上大约 10 倍，你的做法就会完全不同。

I think there's so many things that I feel are not yet priced in. I think the majority of actions on the internet will be taken by agents. Models are going to become more cheaper and faster at rates that are quite incredible. We will finally be able to integrate all modalities together in a way that is very seamless. I think those three, I feel like often, when I look at what people are building out there, it's like, no. It's just like you're not quite getting it. So you're almost there, but if you were just pushing yourself and just really imagining all of these being roughly 10 times better than it is today in a year, it's like you would build in a different way.

Tibo SottiauxOpenAI

没错。

That's right.

Tibo SottiauxOpenAI

有，很多。首先，如果你想让你的产品在智能体那里取得成功，你就得按一定的规模去构建。比如我们和 Notion 合作得很紧密，他们做了自己的 MCP 之后，突然就发现，哦，原来所有这些能真正干活的智能体都能用这个 MCP 了。于是他们迎来了大量流量，这显然给系统带来了很大压力，你得想清楚这里面的经济账。

Yeah. There are many. First of all, if you want your product to be successful for agents, it's like you have to build for a certain level of scale. We've worked very closely with Notion, for example, and when they build their MCP, suddenly, it's like, oh, well, it's available to all these agents that can actually do the work and use that MCP. And so they saw a ton of traffic come in, and that puts, obviously, a lot of strain on the system. You have to figure out the economics of that.

Tibo SottiauxOpenAI

所以如果你在做产品，就会有这样一种张力：要不要做一个界面？你可以先拖一阵子，但这是不可避免的，对吧？大多数东西都会由智能体来使用。我觉得朝着那个未来去构建非常重要。另一方面，为人类打造真正令人愉悦、能充分利用各种模态的全新体验，我觉得这一块投入还远远不够。

And so there's this tension between if you're building products, it's like, do you build an interface or not? You can hold that back for a while, but it is inevitable. Right? The majority of things are going to be used by agents. Building towards that future, I think, is very important. The other thing is on the flip side, building absolutely delightful new experiences for humans that really benefit from older modalities is something that I feel like is underinvested in.

Tibo SottiauxOpenAI

我原本以为我们要到一两年后，才能达到今天这些模型的能力水平，比如 Astra。所以我不得不修正自己的预期。另外，我也没想到自己会这么依赖语音，我现在很多事情都是靠听写，或者直接跟我的智能体通话来完成的。我没预料到这种感觉会好这么多，所以在这点上我也改变了看法。

I was expecting us to reach the level of capabilities of the models that we have today, say like Astra, maybe in a year or two. And so I had to revise my prioriter. Also, I did not expect to rely so much on voice, and I do so much through dictation or just calling my agent. I had not anticipated that, how much better that feels. So I had to change my mind there.

Tibo SottiauxOpenAI

还有，你刚才聊到招聘，有一点我之前没意识到：年轻一代身上有着惊人的才华和能量，他们会是最先拥抱这一切变化、并想清楚如何驾驭它的人。所以这也改变了我对招聘以及招聘策略的看法。

Then, you talked about hiring, and one thing that I hadn't realized is the incredible talent and energy just younger generations have and how they would be the ones embracing all this change first and figure out how to harness all of it. And so that changed, also, my opinion on hiring and the strategy there.

Tibo SottiauxOpenAI

对，而且我觉得还有那种超快的吸收和学习能力。

Yeah, and I think just also the ability to just absorb and learn super fast.

Tibo SottiauxOpenAI

对我来说，把控前沿的节奏，就是真正提前投入，而且投入得比我们自认为应该投入的还要多得多：投入模型的对齐和安全，投入安全防护，投入护栏。我们在二级监控上花的算力越来越多。也就是说，一部分算力投入到主系统，也就是正在干活的那个智能体；另一部分算力则用来监控这个主智能体，确保它不会采取风险过高的操作，一旦看起来它可能遭到了提示注入，就打断它、进行干预。

For me, pacing the frontier is just really investing ahead and also investing way more than we even think we ought to, investing in the alignment and safety of the models, investing in security, investing in the guardrails. We are spending more and more compute on secondary monitoring. So you have all of the compute that is going into the primary system, the agent that is doing work, and then you have all of the compute that is going into monitoring the primary agent to make sure that it's not taking too high risk of actions or not interrupting it if anything looks like maybe it's getting prompt injected, and just intervening with that.

Tibo SottiauxOpenAI

我们在 API 技术栈上的大部分投入，其实都用在了安全技术栈上。所以对我来说，这就是把控节奏……你要确保自己极其稳固。这就是我们在做的。我们还没有发布比 Astra 更进一步的能力跃升。我们发布的是一个水平相近、接近 Astra 智能，但效率高得多的版本。所以我们对此很放心，但我们会继续在这方面大力投入。到目前为止，我对我们的做法感觉很好。我想新闻里也报道过，我们有 6.1 Astra，但最后没有发布。这是我非常引以为豪的一件事。

The majority of our investment on the API stack is actually going into the safety stack. And so to me, that is facing ... It's like you want to make sure that you're extremely hardened. That's what we're doing. We haven't yet released the next step-up in capability beyond Astra. We have released a level that is similar, near Astra intelligence, but much more efficient. And so we feel very comfortable about that, but we're going to continue to invest a ton there. I feel very good about our approach so far. I think it was also in the news that we had 6.1 Astra, and then we didn't release it. That is something I'm very proud of.

Tibo SottiauxOpenAI

是的，而且，我觉得这显然是一份沉甸甸的责任。但如果你想想 OpenAI 的激励机制：我们在为 12 亿人做产品，我们不能把这搞砸，对吧？我们非常非常认真地对待自己交到这么多人手里的东西，其中包括很多本身并不懂技术的人。你必须把它做对，所以在这上面我们不赌。

Yes, and it's also, it's, I feel, obviously, a deep responsibility, but also just if you think about the incentives for OpenAI, we build products for 1.2 billion people, we can't screw that up. Right? We take it very, very seriously, what we put in the hands of so many people, including people who are not very technical themselves. You have to get it right, and so we don't gamble there.

Tibo SottiauxOpenAI

应用现在的状态啊。我觉得模型已经差不多了，但还没到能让应用几乎完全消失的程度。我迫不及待想真正达到那种极致简洁的本质。连我自己都会被模型选择器、推理强度、要不要用多智能体或者 ultra、它到底是干嘛的这些东西搞得很累。你差不多得有个模型选择器博士学位才行，对吧？我只想尽快把这些全都去掉。

Current state of the app. I feel like the models are almost there, but not quite for the app to almost completely disappear. I can't wait to just really get to this super essence of simplicity. I myself even get fatigued with the model picker, and the reasoning efforts, and whether to use multi-agent or ultra, or what it even does. You kind of need a PhD in model pickers. Right? I just want to get rid of all of that as quickly as we can.

Tibo SottiauxOpenAI

嗯，谢谢你邀请我。嗯。

Yeah. Thank you for having me. Yeah.