Muse以实用场景起飞
"If you were looking at the third-party reports, you would see that this might be the fastest-growing agent or AI app that I think has existed yet."
"Then that's why I was saying it's kind of like become an eval for this too."
购物、训练指导等真实任务既推动增长，也成为检验智能体能力的评测。
人人都在无感写代码
"They're just not going to know that they're writing code."
"It just pulls out frames from it and it's like too much on my back leg there."
智能体会把自然语言要求转成视频处理等程序，用户无需直接接触代码。
多模态决定消费体验
"I think most consumers have this highly multimodal experience."
"The glasses basically will eventually just be a pure multimodal product, right?"
消费者通过声音、图像和视频理解世界，因此多模态不是附加能力，而是产品核心。
模型本身就是产品
"So one of the things that we say internally is like, the model is the product."
"But the capabilities, the personality, a lot of the stuff is like you're training into the model."
智能体的能力、性格和主动性主要由训练决定，传统界面只是较薄的一层。
消费产品必须去复杂化
"I just want to make it simple."
"Like, I just like, how is somebody who's not carefully reading API design specs supposed to know what they're supposed to pick?"
普通用户不该理解模型命名和推理档位，系统应自动选择合适能力。
主动记忆是智能体底座
"We call it active memory."
"How well does it remember things that you mentioned to it over days, weeks, or months, and then leverage those in the process of doing future tasks?"
记忆不只是上下文长度，而是跨越数周或数月保存信息并用于后续任务。
内部评测服务真实体验
"Ultimately, we care about making useful models."
"And so you're kind of outsourcing all of your product thinking to that PhD student who happened to make a viral benchmark."
公共榜单适合确认研发方向，但成熟产品应围绕自身用户场景设计评测。
人才密度胜过团队规模
"And the very first thing is that you don't need a lot of people, and talent density is the thing that matters."
"We're not peanut buttering it across a lot of competing efforts."
小型高密度团队配合集中算力，比把资源平均分给大量项目更有效。
模型能力要多轴扩展
"I mean, if I were to give, say, like one broad statement about AI, it's just scale everything, you know?"
"But, um, uh, but in particular, you have to scale everything."
预训练、数据、强化学习、算力和智能体框架需要同步扩展，单轴放大并不理想。
长周期迫使提前下注
"It's nice when you have metrics, but I think the, like, state of being an entrepreneur is, like, how well do you make decisions in an information-poor environment, right?"
"It's like the opposite of Lean Startup."
数据中心和前沿模型耗时多年，企业必须在结果得到验证前做出巨额承诺。
智能体将成为新买家
"Like, now agents are effectively going to have accounts."
"Because it's not just the people anymore."
智能体不仅替人下单，还可能创造原本不存在的消费需求，扩大商业市场。
核心技术决定平台上限
"The companies that did the best kind of had the core technology to be able to serve people the best there."
"But with the internet, there were, like, okay, 5 or 6 very big companies and then a whole ecosystem of other companies that are great."
未来仍会有多家赢家，但掌握前沿模型与端到端技术的公司更能定义产品方向。
"I kind of disagree."
为什么非共识 · 多数人觉得人工智能产品会迅速收敛成同一种智能体，他说差异只是藏进了模型行为、性格、工具使用和评测体系。
"Like if you— in a lot of cases, less is actually more when it comes to data in RL."
为什么非共识 · 多数人觉得训练数据越多越好，他说强化学习中少量高质量环境可能更有效，因为错误奖励会被模型放大。
"And so you're kind of outsourcing all of your product thinking to that PhD student who happened to make a viral benchmark."
为什么非共识 · 多数人觉得公共榜单能客观决定模型优劣，他说盲目追榜等于外包产品判断，因为评测未必对应真实用户需求。
"Like, you cannot do that in AI."
为什么非共识 · 多数人觉得应先小规模验证再逐步投资，他说前沿人工智能不能照搬精益创业，因为算力建设和模型研发必须提前多年下注。
"I think that that would be— from my perspective, I think that is probably a riskier configuration for the future than lots of people having access to different things."
为什么非共识 · 多数人觉得少数机构集中控制强模型更安全，他说多样化和广泛可用反而更安全，因为能形成制衡并降低单点风险。
If you were looking at the third-party reports, you would see that this might be the fastest-growing agent or AI app that I think has existed yet.
从第三方报告来看，这可能是迄今增长最快的智能体或人工智能应用，至少我这样认为。
What are your guys' personal favorite Muse use cases, either yours or other people's?
你们个人最喜欢用 Muse 做什么？可以谈谈你们自己的用法，也可以谈谈其他人的用法。
My favorite thing is I have my Muse, whose name is Euler, constantly source increasingly obscure teas from increasingly obscure vendors of tea.
我最喜欢的用法是让我的 Muse——Euler——不断搜罗越来越小众的茶叶，而且要从越来越小众的茶商那里寻找。
So at this point he's, on a weekly basis, my Muse is buying some random tea from some farm in Taiwan or some farm in Japan and sourcing it over and bringing it back.
现在，我的 Muse 每周都会从台湾或日本的某个茶园寻找一些稀奇的茶叶，购买后再带给我。
My favorite both use case and probably eval for the product is having it coach me in MMA.
我最喜欢的应用场景，同时也可能是对这款产品最好的测试，就是让它指导我练习综合格斗（MMA）。
It watches the cameras when I'm in there, and it gives me feedback in between the rounds, which is a very hard problem, it turns out, because not only does it have to...
我在里面训练时，它会查看摄像头画面，并在每个回合之间向我提供反馈。事实证明，这是一个非常困难的问题，因为它不仅需要……
It basically has to process all the video quickly enough because you only get 1 minute in between the rounds.
它必须足够快地处理完所有视频，因为每个回合之间只有 1 minute。
I've wondered about this too, because looking at videos is a really good AI use case.
我也一直在思考这个问题，因为视频分析是一个非常好的人工智能应用场景。
Because really basically what it's doing, it's like, you know, within the agent, I'm not seeing any code, but it basically is writing this whole video processing, image processing pipeline.
在这个智能体中，我看不到任何代码，但它实际上是在编写一整套视频处理和图像处理管线。
And it's, I think, a good example of how stuff is going to go in the future where basically everyone is actually going to be writing tons of code.
我认为，这很好地说明了未来的发展方式：届时，基本上每个人都会编写大量代码。
It was definitely one of his first use cases actually of even the initial prototypes. Yeah.
这绝对算是他当初对最早期原型的首批使用场景之一。
In the model's training data, like the fine-grained, this is what an MMA fighter looks like when they're appropriately balanced, and this is when they're like too much on the back foot.
模型的训练数据中是否包含如此细粒度的内容，例如综合格斗（MMA）选手保持正确平衡时的状态，以及重心太靠后脚时的状态？
Just, I'm wondering, how has it consumed so many images of appropriate versus inappropriate MMA posture?
我很好奇，它究竟看过多少图像，才能分辨正确和错误的综合格斗（MMA）姿势？
Again, does it know from seeing lots of images of MMA or does it know because it understands the concept of where you're meant to be balancing?
换言之，它是因为看过大量综合格斗（MMA）图像才知道这一点，还是因为它理解重心本来应该放在哪里这一概念？
It's read books about MMA, and then it sees that you're too much on the back foot, or—
它是通过阅读综合格斗（MMA）方面的书，判断出你的重心太靠后脚，还是——
So, I mean, I think that that's like kind of making good multimodal models.
因此，我认为这基本就是打造优秀多模态模型的方式。
I mean, this gets into like, I think, a core approach thing that we have, which is we're training the models end-to-end to be good personal agents, right?
这涉及我们的一项核心思路：我们正在对模型进行端到端训练，使其成为优秀的个人智能体。
First of all, I think just superintelligence overall is going to allow us to run the company better and build all kinds of different things.
首先，我认为从整体来看，超级智能将使我们能够更好地运营公司，并打造各种不同的事物。
But for the products that we're building, there's certain capabilities and certain mixes of skills that basically are just much more applicable to these cases.
但对于我们正在打造的产品而言，有些能力和技能组合更适用于这些场景。
So I mean, this is one of the things that Alex focused on from the beginning was like, all right, we really want to make sure that our multimodal program is good because that's disproportionately important for the types of things that people use all these services for.
这是 Alex 从一开始就重点关注的事情之一。当时的想法是，我们必须确保多模态项目足够出色，因为对于人们使用这些服务完成的各种事情而言，多模态能力尤为重要。
And I think ultimately, a lot of this, it comes down to how are you focused on it, right?
归根结底，这在很大程度上取决于对这件事的重视程度。
I think because of the fact that we wanted to build personal superintelligence.
我们之所以重视这件事，是因为我们希望打造个人超级智能。
And fundamentally, I think most people in the world really experience the world in a highly multimodal way.
从根本上说，世界上大多数人确实是以高度多模态的方式体验世界的。
They like talking to other people and listening to things and seeing videos and filming things of their own.
他们喜欢与他人交谈、聆听各种内容、观看视频，也喜欢亲自拍摄内容。
And especially earlier this year, there was actually a point of tension almost in the industry because I think so much of the consensus view within the AI industry was that it was just all about coding and coding was all that mattered.
尤其是在今年早些时候，行业内一度出现了一种紧张局面。当时人工智能行业内部的主流共识在很大程度上认为，一切都围绕编程展开，编程才是唯一重要的事情。
And I think that for the success of a product like Muse and the success of a lot of these consumer use cases, it really depends on the ability to be strong and multimodal. Yeah.
Muse 这样的产品以及许多面向消费者的使用场景能否成功，实际上都取决于其是否具备强大的多模态能力。
I mean, a lot of the initial theory that we had was that, I mean, most people don't actually like reading that much, right?
我们最初的许多设想都基于一个事实：大多数人其实并不那么喜欢阅读。
So even the text-based interaction model, I think it's something that people do.
因此，即使是基于文本的交互模式，也只是人们可能采用的方式之一。
But when you think about stuff like the glasses that we're working on, or even our other products that are very visual, the glasses basically will eventually just be a pure multimodal product, right?
但对于我们正在研发的眼镜，以及其他视觉属性很强的产品而言，眼镜最终将成为一种完全多模态的产品。
It's like you can let it hear what you are listening to, and it can talk to you, and it can see what you see if you let it.
例如，你可以允许它听到你正在听的内容；它也可以与你交谈；如果获得你的允许，它还能看到你所看到的事物。
And then eventually, I don't even know if we're going to have an app store on the glasses.
最终，我们甚至未必会在眼镜上设置应用商店。
It'll just be like a multimodal generation that kind of creates whatever you need right in your vision.
它或许会通过多模态生成，直接在你的视野中创造出你所需要的一切。
And I imagine something like that will probably end up powering a lot of the apps that we build too.
这样的技术最终也可能为我们打造的许多应用提供支持。
I want to ask you guys about the product space here in AI and try something out on you.
我想了解你们对人工智能产品空间的看法，也想请你们评估一个想法。
If I was to describe to you, say, with Stripe, what I think a winning product looks like, it would take me pages and pages to talk about because we have the developer experience and we have to talk about all the international heterogeneity of payment methods. Then we have to talk about stablecoins.
以 Stripe 为例，如果要描述我心目中一款成功产品的形态，可能需要一页又一页的篇幅，因为我们既要讨论开发者体验，也要讨论全球不同地区支付方式的巨大差异，还要讨论稳定币。
Whereas I notice a phenomenon in AI where it feels like the winning products collapse very quickly onto an agreed-upon product space that everyone is pursuing, and so transformers worked, and then like the chat agent UI really worked.
但我注意到，人工智能领域存在这样一种现象：成功的产品似乎很快就会趋同于一个获得普遍认可的产品空间，也就是所有人都在追逐的方向。Transformer 成功了，聊天智能体这类用户界面也确实奏效了。
You mentioned OpenClaw earlier, the kind of OpenClaw persistent agent, has a computer, able to do things.
你之前提到过 OpenClaw。OpenClaw 这类常驻智能体拥有一台计算机，能够完成各种任务。
But there's something about AI where the winning products tend to be very simple, pursued by all the companies that are taking it seriously.
但人工智能有一个特点：最终胜出的产品往往非常简单，而且所有认真投入这一领域的公司都在追逐同一种产品。
There's, you know, huge amounts of frontier work that goes into making a good version.
为了做出一个优秀的版本，背后需要投入大量前沿研究工作。
But the winning brand—now you're talking about the glasses or the future versions.
但最终胜出的品牌——你现在说的是眼镜，还是未来的版本？
But somehow there's a very small number of clearly correct products in AI as opposed to it being this, like, very broad and complicated product space.
但不知为何，人工智能领域显然正确的产品只有寥寥几种，不像其他领域那样拥有非常广阔而复杂的产品空间。
Well, first of all, I think part of the theory from the beginning was that this kind of agent or whatever form it took was going to be quite a different thing from this text-based assistant, right?
首先，我认为从一开始，部分设想就是，这种智能体——无论它最终采取什么形态——都会与文本助手截然不同。
It reaches a ton of people, but I mean, it's like the same thing that all these companies have.
Meta AI 覆盖了非常多的用户，但它其实与其他公司拥有的同类产品相差不大。
And we, I think, specifically think that that is used for very informational purposes, but people don't think about it as like a thing that gets stuff done for them.
而且我们尤其认为，它主要用于获取信息，人们不会把它视为能够代替自己办事的工具。
You know, I think that it's only kind of this most recent wave that gives the agent a dedicated VM, which I think is actually incredibly important for making you feel confident that the security and privacy of your information is—it's not just like, okay, it's not connecting to some elastic compute that just everyone uses.
直到最近这一波产品出现，智能体才开始配备专属虚拟机。我认为，这一点对于让用户相信自己的信息在安全和隐私方面得到了保障极其重要。它并不是接入某种由所有人共用的弹性计算资源。
But my claim isn't that it's easy to implement, nor that it's obvious—
但我并不是说它很容易实现，也不是说答案显而易见——
Because again, my claim is that now that people see Muse, everyone looks at that and says, "Oh yeah, that's clearly how you want the AI agent product to work, and that the future AI agent products that we'll see will be fairly Muse-shaped."
因为还是那句话，我的观点是，现在大家看到 Muse 后，都会看着它说：“哦，对，人工智能智能体产品显然就该这么运作，而且我们未来会看到的人工智能智能体产品大致都会是 Muse 这种形态。”
My take on this is, I think we're just so early in the realm of what AI products will look like that I think a lot of the interesting differentiation about the products kind of gets collapsed in how we view the products today.
我的看法是，我们还处在探索人工智能产品形态的极早期阶段，所以我认为，产品之间很多有意思的差异，在我们今天看待这些产品的方式下，多少被掩盖了。
I think the traditional way in which we wrote down and think about product specs, that is...
我觉得，我们过去记录和思考产品规格的传统方式，也就是……
Those aren't the relevant dimensions and levers and axes to consider the important design decisions on AI-native products.
那些并不是我们在考量人工智能原生产品的重要设计决策时，真正需要关注的维度、杠杆和轴线。
And so, Muse, for example, I think so much attention was put in the design of the model for the product.
所以，以 Muse 为例，我觉得它在设计上投入了大量精力，去为产品打造模型。
And there were a lot of very intentional decisions around, in all these situations, how exactly should the model behave?
而且针对所有这些情境，我们都做了很多非常有意识的决定，比如模型究竟应该如何表现？
And probably in the same way that the full spec of how we wanted the model to behave in all these scenarios would be an extremely long document.
同样，如果要把我们希望模型在所有这些场景中如何表现的完整规格全部写下来，那很可能会是一份极其冗长的文档。
But because of just how we think about products today, it's just AI.
但按照我们如今思考产品的方式，它就只是人工智能。
It's kind of like how when mobile first started happening, it was like, "Oh, it's just an app and everything is just an app." Well, it turns out there's a lot of nuance in what exactly an app is.
这有点像移动互联网刚开始兴起时，大家会说：“哦，它就只是个应用，一切都只是应用。”但事实证明，应用究竟是什么，里面其实有很多细微差别。
And we're kind of in this phase where, oh, it's just AI, or it's just an agent, or it's just an agent with a VM or something like that.
而我们现在就处于这样一个阶段：哦，它就只是人工智能，或者只是个智能体，又或者只是个带虚拟机的智能体之类的。
One thing that we talked about from early on, is that basically there is this pattern where I think the products are getting thinner and the model is getting to be more of the product, right?
我们从一开始就讨论过的一件事是，基本上有这样一种趋势：我觉得产品正在变得越来越轻，而模型本身正越来越成为产品，对吧？
So one of the things that we say internally is like, the model is the product. Yes.
所以我们内部常说的一句话就是：模型就是产品。对。
So, and when you're talking about an agent, a lot of the complex product work that went into Muse is basically around the privacy and security model for handling your data and the things that get connected to it.
所以，当你谈到智能体时，Muse 背后大量复杂的产品工作，基本都围绕着处理你的数据以及接入其中的各种内容时所采用的隐私和安全模型。
And that, I think, is something that we're just very proud of that we really focused on.
这是我们非常自豪的一点，我们确实将重心放在了这方面。
But the capabilities, the personality, a lot of the stuff is like you're training into the model.
不过，产品的能力、个性以及其他许多方面，都是通过训练融入模型的。
And whether they're the people in the lab who are working on that or the people on Nat's team who are working on the product, I think when Nat is thinking about what is the key stuff that we need to make the product feel good, he's talking about post-training the model, right?
无论是实验室中从事这项工作的人员，还是 Nat 团队中负责产品的人员，当 Nat 思考需要做好哪些关键事项才能带来良好的产品体验时，他所谈论的其实是模型的后训练。
I think there's like, there's still very viscerally, I think from a user, there's a feeling of using Muse over time.
从用户的角度来看，随着时间的推移，使用 Muse 依然会带来一种非常直观而强烈的感受。
And maybe some meaningful portion of that is just the reliability of Muse to handle certain tasks.
其中相当一部分可能来自可靠性，也就是 Muse 处理某些任务时的可靠程度。
And when it talks to you, how do you feel about how it talks to you?
此外，它与你交谈时，其说话方式会带给你怎样的感受？
And I'm starting to see more and more of this actually show up on the X feed, which are people are starting to tease out these, like...
我发现，X 的信息流中开始越来越多地出现这类讨论，人们开始梳理这些问题，例如……
They're all proactive agents, but what are the actual meaningful differences between them?
它们都是主动型智能体，但彼此之间真正有意义的区别是什么？
And these, I think, are the interesting product decisions, which, you know, the funny thing is how do you specify these product decisions?
我认为，这些才是有意思的产品决策。但值得思考的是，应当如何定义这些产品决策？
Well, it's ultimately in model design decisions, which ultimately boil down to training methods or data or evals or, you know, all these things.
归根结底，这些都是模型设计决策，而模型设计决策最终又取决于训练方法、数据或评测等因素。
I think somewhat infamously, a lot of the AI products have been designed by a highly technical group that do not have these sort of randomly sampled American in their mental model.
许多人工智能产品多少有些声名不佳，因为它们是由一群技术水平很高的人设计的，而这些人所设想的用户画像并不包括随机抽样出来的普通美国人。
Some of the times the numbers go backwards and sometimes it's just like a totally different word and you have another set of words.
有时数字会反向变化，有时又会突然换成一个完全不同的词，随后还会出现另一套词。
And like, I just like, how is somebody who's not carefully reading API design specs supposed to know what they're supposed to pick? Like, it's just far too confusing.
我不禁想到，一个没有仔细阅读 API 设计规范的人，怎么可能知道自己应该选择哪个？这实在令人摸不着头脑。
Um, yeah, the new model—Muse Wonder—it's a new class of model that is at the frontier.
这个新模型——Muse Wonder——是一类全新的前沿模型。
Basically, when you're doing these pre-trains, you don't just go to whatever the biggest cluster is that you have.
进行这类预训练时，通常不会一开始就使用手头规模最大的集群。
You want to work your way up the ladder training models that are maybe 4 or 5 times bigger, and then you just keep on repeating that process.
训练规模会逐步提升，可能先训练规模大 4 或 5 倍的模型，再不断重复这一过程。
So the first set of models that we've had Muse Spark was all based on this initial pre-training that we did in, what was it, December?
我们最早推出的那批模型 Muse Spark，全都基于第一次预训练。那次是在什么时候，12 月吗？
I think we've all just been quite surprised and impressed with how much intelligence and capability the team has been able to build into a model which is like frankly not that big, right?
我认为，我们所有人都非常惊讶，也很佩服团队竟然能将如此强大的智能和能力，融入一个其实并不算大的模型中。
I mean, and then after that we were working on this next model, and the pre-training has been done for a bit, and we've been working on the post-training, and now it is here.
之后，我们开始开发下一个模型。它的预训练已经完成了一段时间，我们则一直在投入后训练工作。现在，它终于推出了。
So I think you you will feel that in terms of the nuance in which it can think.
我认为，从它思考时所展现出的细腻程度中，就能感受到这一点。
A lot of the areas that we focused on were making it the best personal and professional agentic model.
我们重点投入的许多工作，都是为了让它成为最出色的个人及专业智能体模型。
Some of the different things that you can optimize for that's it for us, and then yeah, a lot of the basic capabilities that we focused on building in.
模型可以针对各种不同方向进行优化。对我们而言，重点就是这些方向，以及我们着力构建的许多基础能力。
Very good memory, very good ability to persist through complicated tasks.
它的记忆力非常出色，也非常善于坚持完成复杂任务。
But a ton of work went into this across the lab and the product teams and all of this.
整个实验室和产品团队等各方都为此投入了大量工作。
And I think it's also just sort of a testament to the the team being able to methodically figure out the process to scale these models.
我认为，这也恰好证明了团队有能力有条不紊地探索出扩展这些模型的方法。
So I think this is also just, you know, a milestone on the way towards the time between each subsequently bigger class of models, I think will now start to compress.
我认为，这也算是迈向一个重要节点的里程碑：从现在开始，此后每一代更大规模模型之间的时间间隔将逐渐缩短。
But the team basically had to go build up like different tiers of models to get to this one.
不过，团队基本上必须先构建不同层级的模型，才能打造这个模型。
And now that we are kind of at the frontier, now everything that we do going forward will just be pushing the frontier, as opposed to there's a lot of background work that we needed to do before that.
现在，我们已经站在前沿。接下来所做的一切都将继续推动前沿发展，而不像以前那样，还需要先完成大量基础工作。
I mean, we've seen like incredible performance across a slate of evaluations on personal, professional agent tasks.
我们看到，它在一系列涵盖个人和专业智能体任务的评测中都表现得非常出色。
And then there's a number of, I think, specific highlights which we're really excited about.
此外，还有几个令我们非常兴奋的突出亮点。
So active memory, like Mark mentioned, is one area where we think that to build personal agents, personal professional agents, like memory is just one of the clear requirements to build reliable agents and is one of the areas where we've invested, I think, disproportionately actually compared to many of the other efforts.
例如马克刚才提到的主动记忆。我们认为，要构建个人智能体以及专业智能体，记忆显然是打造可靠智能体的必要条件之一。这也是我们实际投入远超其他许多项目的领域之一。
And we have our own internal evals of Memory Bench which demonstrate a lot of the capabilities there.
我们还有自己的内部评测 Memory Bench，展现了它在这方面的多项能力。
How well does it remember things that you mentioned to it over days, weeks, or months, and then leverage those in the process of doing future tasks?
对于你几天、几周甚至几个月前向它提过的事情，它能记得多牢，又能否在执行未来的任务时利用这些信息？
Very happy OpenClaw user, but it ultimately does this compaction where it has like the long-term memory and the working short-term memory or whatever.
我使用 OpenClaw 时很愉快，但它最终会进行这种压缩，将长期记忆、工作短期记忆等内容压缩起来。
And I think once people use OpenClaw for a while, they find the memory pretty buggy.
大家使用 OpenClaw 一段时间后，就会发现它的记忆功能存在不少 bug。
But so this is outside the model, like the model exists and it's—
那么，这是模型之外的东西吗？例如模型本身已经存在，然后它——
And it's about how it's able to leverage various sources to be able to conduct tasks in the future, as well as understand its ability to do compaction, how it should do that compaction most effectively, et cetera.
关键在于，它如何利用各种信息源执行未来的任务，同时了解自身的压缩能力，以及如何以最高效的方式进行压缩等。
Active focus is another, which is effectively just the ability for the model to, when conducting tasks, be able to carry them through, especially as we've seen Muse users as well as Muse Code users, give higher and higher complexity and more and more ambiguous and more and more complex tasks, we want the model to be able to conduct those and do those very effectively.
另一项能力是主动聚焦，也就是模型在执行任务时能够将任务贯彻到底。我们已经看到，Muse 用户和 Muse Code 用户提出的任务复杂度越来越高，任务本身也越来越模糊、越来越复杂，因此我们希望模型能够非常高效地执行并完成这些任务。
And this is one of the things actually in early dogfooding and testing that a lot of people have noticed about the model that is sort of special and stands out with respect to other models that they've tried or used.
在早期的内部试用和测试中，许多人都注意到这个模型具有某种独特之处。与他们尝试或使用过的其他模型相比，它显得格外突出。
It's basically like the ability to conduct complete tasks over very long horizons.
它主要指在很长的时间跨度内完整执行任务的能力。
And then when faced with various setbacks along the way, the ability to identify other solutions to those problems in an aligned way.
当执行过程中遇到各种挫折时，模型还要能够以符合目标的方式找到其他解决方案。
It's actually a big step up in coding capabilities, so frontier coding capabilities.
它的编码能力也得到了大幅提升，达到了前沿编码能力水平。
So it is actually state-of-the-art on DeepSWE and, SWE-Marathon and does very well on other benchmarks like Terminal-Bench 4 and many others.
它在 DeepSWE 和 SWE-Marathon 上达到了最先进水平，在 Terminal-Bench 4 等许多其他基准测试中也表现得非常出色。
So we think that users who use it for coding, which will be a minority of all of our users, but the ones who do will see a big improvement and a big step up.
我们认为，使用它编程的用户虽然只占所有用户中的少数，但会看到显著的改进和大幅提升。
Okay, so we should also just be using Muse Wonder Full for coding at Stripe?
那么，我们在 Stripe 编程时是否也应该直接使用 Muse Wonder Full？
Okay, so this is both just your latest frontier model, like people can come, you know, developers can come directly to you, they can use it on OpenRouter, like they can just consume it as a model, and it's also a big upgrade inside of Muse that people will just feel as active users of Muse.
因此，这既是你们最新的前沿模型，开发者可以直接通过你们或 OpenRouter 使用，将它作为模型调用；同时，它也是 Muse 的一次重大升级，Muse 的活跃用户会直接感受到这种提升。
Well, what we're going to do in Muse, because it's a lot bigger and it's more expensive to serve, we're going to start off primarily in paid subscriptions.
由于它的体量大得多，服务成本也更高，因此我们最初主要会在 Muse 中通过付费订阅提供它。
So what we want to do with Muse as a product, we want it to be free for everyone to use.
对于 Muse 这款产品，我们希望所有人都能免费使用。
So we have this huge amount of free tokens that people can use with with Muse Spark as the model.
因此，我们提供了大量免费 token，用户可以使用以 Muse Spark 为模型的版本。
But if you want to use Wonder for most of the hard tasks, then at least for now, you'll use the paid version.
但如果想使用 Wonder 处理大多数高难度任务，那么至少目前需要使用付费版本。
But, you know, I mean, eventually we'll, you know how quickly the field moves.
不过，这个领域发展得非常快，我们最终会……
And we hope to, you know, at some point not too far from now, have a model that is as good as Wonder, but in the Spark form factor.
我们希望在不久后的某个时候推出一个效果与 Wonder 一样好、但采用 Spark 形态的模型。
And that will be a great thing to be able to serve hundreds of millions of people around the world.
这将是一件很棒的事，可以为世界各地的数亿用户提供服务。
Okay, so you go to Muse, you subscribe to the—what's it called—like the premium or pro plan, or...
那么，打开 Muse 后，需要订阅高级版、专业版之类的套餐，还是……
And then it'll also be available via API, on the Meta Model API and OpenRouter.
它也会通过 API 提供，可在 Meta Model API 和 OpenRouter 上使用。
And then Muse Wonder Full will be available to subscribers of Muse Code.
此外，Muse Code 的订阅用户也能使用 Muse Wonder Full。
And so we're working through how to serve that most effectively to as many people as possible.
因此，我们正在研究如何以最高效的方式，为尽可能多的人提供服务。
I mean, this is like, and we really with Muse Spark, we were very happy with the performance that we were able to deliver as we were getting up to this point.
Muse Spark 已经让我们对迄今实现的性能非常满意。
But like, this is, the beginning of, like, MSL kind of being at the frontier.
但这标志着 MSL 开始跻身前沿。
We feel like the models are quite competitive with the best models that are out there. And I think that this is—
我们认为，这些模型与目前最优秀的模型相比很有竞争力，而且我认为——
Okay, so you think it's like Astra, Fable, and Wonder altogether at the frontier?
那么，你认为 Astra、Fable 和 Wonder 都达到了前沿水平？
And I think actually one of the things that we're especially proud of is we have taken the time, as we've mentioned, to optimize models for the product form factor of Muse.
我们尤其自豪的一点是，正如刚才提到的，我们花了时间针对 Muse 的产品形态优化模型。
We think it really is the best model in the world for the use case of personal professional agents.
我们认为，对于个人专业智能体这一用例，它确实是世界上最好的模型。
You know, a metric that becomes a target ceases to become a good metric.
一项指标一旦成为目标，就不再是一项好的指标。
You hear complaints that models are benchmaxed and that at a certain point, kind of overfitting to the benchmarks actually makes the models slightly worse.
有人会批评模型都在针对基准测试刷榜，而且达到一定程度后，对基准测试的过度拟合反而会使模型略微变差。
I mean, as we've gone on, we steer the models more towards our internal benchmarks for the use cases that we're focused on.
随着工作的推进，我们会让模型更多地针对内部为重点用例设计的基准测试进行优化。
So, I mean, early on when we were kind of just doing the science of trying to climb the ladder, we cared more about what the external benchmarks were just to make sure we were on the right track.
刚开始时，我们基本上只研究如何不断提升模型能力。当时，我们更重视外部基准测试的结果，以确保方向正确。
And then I'd say, you know, each release that goes by, the larger share of what we're looking at are internal ones.
随着一次次发布，内部评测在我们所关注的内容中占据了越来越大的比重。
And one of the things that we actually quantified in our research blog post as part of Muse Wonder is this metric for big model smell, which is kind of, you know, big model smell and feeling of benchmaxing are kind of like inverses of one another.
事实上，我们在研究博客文章中量化的一项指标，也是 Muse Wonder 的一部分，即“大模型感”指标。“大模型感”和“刷榜感”有些像是彼此的反面。
This idea that you could just like tell when you're using a really big model, it just... You can tell.
当你使用一个真正的大模型时，能够直观地感受到这一点。
And so we, the way we quantify that is actually almost explicitly what you're mentioning, which is the gap between in-distribution performance and out-of-distribution performance. So what is the...
我们量化它的方式，几乎就是你刚才提到的分布内表现与分布外表现之间的差距。因此——
We kind of measure how well does the model perform on tasks that are in distribution of our RL and post-training versus tasks that are out of distribution.
我们大致会衡量模型在两类任务上的表现：一类是处于我们的强化学习和后训练分布内的任务，另一类是分布外任务，并对两者进行比较。
What does that gap look like between the Muse Spark line and the Muse Wonder line?
Muse Spark 系列与 Muse Wonder 系列之间的这一差距具体如何？
And we actually see this huge step up from the Muse Spark to the Muse Wonder line because of, as Mark mentioned, like, it's just a much better base model powering this new line.
事实上，我们看到 Muse Spark 到 Muse Wonder 系列实现了巨大跃升。正如 Mark 提到的，驱动这个新系列的基础模型确实强大得多。
And there's an entire je ne sais quoi to that, but I think we've been able to actually quantify this effect.
这背后存在一种难以言喻的感受，但我认为，我们已经能够真正量化这种效应。
And it's an obvious dimension that we're excited to continue improving on.
这显然也是一个我们非常期待继续改进的方向。
But if you're saying that external benchmarks have ceased to become as important, and you care much more about your internal benchmarks.
但你的意思是，外部基准已经不再那么重要，而你们更重视自己的内部基准。
Is that kind of an argument against benchmaxing, that you think that just trying to get the top score on the external benchmarks that everyone talks about is maybe not going to lead you to develop a great model?
这是否可以视为一种反对“刷榜”的论点？换言之，你认为一味追求大家都在讨论的那些外部基准的最高分，可能无法帮助你开发出一款出色的模型？
I mean, part of the whole strategy is personal superintelligence and just trying to build different types of products than other people are trying to build.
整个战略的一部分是实现个人超级智能，而我们希望打造与其他人不同类型的产品。
And though I would say a personal superintelligence is not like a super niche interest, like it's probably something that other models want to be able to do.
不过，个人超级智能并不是特别小众的研究方向，其他模型可能也希望具备这种能力。
I mean, one of the things that we talk about is this idea of discretion, right?
我们谈到的一个问题，就是所谓的“谨慎披露”。
So if you want to like use the model to get something done, It needs to be able to know a lot about you and get things done in the world while disclosing as little as possible.
如果你想用模型完成某件事，它就需要充分了解你，在现实世界中把事情办成，同时尽可能少地透露信息。
So, you know, if you're— you have an allergy or you're pregnant or something, you're trying to, like, get a, you know, like, a table at a restaurant and you don't want to, like, have to tell them your medical condition. Yes, yeah.
如果你有过敏症、处于孕期或存在类似情况，并且想预订餐厅座位，你可能不愿向对方透露自己的健康状况。
And I mean, humans have common sense and the models so far If you're building a coding agent that works in an environment where everyone has access to the same context, discretion isn't particularly important.
人类具备常识。但到目前为止，如果你构建的是一个编程智能体，而它所在工作环境中的每个人都能访问相同的上下文，那么这种权衡判断的能力就没有那么重要。
So there are all these things that I think are just kind of more important for us.
因此我认为，对我们而言，还有许多类似的问题更为重要。
Aspects of coding where the code needs to work, but you don't— the consumers who are using it may not even ever see the code.
在编程的某些方面，代码必须能够正常运行，但使用它的消费者甚至可能永远看不到代码。
I mean, that probably matters more in a personal superintelligence context rather than serving software developers.
我认为，与服务软件开发者相比，这一点在个人超级智能的情境下可能更加重要。
So we want models when you put them into people's hands, they have a great experience using.
因此，我们希望把模型交到人们手中时，能够为他们带来出色的使用体验。
And I think that is the most important metric for any model that we or other people produce.
我认为，这是衡量我们或其他人所开发的任何模型时最重要的指标。
I think that really focusing on the external benchmarks, if you really dig into the supply chain a little bit, it's some PhD student who made these benchmarks, a benchmark as part of their, as the course of them doing their PhD.
如果深入追究背后的来龙去脉，就会发现，对外部基准测试的过度关注，实际上源于某个博士生在攻读博士期间，将制作这些基准测试中的某一项作为其博士研究的一部分。
And so you're kind of outsourcing all of your product thinking to that PhD student who happened to make a viral benchmark.
于是，你实际上把所有产品层面的思考，都外包给了那个碰巧制作出一项热门基准测试的博士生。
A huge part of what a lot of people at Meta and MSL are focused on is crafting amazing evals.
Meta 和 MSL 的许多人都在重点投入一项重要工作，即精心设计出色的评测。
It's like those public ones are not a crazy place to start, but it's just not where you want to end up.
从这些公开评测入手并无不妥，但肯定不能止步于此。
But then when you're like really in this mode, as we are now, where it's like, you know, we want to be very intentional about how we're improving our model over time to build better and better, better products.
但当你真正进入我们目前所处的模式时，目标就是有意识地规划如何持续改进模型，打造越来越优秀的产品。
Like a lot of the cases that Mark just mentioned, like these are actually detailed very rigorously in like the spreadsheet of all the little capabilities and nuances of the product that we think really matter to the end user.
例如 Mark 刚才提到的许多情况，都在一份电子表格中得到了极其严谨细致的说明。表格列出了产品在能力和表现上的所有细微差别，而我们认为，这些差别对最终用户确实很重要。
And maybe each of them matters like 2% to the user, but in aggregate, that's the obsession that will ultimately result in these models that feel magical.
也许每一项对用户的重要性都只占 2%，但这些项目叠加起来，正是这种执着最终造就了那些令人感觉如同魔法一般的模型。
Can you guys describe for me just what goes into training an RL model?
你们能否介绍一下，训练一个 RL 模型涉及哪些环节？
I'm just curious in that I think a lot of people listening to this have a rough idea where there's there's a pre-training and a post-training step.
我对此有些好奇，因为我认为许多听众都只有一个大致概念，知道训练分为预训练和后训练阶段。
And after a while you start to develop a sense for what is catness and what is dogness and what is fur and everything like that.
经过一段时间，模型会逐渐理解猫和狗各自的特征、什么是皮毛，以及其他类似的概念。
And so that's kind of the pre-training step where you just hoover up data and maybe you can talk about the data acquisition and just kind of what goes into that.
这就是预训练阶段，模型会吸收所有这些数据。也许你可以谈谈数据获取，以及这个阶段大致涉及哪些内容。
And then there's this RL stage, which is the newish advance over kind of the past few years of AI.
接下来是强化学习阶段，这是过去几年人工智能领域相对较新的进展。
And again, it sounds like that is where a lot of the specific nuance of this model has come in, where you guys have chosen better domains, better hills to climb for the model.
听起来，模型的许多具体细节也是在这个阶段形成的。你们会选择更好的领域，也就是更适合模型攀登的山头。
But just at a high level, describe how many people, how long it takes, what the steps are, how you think about those in—
不过，请先从宏观层面介绍一下：有多少人参与、需要多长时间、包含哪些步骤，以及你们如何看待这些问题。
so how many people are on this whole effort, like when you develop these models?
整个项目共有多少人参与？例如，在开发这些模型时，有多少人参与？
So this, I think you're getting into what is interesting about MSL and the approach that we've taken, which is, I mean, From the very beginning, I think when I first reached out to you, I think we were just looking at this because we were like, just like, wow, okay, how far we've come.
这正体现了 MSL 的有趣之处，以及我们所采取的方法。我记得从一开始，也就是我第一次联系你的时候——我们刚才还在回顾那段经历，并感叹自己竟然已经走了这么远。
And I was just like, I pulled up the screenshot of like when I first reached out to Alex.
随后，我调出了第一次联系 Alex 时的截图。
My first thing was like, hey, do you have time to catch up and talk?
我发出的第一句话大概是：“嘿，你有时间叙叙旧、聊一聊吗？”
And then after that call, 2 days later, Alex sends me this long thing, which is like, well, here's what I think I would do if I were in your position.
那通电话结束 2 天后，Alex 给我发来一大段话，大意是：“如果我是你，我会这样做。”
And the very first thing is that you don't need a lot of people, and talent density is the thing that matters.
第一条是：你不需要很多人，真正重要的是人才密度。
So I'd say, like, from the beginning of the plan, this was, like, the core of your strategy, right?
因此，从制定计划之初，这就是你策略的核心吗？
It's basically, like, you actually don't need a massive number of people, but what you do need is for, like, each of the people to be as good as possible, which kind of informed this kind of crazy recruiting push and, like, very intense filtering that we had to take, you know, some of the best people from inside Meta and a very intense process to get people from outside Meta.
归根结底，你并不需要一支非常庞大的团队，但需要确保每个人都尽可能优秀。这也促成了这场近乎疯狂的招聘行动和极为严格的筛选：我们必须从 Meta 内部挑选一些最优秀的人才，同时通过一套非常严格的流程从 Meta 外部招聘人才。
But I actually think like a lot of what has made MSL special and why a lot of people have wanted to join is because I think it's in some ways it's just like more fun to work on a talent-dense, smaller effort.
我认为，MSL 之所以特别，并能吸引如此多的人加入，很大程度上是因为与一支人才密度高、规模较小的团队共事更有意思。
In some ways it has all the benefits of a startup with the stability and resources of a larger company.
在某种程度上，它既具备初创公司的所有优势，又拥有大公司的稳定性和资源。
And this was like this environment that I think this was like point one in your kind of plan for what, um, for what Alex was like, this is what we need to go do.
我认为，这样的环境正是你那份方案中的第一项计划。Alex 当时表达的意思是：这就是我们接下来要做的事。
So I don't know the answer of like how big is a pre-training team, you know, it's, it doesn't need to be that big.
因此，我也很难回答一支预训练团队应该有多大。事实上，它并不需要很多人。
But I think if you look at, you know, some of the other labs, um, it's probably quite big in some other places.
不过，如果观察其他一些实验室，就会发现那里的团队可能相当庞大。
So, um, I think part of the reason why, uh, these models are doing well is because you You have people who we're betting on and they have a lot of resources because we're taking our compute.
我认为，这些模型之所以表现出色，部分原因在于我们把赌注押在一些人身上，并将算力集中提供给他们，使他们获得大量资源。
I mean, internally, like, Avocado was the code name for the Muse Spark line, and then Watermelon was the code name for this one.
在内部，Avocado 是 Muse Spark 系列的代号，Watermelon 则是这次这个模型的代号。
But, um, I think the only fruit that is technically larger is the durian, and you don't want to go to the durian. No, no, no, no.
不过，严格来说，唯一更大的水果可能就是榴莲，但你肯定不想用榴莲。绝对不行。
But no, yeah, I think as Mark mentioned, like, you know, each of the— I mean, each of the stages I'll describe, like, it's tens of people working on it at Meta versus, you know, hundreds at—
不过，正如 Mark 提到的，我接下来会讲到的每个阶段，在 Meta 都只有几十个人参与，而其他地方可能有几百人参与。
So Meta is a huge company, but this is like a really tiny effort inside of Meta.
所以，Meta 虽然是一家庞大的公司，但这个项目在 Meta 内部的规模其实很小。
You take a huge amount of data for pre-training, and here you're— And it's here, like, roughly, it's like all the, you know, public internet and like all the books you can get your hands on and just like as much data as possible. But different types.
你会使用海量数据进行预训练。大致来说，这些数据包括所有公开的互联网数据、你能找到的所有书籍，以及尽可能多的其他数据，但数据必须涵盖不同类型。
So if you want it to be good at perception, then you want to make sure it has a lot of images and audio and video and this kind of stuff.
所以，如果希望它擅长感知，就必须确保它接触过大量图像、音频、视频等内容。
Um, you know, if you want it to be good at certain domains, you don't just want to RL that in at the end.
如果希望它在某些领域表现出色，就不能只依靠最后的强化学习将这些能力灌输给它。
When you talk about this being the best personal agent model, was there data you had to get for the pre-train where it's like huge amounts of restaurant reservation, call booking?
当你说这是最好的个人智能体模型时，你们在预训练阶段是否必须获取某些数据，例如海量的餐厅预订、电话预约等数据？
You just kind of want to dump in large volumes of the modalities that people are going to care about.
你需要大量输入人们会关心的各种模态的数据。
I mean, I don't actually think we talked about this when we were talking about where Muse Wonder excels.
此前讨论 Muse Wonder 擅长哪些方面时，我们似乎确实没有谈到这一点。
But some of the areas that it's like, all right, what do people care about, right?
但它擅长的一些领域，实际上正是人们所关心的领域。
It's, I mean, there are things like shopping that people care about a lot, personal advice that people care about a lot.
例如，人们非常关心购物，也非常重视个人建议。
Yeah, and so those types of things, I mean, you can RL them to some degree, but you're like, just want to put in a large amount of data.
因此，这类能力在一定程度上可以通过强化学习来训练，但你仍然会希望直接输入大量数据。
Like, you can imagine health is a very— there's a lot of data out there about health.
以健康领域为例，外界有大量关于健康的数据。
And, um, in, in pre-training, you just wanted to like see all these different things.
在预训练中，你会希望让它接触所有这些不同的内容。
So, so that— so as, um, as Mark was mentioning, like, the, uh, we, uh, one of the highlights of Muse Wonder actually is its ability on domain various domain expertises. Yes.
因此，正如 Mark 刚才提到的，Muse Wonder 的一大亮点是它在各个领域具备的专业能力。
And I think our philosophy here is that, you know, the best personal agent should feel like a team of experts in your corner.
我认为，我们的理念是：最好的个人智能体应该让你感觉身边有一支专家团队在支持你。
And so, um, on a bunch of these domains, we, you know, Muse Wonder is actually state of the art on, I think, um, a lot of the finance evals, as well as a lot of the, um, uh, legal evals.
在其中许多领域，Muse Wonder 实际上已经达到最先进水平，包括多项金融评测和法律评测。
And that, you know, these, these domain areas, that starts in pre-training.
这些领域的能力是从预训练阶段开始培养的。
Those, you know, we look for a lot of domain data to put into pre-training.
针对这些领域，我们会寻找大量领域数据，并将其加入预训练。
But a lot of the, a lot of the art of pre-training— I mean, pre-training is, um, there's this one Rune tweet where it's like pre-training is full of the researchers who like chew glass and stare out into the distance and into the, into the abyss or whatnot.
预训练包含许多精妙之处。Rune 曾发过一条推文，称预训练团队的研究人员都喜欢嚼玻璃、凝视远方和深渊。
But, but I think there's a truth to that because it is like, um, It is both the highest stakes of the research areas because you, you know, we invested a huge percentage of our compute into training Muse Wonder, into pre-training Muse Wonder.
我认为这番话确实有道理，因为预训练是风险最高的研究领域之一。我们将很大一部分算力投入了 Muse Wonder 的训练，尤其是 Muse Wonder 的预训练。
But yeah, troughs of uncertainties, you know, all the 7 stages of grief, you know, happen while you're pre-training a model.
不过，我们也经历过一阵阵不确定，甚至完整走过了悲伤的 7 个阶段。这些情况都会在预训练模型时发生。
And it is the most sort of like fundamentally scientific in an interesting way, because, you know, you just don't have very many samples.
从某种有趣的角度来看，这也是最接近基础科学的工作，因为可供参考的样本非常有限。
So, like, your ability to infer exactly what will happen scientifically really depends on your ability to sort of like, you know, big brain reason about everything that's happening in this process. Okay.
因此，能否从科学上准确推断将会发生什么，实际上取决于你能否充分思考并理解这个过程中发生的一切。
So in the pre-training stage, everyone who's training frontier models basically is trying to get as much data as possible for the pre-train with maybe a bent towards If you want Muse Wonder to excel at legal and finance and tax questions and things like this, then you want to have as much of that specific corpus as possible in the pre-train.
在预训练阶段，每个训练前沿模型的人基本上都会设法为预训练获取尽可能多的数据，同时也可能有所侧重。例如，如果你希望 Muse Wonder 擅长回答法律、金融和税务等方面的问题，就会希望在预训练中加入尽可能多的相关特定语料。
Yeah, you still want as much intentionality about where ultimately you want your model to be, but again, it just, you know, benefits from a lot more data.
你仍然需要尽可能明确自己最终希望模型达到的状态，但更多的数据终究会让模型受益。
Yes, but my point is, like, is there data where it's like, we can't distract the model with that, we can't clutter up its brain attic with knowing that stuff, or is it just more data is better?
是否存在某些会分散模型注意力的数据？我们是否应该避免让这些内容塞满模型的脑内阁楼，还是数据始终越多越好？
No, it's not even that, it's just like, I mean, I think that there are all these questions about, like, okay, well, we put a lot of these tokens in, is there any zero-sumness in a bunch of places there isn't, or it's actually positive transfer, but then like depending on how it's set up, like you can have some negative transfer.
不完全是。我认为这里存在许多类似的问题。例如，当我们加入大量这样的 token 时，是否会产生某种零和效应？很多情况下并不存在零和效应，甚至还会产生正迁移；但根据具体的设置方式，也可能出现负迁移。
But in general, one of the bigger challenges of scaling bigger models is it's not just about how many flops and how much compute.
但总体而言，扩展更大模型时，一个较大的挑战是不能只考虑浮点运算量和算力。
I mean, this is like part of what, when the original kind of scaling laws for pre-training were written, I mean, this is sort of what came out of that is is that.
这也是最初提出预训练领域的那类扩展定律时得出的部分结论。
So if the internet isn't growing that quickly and you want to scale bigger models that can be more intelligent, I think one of the interesting problems is, all right, where's the data coming from? Yes.
如果互联网的数据量增长得没有那么快，而人们又想扩展出更大、更智能的模型，那么一个值得探讨的问题就是：数据从哪里来？
So I think that probably gets to the end of what we're going to talk about on that.
关于这个话题，我们差不多就讨论到这里。
But we could go to the other— Excellent CEO energy over here.
不过，我们可以转向另一个话题——这里真是充满了优秀 CEO 的气场。
But so post-training, so then I think Post-training is kind of— is a very exciting field now, I think, and is an area where I think actually, I think almost on a continual basis, the entire field is learning about the optimal recipes to be able to post-train these models.
谈到后训练，我认为它如今是一个非常令人兴奋的领域。整个行业几乎始终都在探索对这些模型进行后训练的最佳方案。
So that's like a lot of when you take this kind of raw thing and you give it both the question and the answer that you would want it to have, right?
获得这种比较原始的模型后，需要同时向它提供问题和希望它给出的答案。
If like, if you're having a chat conversation, how would you want it to respond in this case?
例如，在聊天场景中，你希望它如何回答？
So when you're training things like personality, the SFT is actually like a very important part for molding it.
因此，在训练个性等特征时，SFT 是塑造模型的一个重要环节。
But then I think most of the advances in the recent years have been in RL.
不过，我认为近年来的大多数进展都出现在 RL 方面。
So I think one mental model that can be helpful is like, And SFT is where you kind of initialize a lot of— SFT is supervised fine-tuning.
一种可能有帮助的思维模型是：SFT 用于完成许多初始化工作。SFT 指监督微调。
Where you initialize a lot of the behaviors that you want the model to be able to exhibit later on.
在这个阶段，需要初始化许多希望模型日后能够展现的行为。
And then, you know, maybe at the end of SFT, it has those behaviors kind of baked in, but it'll only do it like 10% of the time.
SFT 结束时，这些行为可能已经被固化到模型中，但模型可能只有 10% 的时候会表现出这些行为。
And then in reinforcement learning, you really sort of give the model reinforcement in various situations to bring the elicitation of those behaviors dramatically up.
随后进入强化学习阶段，需要在各种不同的情境下向模型提供强化反馈，从而大幅提高这些行为被激发的概率。
And I think that the— yeah, so I think there's a bunch of interesting things that are happening in RL and post-training.
因此，我认为 RL 和后训练领域正在出现许多值得关注的进展。
I think one is that just the scaling of RL and post-training has been one of the clear axes, just spending more compute and gathering more data.
我认为，其中一条明确的路径是扩大 RL 和后训练的规模，即投入更多算力并收集更多数据。
So version 1 to 1.1, 1.2, 1.3, we didn't change anything about pre-training, right?
从 1 版到 1.1、1.2、1.3，我们完全没有改动预训练。
So, um, so you, when you think about the early versions of a model of a new class, they're actually relatively less, like the RL to pre-training ratio is going to be lower for RL than the later versions.
因此，对于一个全新类别模型的早期版本，其强化学习规模相对较小。具体而言，早期版本中强化学习相对于预训练的占比，会低于后续版本。
But for a lot of companies, like the big number is the pre-train and then the—
但对很多公司而言，大版本号代表预训练。
It's true that Muse Spark and Muse Wonder are different pre-trains.
Muse Spark 和 Muse Wonder 的预训练确实不同。
And we decided to kick the can down the road on this question for Muse Wonder by not giving it a number yet.
关于 Muse Wonder 的这个问题，我们决定暂时搁置，暂不给它版本号。
But when we have our first dot release, we will have to decide that.
但在第一次发布小版本更新时，我们就必须做出决定。
But yeah, no, I think the RL stage, part of what has been— part of what's very interesting is first is just producing the environments and producing the tasks.
强化学习阶段有一个非常有意思的方面：首先要构建环境和设计任务。
And I think that is one of the areas where there's a lot of intentionality in training these models.
我认为，这也是训练这些模型时需要投入大量精力进行明确思考的领域之一。
Like if you— in a lot of cases, less is actually more when it comes to data in RL.
例如，在强化学习数据方面，很多时候，少确实胜于多。
And so I think I think the sort of, like, the feel and the science of it, like, is pretty flipped, generally speaking, from pre-training.
因此，总体而言，强化学习所依赖的直觉和科学方法几乎与预训练完全相反。
The other is that I think it can give very magnified kind of instruction to the model, if that makes sense.
此外，我认为强化学习能够以一种高度放大的方式向模型传递指令。
So I think that, like, a lot of safety cases or alignment cases that we care a lot about are, like, you know, and, you know, where you would see when you do large high-compute RL runs on frontier models, are where you apply a lot of RL or optimization pressure in scenarios where the only way for the model to succeed is to do something misaligned or do something unsafe.
因此，在许多我们非常重视的安全或对齐案例中，前沿模型经过高算力、大规模的强化学习训练后，会面临这样的场景：大量强化学习或优化压力被施加于模型，而模型要想成功，唯一的办法就是做出不符合对齐要求或不安全的行为。
And so that's another thing that we had to be very careful about and thoughtful about in training Muse Wonder is ensuring that we curated the RL environments very thoughtfully to ensure that all of them were eliciting and driving towards more aligned behavior from the model.
因此，在训练 Muse Wonder 时，我们必须非常谨慎并进行认真思考，审慎筛选强化学习环境，确保所有环境都能引导并推动模型展现出更符合对齐要求的行为。
So the point that you're saying about quality, I think, is an important one because I think a lot of the previous questions back and forth were about the quantity and have the flavor of like, oh, you need to scale the quantity. Yes.
你刚才提到的质量问题非常重要，因为此前反复讨论的许多问题都集中在数量上，给人的印象是必须扩大数量。这一点没错。
But I mean, I do think like a huge part of this is just, is, is the point that you just made of like the even a smaller amount of high quality.
但其中一个非常重要的方面是，正如你刚才所说，即使高质量数据的数量少一些也可以。
Yeah, I think, I think in RL, like we, you know, as we've seen pretty big jumps in our Muse Spark line, I think these have been like accompanied with, you know, real meaningful improvements to the underlying recipe.
在强化学习方面，从我们新的 Spark 系列中可以看到相当显著的跃升；这些跃升也伴随着底层训练配方真正有意义的改进。
And each of these improvements in the recipe have been like, I think the improvement between Muse Spark 1 to Muse Spark 1.3 feels night and day.
训练配方的每一次改进都带来了显著提升。Muse Spark 1 和 Muse Spark 1.3 之间的提升简直是天壤之别。
I think probably to most people, if you ask them, do these feel like they came from different pretrains?
我认为，如果询问大多数人，这两个模型是否给人一种来自不同预训练的感觉，他们会给出肯定的答案。
But, you know, this sort of like, these two things are happening at the same time, which is also, by the way, for us, this is why we're seeing so much progress in general.
但这两件事正在同时发生。对我们而言，这也是总体上能够取得如此巨大进展的原因。
Like, I think it's the fact that we're scaling both pre-training and scaling our recipes in RL kind of like concurrently. At a nice clip.
我认为，原因正在于我们既扩大了预训练规模，也扩展了强化学习的训练配方，并且同步推进了这两方面，速度也相当可观。
I mean, like making sure that your agent harness is good and that you can co-optimize these things is probably just another aspect of, or dimension for scaling.
例如，需要确保智能体框架足够完善，并且能够对这些方面进行协同优化。这大概只是规模扩展的另一个方面，或者说另一个维度。
But if I'm understanding you guys correctly, are you saying that alignment is not about coming up with some alignment-specific RL steps and instead about making sure the entire RL process is kind of steering The model—
但如果我没有误解你们的意思，你们是说，对齐并不是设计一些专门用于对齐的强化学习步骤，而是要确保整个强化学习过程都以某种方式引导模型——
I mean, a lot of our theory is like, okay, we care more about it being aligned to, you know, we're trying to give people agents, right?
我们的许多理念都建立在这一点上：我们更在意它与谁对齐，因为我们正努力为人们提供智能体。
And it needs to, if you ask it to do something, you need to have a high confidence that it's going to do what you ask it to do.
如果你要求它做某件事，就必须非常确信它会按照你的要求去做。
You don't want a moralizing lecture for why you're wrong for asking it to do the thing.
你不会希望它反过来向你进行道德说教，告诉你为什么不该让它做这件事。
Obviously, if you're trying to do something that's catastrophic, it needs to gracefully, without pissing you off, like, not do it.
当然，如果你想做的事情会造成灾难性后果，它就不应执行，但需要委婉处理，不要激怒你。
But in general, you know, we view alignment as more of a question of instruction and values alignment with the person who it's intending to serve.
但总体而言，我们更多地将对齐视为指令对齐和价值观对齐的问题，对齐的对象是它打算服务的人。
But one of the more interesting things I think that is maybe a question related to your original question is also just how long the process of training one of these models takes.
不过，另一个更有意思、或许也与你最初的问题有关的方面，是训练这类模型究竟需要多长时间。
Like Muse Wonder, I mean, in general, we've planned Muse Wonder since last You know, August.
以 Muse Wonder 为例，总体而言，我们从去年 8 月就开始规划 Muse Wonder 了。
Yeah, so yeah, I remember the offsite when we were— so this is like when they first presented this.
对，我还记得那次外出会议，那是他们第一次展示这个项目。
Yeah, so I think all of the stages that we just talked about are like, you know, you really work on them for quite a long time.
我认为，我们刚才谈到的所有阶段都需要投入相当长的时间。
And, and, um, it's not that we, we didn't start working on Muse Wonder when we finished Muse Spark.
我们并不是等到 Muse Spark 完成之后，才开始开发 Muse Wonder。
It's that we've actually just been planning and working on this for quite a long time.
实际上，我们已经规划并推进这件事相当长一段时间了。
And we really view Muse Spark, if anything, as like I think we mentioned this in the first blog post where we wrote about it.
我们确实是这样看待 Muse Spark 的——我记得，我们在第一篇介绍它的博客文章中提到过这一点：
Yeah, and in terms of how long it takes, I mean, we're also even just talking about the research part.
对。至于这需要多长时间，我们现在甚至还只是在讨论研究部分。
I mean, there's also the compute part where we've been investing in building up, first of all, a huge amount of compute overall.
此外还有算力部分。我们一直在投入建设，首先要建设规模极其庞大的整体算力。
But a lot of, you know, part of our philosophy on this is that you can kind of measure the compute strength of a lab by how big its largest single computer is, supercomputer, for being able to do these large training runs.
不过，我们在这件事上的一个重要理念是：一家实验室的算力实力，大致可以通过其规模最大的单台计算机来衡量，也就是用于开展这些大型训练任务的超级计算机有多大。
So, I mean, we built this Prometheus cluster, we call it, in Ohio.
我们在俄亥俄州建设了一个名为 Prometheus 的集群。
And now we're building this Hyperion cluster in Louisiana that's going to scale to 5 gigawatts.
现在，我们正在路易斯安那州建设 Hyperion 集群，其规模将扩大到 5 吉瓦。
Because like building a whole like data center site from scratch, at least the plans for those, they're coming online now.
从零开始建设一整个数据中心园区需要很长时间。这些项目至少很早就已开始规划，现在才陆续上线。
But if you want to go really end-to-end on doing this stuff, then, I mean, obviously the kind of full stack, and I mean, Meta is a full-stack tech company.
但如果你真想把这些东西端到端地做下来，那，我是说，显然就要覆盖整个技术栈，而 Meta 本身就是一家全栈科技公司。
I mean, we build our own data centers, we build our own chips, we build our own low-level infrastructure.
我是说，我们自己建数据中心，自己造芯片，也自己搭建底层基础设施。
And now this is kind of the next component, which, I mean, arguably is probably the most important component in all of this for the next wave of technology is kind of building really good models and then delivering them to people.
而现在，这是接下来的一个组成部分，我是说，可以说，这很可能是下一波技术浪潮中最重要的组成部分，也就是打造真正优秀的模型，然后把它们交到人们手中。
But I mean, yeah, if you want to go fully end-to-end, I mean, this is a long process.
不过，我是说，对，如果你想实现完全的端到端，我是说，这是一个漫长的过程。
I do think this is like a— I mean, maybe it's well appreciated by folks in the tech industry, but It really is, you know, it feels like models are coming out constantly.
我确实觉得这就像是——我是说，也许大家都很清楚这一点，尤其是科技行业的人，但实际上，你知道，感觉模型一直在不断推出。
So maybe you'd have the mental model that like, oh, a company like launches a model and then they figure out how they're going to make the next model and, or something like that.
所以你可能会觉得，哦，一家公司先推出一个模型，然后再琢磨下一个模型要怎么做，诸如此类。
But it's actually these like very long planning processes and these sort of like very long work streams that kind of all come together at the very end and then result in these beautiful models.
但实际上，这些规划流程都非常漫长，各条工作线也会推进很久，直到最后关头才全部汇聚到一起，最终造就这些出色的模型。
Like ultimately you start planning the model in August and you don't know until more than a year later. if it's good.
比如，你从8月开始规划模型，却要等到一年多以后才知道它好不好。
And, you know, Silicon Valley famously likes to run on metrics, but, like, ultimately, do you just need someone who is the aesthete and says, I'm sorry, I don't like it?
而且，你知道，硅谷向来喜欢用指标说话，但说到底，是不是还得有那么一个有审美的人站出来说：“抱歉，我不喜欢”？
It's nice when you have metrics, but I think the, like, state of being an entrepreneur is, like, how well do you make decisions in an information-poor environment, right? So—
有指标当然很好，但我觉得，创业的本质就在于，你能否在信息匮乏的环境中做出正确的决策，对吧？所以——
Yeah, I think that this is— and I think Mark gets a huge amount of credit for this, like, Because of how long everything takes, like, there were many points along over the past year since I've been here, but also, you know, as you mentioned, like, long before that, where you have to make very large investments or make very big decisions about, you know, how you're going to invest or develop the technology or, you know, what decisions you're going to make
对，我觉得这——而且我觉得 Mark 从中获得了非常多的过去一年里，自从我来到这里后，有很多时候，当然还有，你知道，就像你提到的，在那之前很久，也有很多时候，你必须进行巨额投资，或者做出非常重大的决策，比如要如何投资或开发这项技术，或者，你知道，要做出什么样的决定，而此时距离出现任何证据来证明这东西到底能不能奏效，还早得很。
long before there's proof points around whether or not even the thing is working.
所以我觉得，从某种意义上说，这就像是
And so I think this is like, in some sense, it's like entrepreneurship on hard mode.
高难度模式下的创业。它就像是精益创业的反面。
Like, Lean Startup is like, yeah, you like see if you get something working and then you double down and then you see if that works, then you double down.
然后加大投入，再看能不能奏效，如果奏效就继续加码。但在人工智能领域，你根本没法这么做。
You have to like we've— I mean, we've been investing into these data centers since, you know, for years and years and years.
你得——我们已经——我是说，我们一直在这些数据上投入我们不得不——
And you'll be radically behind if you wait to see if things are working.
所以我们会聊到这些，比如去年年底，我们就说，
So we have these conversations where it's like, At the end of last year, we're like, all right, we are quite excited about the progress.
我们对目前的进展相当兴奋，尽管还没有看到模型。
Like, it's like, I'm going to, like, you know, we're going to write this check for, like, I don't know, like $100 billion, $200 billion.
大概价值$100 billion或$200 billion的支票。这件事实在太夸张了。
It's like, um, yeah, so it was, it was so extreme that, you know, we had, we had like the whole exec team.
我们把整个高管团队都召集了过来。我觉得 Susan 把你吓得不轻。
Yeah, well, yeah, no, well, it was like The joke every time we talked about it would just be like, yeah, and we hope it's all going to work and be good.
对，希望一切都能顺利，结果也能不错。about it would just be like, yeah, and we hope it's all going to work and be good.
But like, there were literally, like, at the end of last year, other than me giving a vibe report on how our research was going and, you know, I could sort of show some plots and whatnot, but it was like, I mean, I was feeling good about it, but it was certainly not clear at all that it was going to work.
但说实话，去年年底时，除了根据感觉谈谈我们的研究进展，以及展示一些图表之外，我也无法提供更多依据。当时我认为情况不错，但完全不能确定最终能否成功。
You guys are saying that the essence of training a frontier model is you have to push your chips all in before you've been dealt the cards.
你们的意思是，训练前沿模型的本质，就是必须在牌还没发到手上之前押上全部筹码。
You just have to decide that you're going to develop a frontier model and you don't get to meter the inputs. At this scale.
你只能先下定决心开发一个前沿模型，无法逐步控制投入。在这种规模下，这是做不到的。
I mean, if we, like, back a few years ago, the clusters didn't have to be this big, but now they do.
如果回到几年前，集群还不需要这么大，但现在必须达到这样的规模。
So I think this is part of the excitement of booting up a frontier lab now, right?
我认为，这正是当前从零搭建一家前沿实验室令人兴奋的地方。
And yeah, Susan pulled me aside many times and was like, you know, I just want to tell you, if this doesn't work, like, you know, this is going to be really, really bad.
Susan 好几次把我拉到一边，对我说：“我想让你明白，如果这件事没有成功，后果会非常、非常糟糕。”
Okay, so it's roughly a $15 billion effort to train a bunch of models.
好，所以训练一批模型大致需要投入 $15 billion。
If you're talking about building the CapEx, it's probably closer to, you know, I mean, you pay for it upfront, so it's probably, you know, $50 or $60 billion.
如果说的是建设所需的 CapEx，因为这笔钱需要预先投入，可能更接近 $50 或 $60 billion。
So, but then that's just kind of, one cluster for one part of training.
不过，这还只是一个集群，只负责训练中的一个环节。
It's like, that's like probably an equal amount of it is because for each change that you make, you're testing all the ladders all the way up to the big, the big scale.
此外还有规模大致相当的工作量，因为每做一项改动，都必须在所有规模阶梯上逐级测试，直至大型和超大规模。
Yeah, definitely the whole, I mean, I've seen the financial rollups, tens of billions of dollars is a good estimate for—
对，确实如此。我看过汇总财务数据，数百亿美元是一个相当可靠的估计——
So I actually think like some of the more complicated questions that we had were like, all right, so we're training this model and we know, we know we want to train the model cuz we, this is like key, but like how much product demand do we think we're going to have?
所以我其实认为，有些更复杂的问题在于：我们正在训练这个模型，也知道自己想要训练它，因为这很关键，但问题是，我们认为产品需求会有多大？
Um, like, and as in you're trying to build data centers to serve Muse for a product that doesn't exist yet. Yeah.
也就是说，你需要建设数据中心来为 Muse 提供服务，但这个产品目前根本不存在。对。
So this is probably where, like, when we were talking about it last year, we were just like, huh, there was a moment where I think a bunch of people were like, yeah, maybe, maybe we should just like be a little more conservative.
所以，情况可能就是这样。去年我们讨论这件事时，曾有那么一刻，我觉得很多人都在说，也许我们应该稍微保守一些。
And But it was like, I mean, even, I mean, I too am an entrepreneur and founder and have, you know, at various points in my life taken risk.
不过在当时……即使是我，也是一名创业者和创始人，并且在人生的不同阶段冒过险。
I remember the visceral feeling at the time was like, this is a fuck ton of risk.
我记得当时最切身的感受是：这他妈风险也太大了。
Yeah, yeah, it's only— well, I do think this is like one of the like things that I think, Mark, is very impressive about you, which is I think you have really internalized this concept of You know, it's about how big it is when you win, not like, you know, failure is okay.
我确实认为，Mark，这是你身上非常令我佩服的一点：你已经真正把这个理念内化了。关键在于成功时能取得多大的成果，而不是……失败并不要紧。
Mark has fully internalized power laws in a way that many people have not.
Mark 已经彻底内化了幂律，而很多人还没有。
One is if AI works overall, plenty of other people want all our compute.
第一，只要人工智能整体上行得通，就会有很多其他人想要我们的全部算力。
So they're willing to pay, like, I mean, if like the cost is whatever, $15 billion a gigawatt, we could get like a lot more for that.
因此，他们愿意付钱。例如，如果成本是每吉瓦 $15 billion，我们可以卖出高得多的价钱。
So I think that you, in order for it not to work out, it's not just that we have to fail, AI overall has to fail.
因此，要让这件事最终无法成功，不能只是我们失败，而必须是人工智能整体都失败。
And then even if that happens, we still have all these other uses to productively use compute.
即使真的发生了这种情况，我们仍然有其他各种用途，可以有效利用这些算力。
So in the grand scheme of things, we were definitely building out this compute for inference ahead of knowing that the product was going to be ready.
因此，从大局来看，我们确实是在尚不确定产品能否准备就绪时，就提前建设了这些用于推理的算力。
But you were AI-pilled enough that you said GPUs online and powered will be useful to someone.
但你当时已经足够笃信人工智能，相信只要 GPU 上线并通上电，总会有人用得上。
I mean, it's like, I think, and you can look at the kind of the trajectory where it's like, all right, first there's ChatGPT as the text-based assistant.
从整个发展轨迹来看，最先出现的是 ChatGPT，一个基于文本的助手。
And it just seemed like this is the thing that Meta should build in the world.
而且，这似乎正是 Meta 应该为这个世界打造的产品。
And I think that that was just like pretty clear to you and me and Nat, like from the beginning, right?
我认为，从一开始，你、我和 Nat 都非常清楚这一点。
I mean, there's always like the real question for me on the Muse launch even was you never know when you launch something how it's going to be received.
对我而言，即使在 Muse 发布时，真正的问题也始终在于：产品发布后会得到怎样的反响，永远无法预知。
So there's always the case where you get a home run, which in this case it was, right? Yes.
有时产品会一炮而红，这次就是如此。
But the normal case would be you kind of— it's fine, and you have to like iterate a few times to get product-market fit before you really grow.
但通常情况下，产品发布后的表现只能算尚可，需要经过几轮反复迭代，真正找到产品与市场的契合点后，才能实现大规模增长。
In some ways, like, I actually think the product, we maybe could have even grown it faster, but we didn't have all the growth stuff really ready to go when we shipped it because we are like— our median expectation was like, all right, we're probably going to have to like iterate on this a bit, and like, that'll be That'll be fine.
从某种角度看，我认为这个产品原本甚至可以增长得更快。但发布时，我们尚未真正做好增长方面的全部准备，因为当时预期的中位情形是：产品可能还需要再迭代一次，不过这并无大碍。这样的结果也很好。
But I don't know, I think like if it takes a few iterations, it's like not the end of the world. That's normal.
我认为，即使产品需要经过几轮迭代，也不是什么严重的问题。这很正常。
Um, you can have launches where it's like you get like, it's kind of catastrophic and then it takes a while to recover, but that's not good.
也有些产品在发布时会遭遇近乎灾难性的失败，随后需要很长时间才能恢复，但那绝非好事。
But, but it's like, but I mean, even if it took a few iterations, it would have been fine.
不过，即使这个产品需要经过几轮迭代，也没有什么问题。
So I don't know, I guess it just seemed to me like this was going to be— yeah, the numbers were big, but I— it seemed like it was probably—
因此，我认为这件事应该会……数字确实很大，但我感觉它大概会……
Well, one thing that I think is quite notable about Meta is the degree to which, like, first of all, most of the exec team has been there quite a long time.
我认为 Meta 有一点很值得注意：大部分高管团队成员都已经在那里工作了很长时间。
And And, you know, most people have been through crazy situations many times in the past as well. Mm-hmm.
而且，大多数人过去也经历过多次疯狂的局面。
And I think Meta as a company is like an extremely resilient company that has been through like, you know, many situations that also at the time would've seemed incredibly daunting.
我认为，Meta 是一家韧性极强的公司，经历过许多在当时看来同样无比棘手的局面。
And so everyone's kind of like knows the program, so to speak, in terms of, you know, this obviously on paper, and especially when like Susan Li, our CFO, is sort of like pulling me aside in the meetings, like, does, you know, does seem like a particularly scary situation.
因此可以说，每个人都知道该如何应对。显然，从纸面上看，尤其是在当时，这确实是个特别吓人的局面。我们的首席财务官 Susan Li 在会议上把我拉到一旁，问起了这件事。
But I think most people were like, yeah, yeah, no, but we're Meta, we're going to get through it.
但我想，大多数人都会说：“没错，不过我们可是 Meta，我们会挺过去的。”
Like, I think there was like a real camaraderie around that and sort of like quite a nice energy around that, which really I did appreciate because I think it would be very easy for, you know, it's the other executives at the company who are operating and running— Who are paying the bills.
我认为，当时大家确实有一种同舟共济的感觉，整体氛围相当不错。我对此非常感激，因为公司里其他负责运营和管理的高管也在为此买单。
And I think, like, you know, taking a step back, I think one of Mark's jobs as, like, the founder and CEO of the company is to ensure that the company is always pushing towards the most ambitious and expansive version of what it can be.
退一步看，作为公司的创始人兼首席执行官，Mark 的职责之一，是确保公司始终朝着它所能达到的最宏大、最广阔的形态迈进。
And I think that one of the great things about this entire MSL and AI push is that I think this is one of the rawest expressions of that is we knew that AI certainly is a very important technology and it's one of the bigger technologies right now.
我认为，整个 MSL 和人工智能战略的一大亮点，就是它最直接地体现了这一点：我们知道，人工智能无疑是一项非常重要的技术，也是当下最重大的技术之一。
But if we were able to build it effectively and properly integrate it into everything that we do, as well as like, you know, as a full-stack technology company, like build it into, you know, these incredible products that we know will exist in the future, this could be like immensely expansive for Meta.
但如果我们能够有效地打造人工智能，将其恰当地融入我们所做的一切，并作为一家全栈技术公司，把它融入那些我们确信未来必然会出现的出色产品中，就可能极大地拓展 Meta 的边界。
I think it's one of like probably the most exciting products to just like You know, when like late at night and you just think about all the ways it can go.
我认为，它大概是最令人兴奋的产品之一。每当深夜想到它所有可能的发展方向时，这种感受尤其强烈。
Like, I think Nat, Mark, and I just like, you know, shoot across messages late at night of just like, oh, we could do this and we could do this and, you know, we can make this move and we can like build this product.
Nat、Mark 和我经常在深夜互相发消息，讨论我们可以做这个、做那个、迈出这一步，或者打造这款产品。
And I think it's just like, you know, very— It's a very exciting time in technology right now.
这一切都令人振奋。现在确实是科技界一个无比令人兴奋的时代。
I'm struck by, as we talk about Muse and just how it clearly feels like the next platform, an agent operating on your behalf.
我们谈到这些时，有一点让我印象很深：Muse 显然让人觉得它就像下一代平台，是一个代表你行事的智能体。
You've talked for a long time about not wanting to be reliant on other platforms and how kind of the experience of building on Android and iOS kind of burned you on being dependent on other.
长期以来，你一直表示不希望依赖其他平台，也谈过在 Android 和 iOS 上开发的经历。在某种程度上，那些经历让你吃够了依赖其他平台的苦头。
And often when you would talk about, say, Reality Labs, is the importance of owning.
此外，你以前谈到 Reality Labs 时，也经常强调掌握——
the platform of the future and kind of many years working on Reality Labs with the intent of owning the platform of the future.
这关系到未来平台的重要性。多年来，我们一直致力于 Reality Labs，目的就是掌握未来的平台。
Yeah, no, I mean, it's less about wanting to own a platform for its own sake.
这并不是为了拥有平台而拥有平台。
And at some level, you know, if you look at like other social apps, for example, I think part of the reason why Meta is so much better is because we don't just build the app, right?
从某种程度上说，纵观其他社交应用，我认为 Meta 之所以优秀得多，部分原因在于我们并不只是开发应用。
It's like we're like a deep technology company and we do this because we are world-class at machine learning and like all the recommendation systems are operating at a scale that basically few others are in the world.
我们更像是一家深度技术公司。之所以如此，是因为我们在机器学习领域处于世界一流水平，而这些推荐系统的运行规模，也是全球鲜有其他公司能够达到的。
I think it's basically like Just us and ByteDance on that, and and then like okay we just kind of go all the way to the metal.
我认为在这方面，基本上只有我们和 ByteDance，而且我们一直深入到技术栈的最底层。
And to me, being able to also deliver go the last mile through the device, it does come from a while of having these experiences where we couldn't build the experiences that we wanted on phones.
对我而言，我们还必须能够通过设备完成最后一公里的交付。这种想法确实源于过去一段时间的经历：我们无法在手机上打造出自己想要的体验。
But I just think for the future, the the models are going to be by far the most important part of delivering the technology experience.
但我认为，面向未来，模型将远远成为提供技术体验最重要的部分。
And I mean, this is something that I think we kind of bonded on and agreed on, like, very early, is this notion that, like, in every generation of technology companies, there's, like, no, there's usually not one company that kind of is the— is, like, wins everything.
我认为，这也是我们很早就产生共鸣并达成共识的一点：在每一代科技公司中，通常都不会只有一家公司赢得一切。
But with the internet, there were, like, okay, 5 or 6 very big companies and then a whole ecosystem of other companies that are great.
在互联网时代，大约有 5 or 6 家规模非常大的公司，此外还有一个由其他优秀公司组成的完整生态系统。
The companies that did the best kind of had the core technology to be able to serve people the best there.
那些表现最好的公司往往掌握着核心技术，因此能够为人们提供最好的服务。
And I think in this next era, there are going to probably be a small handful of companies that have the ability to build frontier superintelligence.
我认为在下一个时代，很可能只有少数几家公司有能力打造前沿超级智能。
And I think that those companies are going to be the ones that effectively drive the agenda and can build most of the interesting, the most interesting experiences for consumers, for enterprises, hardware, these different kinds of things.
我认为，这些公司将真正主导未来的发展方向，并为消费者和企业打造大多数有意思的体验、最有意思的体验、硬件及其他各种产品。
So, and I just look at it as, okay, like, is it even really a choice, right?
因此，在我看来，这其实已经算不上是一种选择。
It's like, yeah, okay, Meta could, will probably exist into the future regardless of whether we do this.
即使我们不这样做，Meta 未来很可能仍会继续存在。
But if we want to be kind of an interesting company that is like inventing new things that matter to people in the future, I don't see how you can do that if you don't have the ability to build end-to-end your own superintelligence and build them into the products that people use.
但如果我们希望成为一家有意思的公司，能够为未来的人们发明真正重要的新事物，那么就必须具备端到端打造自己的超级智能，并将其融入人们所用产品的能力。否则，我看不出还能如何实现这一目标。
And I think that this is something that you just like believed from very early on, Is basically, it's like, all right, there are going to be like a handful of these.
我认为，这是你从非常早期起就坚信的一点：未来将会出现少数几个这样的超级智能。
Like, it's like there's a kind of personal and consumer-oriented version of this.
其中将有一种偏向个人和消费者的版本。
And like, this lane is open because everyone else seems to just be obsessed so much with just automating knowledge work, which, I mean, isn't even that charismatic version of the future, right?
这条赛道目前仍然空缺，因为其他所有人似乎都只专注于将知识工作自动化，认为仅凭这一点就能实现许多事情。然而，这甚至算不上多么有吸引力的未来图景。
Because it's like, all right, like, a big part of our belief too is that the kind of highest use of superintelligence is invention and creation, not automation.
我们的一个重要理念是，超级智能最具价值的用途在于发明和创造，而不是自动化。
And so, I don't know, it just seemed like there's like a big place in there, but you kind of need to own the stack.
因此，我也无法确定，只是认为这里存在很大的发展空间，但需要掌控几乎整个技术栈。
How is the Muse-mediated future going to shake out with businesses need to— okay, businesses variously need to build support for for Muse, where it might try to computer use its way in, but that might be janky in certain ways.
由 Muse 介导的未来最终将如何发展尚不明确。不同企业都需要构建相应的支持体系来支持 Muse；否则，Muse 可能会尝试通过计算机操作接入，但这种方式在某些方面可能很不顺畅。
So you kind of want everyone to have first-class support for agents talking to things. Yeah.
因此，你会希望所有人都能提供原生支持，让智能体与各种服务交互。
Then the bigger question is they might actively try to block agents from interacting with them.
更大的问题是，这些企业可能会主动阻止智能体与其交互。
And this might be because they have a misguided sense where, like, it's actually in their interest to support agents, but they don't want to.
这可能源于一种错误认识：支持智能体其实符合这些企业的利益，但它们仍不愿意这样做。
You can imagine where maybe they have a very thin moat and it all comes from a business that only succeeds by consumers spending time on the site or only succeeds by overcharging consumers, then they should rationally not want to allow agents because the agents are like, why?
可以想象，有些企业的护城河可能很浅，其业务之所以能够成功，完全是因为消费者会花时间停留在网站上，或者因为它们向消费者收取过高的费用。从理性角度来看，这些企业不会愿意允许智能体接入，因为智能体会质疑其合理性。
Well, I think that it starts off with— I mean, we trained into the models very good browser use, right?
我认为，在初期，我们已经将模型训练得非常善于使用浏览器。
So that way we want to make it so that no one No one has to do anything to integrate with Muse.
这样一来，我们希望任何人都无须进行额外操作，就能与 Muse 集成。
And then, but then of course, for— we're kind of just going down the list of the most important partners, and for people who want to work with us, we'll just get direct API access, right?
目前，我们基本上按照合作伙伴的重要程度逐一推进。对于愿意与我们合作的伙伴，我们会直接获得 API 访问权限。
And then it's like, okay, then it ends up being way more efficient and faster for people to use Muse.
这样一来，用户使用 Muse 时，效率会高得多，速度也会快得多。
It's more efficient because now you don't have an agent crawling your website.
效率之所以更高，是因为智能体不再需要抓取企业的网站。
Um, and then I think the fullness of time is that, you know, just like today every business has, um, a website and social media accounts and email addresses and all these things, every business is going to have agents too.
时机成熟后，情况将会变成这样：正如如今每家企业都有网站、社交媒体账号、电子邮件地址等一样，每家企业也都会拥有自己的智能体。
And, um, and so of course Muse agents will talk to other types of agents, but they'll work really well with our agents, right?
Muse 智能体会与其他类型的智能体交互，但它们与我们的智能体能够很好地配合。
But we have, we have way more businesses that are calling us asking to integrate deeply, and very few really who have tried to block it.
但确实有多得多的企业主动联系我们，要求进行深度集成，真正试图阻止这件事的企业少之又少。
And my guess is that the ones that try to block it, it will end up not being the right strategic thing for them.
我猜，那些试图阻止它的企业，最终会发现这并不是适合它们的战略选择。
So in the long arc, you'd clearly expect that demand will shift from current avenues more and more towards agentic demand.
所以从长远来看，显然可以预见，需求会越来越多地从现有渠道转向智能体驱动的需求。
And, you know, there always will be demand that lives elsewhere, just in the same way that, like, there's still brick-and-mortar demand in commerce.
当然，需求总会存在于其他地方，就像商业领域至今仍有实体店需求一样。
And kind of in the same way, like, I think agentic commerce demand will grow as a portion of the pie.
同样，我认为智能体商业需求在整体中所占的份额会不断增长。
And I think there's sort of these questions for businesses in the long run, like, do they want to participate in that slice or that portion of the economy over time?
我觉得，从长远来看，企业需要考虑这些问题，比如，它们是否想参与未来经济中的这一板块、这一部分？
I think it's, like, very, very exciting to imagine what will happen in this agentic commerce future. Yes.
想象一下在未来的智能体商业中会发生什么，我觉得真的非常、非常令人兴奋。是的。
It's like they're buried amongst— I noticed that The Brothers Karamazov was quite a good literary analysis.
就像那些内容都埋在——我注意到那篇关于《卡拉马佐夫兄弟》的文学分析相当不错。
But yeah, no, I think clearly this is a discontinuous event for how commerce will happen, broadly speaking, or just generally speaking, how the entire how all the business pipes will work, maybe even in a broader sense.
不过，是的，我觉得很明显，这会给商业运作方式带来一次非连续性的变革，广义上说是商业运作方式，或者更笼统地说，就是所有商业管道将如何运作，甚至可以从更宏观的角度来看。
But I think that, you know, you just go back to, this is clearly a lot better for the consumers.
但我觉得，归根结底，这显然对消费者好得多。
One of the exciting things, I'm curious what you guys see too, but I think Muse is one of the first products where you're actually people are fully delegating purchases to the agents.
有一点很令人兴奋，我也很好奇你们看到了什么，但我觉得 Muse 是最早让人们真正把购买任务完全交给智能体的产品之一。
Like, I think, you know, we've been talking a lot about agentic shopping and like the possibility of it, but now it's like really happening.
我们一直在广泛讨论智能体购物及其可能性，而现在它真的开始发生了。
And I think the agents are finding them, you know, in my case, like great teas, or, you know, in whomever— whoever's case, like great products for them to purchase, which is quite exciting.
智能体能够为每个人找到适合购买的优质产品；例如对我而言，它能找到很棒的茶。这相当令人兴奋。
And so I think that there's There's clearly an exciting thing happening here.
显然，这里正在发生一件令人兴奋的事。
I think, I think part of the value prop— well, one, one part of the value prop for the user is just reduced friction.
这项价值主张的一部分，也就是它为用户提供的一个价值点，在于减少麻烦。
Like, hey, I don't feel like click on all these websites and do all this research to try to find the exact products.
例如，我不想为了找到完全符合要求的产品，逐个打开这些网站并做大量研究。
And so I think the like old way of like searching for stuff just feels so obviously primitive now, now that you can do this.
现在既然已经能够采用这种方式，过去搜索商品的方法显然显得过于原始。
Then there's this sort of like back part of the funnel, so to speak, but just like the ease of friction reduction of just, yeah, filling out the form, just like getting it ordered is quite nice.
此外，还有所谓的漏斗后端环节。减少操作阻力所带来的便利，例如填写表单和下单，确实很实用。
And I think the long-term thing that we're quite excited about is greater personalization.
从长期来看，我们非常期待实现更强的个性化。
Ultimately, I think one of the exciting things about Muse is that it has sort of like this holistic context into you as a person and all of the aspects of your life.
归根结底，Muse 令人兴奋的一点，在于它能够从整体上了解你这个人以及你生活的方方面面。
So it sort of like can holistically understand things about your family or things about your coworkers or things about your friends.
因此，它能够从整体上理解与你的家人、同事和朋友有关的信息。
And then sort of like help you co-optimize across all of those to get you the exact right gift for your, you know, best friend in the context of your friend group based on what it knows about your friend group.
然后，它会综合协调所有这些信息，为你选出最适合送给挚友的礼物，同时结合你们朋友圈的情况，因为它了解这个朋友圈。
Like, I think there's just like a lot of exciting opportunities that are somewhat new and novel when it comes to like these personalizations, these recommendations.
在个性化和推荐方面，这里存在许多令人兴奋的机会，其中一些相当新颖独特。
Because historically, a lot of these recommendation systems are quite siloed and quite isolated. Mm-hmm.
过去，许多此类推荐系统彼此分割，而且相当孤立。
And so they're sort of like, they don't see like a holistic view of you as a person.
因此，它们无法从整体上了解你这个人。
And Muse has an opportunity to offer much greater levels of personalization.
而 Muse 有机会提供更高程度的个性化。
Yeah, I think people are already seeing that, where if they ask, like with Muse, if you ask it to buy something for you, quite frequently it'll add in some bit of context that it pulled in from a different conversation.
大家已经开始看到这一点。例如，当用户让 Muse 帮忙购买东西时，它经常会加入一些从其他对话中获取的背景信息。
It's like I'm normally into smoking meats, but I'll do some baking in order to To make, to make Aurelia happy.
我平时更喜欢熏肉，但为了让 Aurelia 开心，我也会做些烘焙。
And so we've gotten into this really cute routine where, um, you know, we, we, I've been playing with Muse voice mode, which isn't fully rolled out yet, but we're, we're playing with it.
我们因此养成了一个特别温馨的小习惯。我们一直在试用 Muse 尚未全面推出的语音模式。
So it knows which one is me and it knows which one is not me based on that.
它知道哪个声音属于我，也能据此判断哪些声音不属于我。
And we just like, you know, throughout the week we're talking about what we'll want to bake that weekend. Mm-hmm.
我们一整周都会讨论周末想烤些什么。
And she'll like give it ideas or like she'll kind of come up with crazy things that are not real recipes, but like bounce it off with Muse.
她会给 Muse 提供一些点子，或者想出一些根本算不上食谱的疯狂创意，再和 Muse 一起琢磨。
And then Muse will just make sure that we, every Saturday when it's time to bake, I get the ingredients from Instacart and they kind of show up and they're on time, which I mean, the biggest thing is like if I were just starting the activity on Saturday, you'd have to get everything and it would take some time.
Muse 会确保每周六开始烘焙时，我从 Instacart 购买的食材都已送到，而且基本都很准时。最重要的是，如果等到周六才开始准备这项活动，我就得先买齐所有东西，这会花费不少时间。
But now everything's ready whenever we're going to go sit down and do it.
但现在，当我们准备坐下来动手时，所有东西都已经备好了。
This weekend it knew that Wonder was launching, so it threw a watermelon in the bag.
这个周末，它知道 Wonder 要发布了，所以推荐了一个西瓜。
It deduced that the recipe did not have enough— there was not enough, I guess, expense in it to get the delivery, so it needed something else.
它判断这份食谱所需食材的消费金额没有达到配送门槛，因此还需要添加一些别的东西。
And it was like, you're going to— And it topped it up with a watermelon.
于是，它询问我要不要再加点东西，随后用一个西瓜凑足了金额。
So, but here's another interesting thing about the whole thing is I didn't really use Instacart that much before.
不过，整件事还有一个有意思的地方：我以前其实很少使用 Instacart。
You know, it's like every once in a while, maybe twice a year, I would like like, use Priscilla's account or something. Yes.
我只是偶尔使用，可能一年两次，例如使用 Priscilla 的账号。
I went and got an Instacart account to give Muse, and now I use it like once or twice a week because like I'm talking with Aurelia with Muse, and we're coming up with these recipes, and it's delivering them, and now it's like a core part of my workflow.
后来，我专门注册了一个 Instacart 账号供 Muse 使用。现在，我每周会使用一两次，因为我会通过 Muse 与 Aurelia 交流。我们一起设计食谱，再由 Muse 安排配送。如今，它已经成为我整个流程中的核心一环。
So I actually think that the opportunity here isn't just, okay, it's going to be 20% more efficient on the conversion rate.
所以，我认为这里的机会并不只是让转化率提高20%。
Like, I am a new Instacart like subscriber, and I just wouldn't have done this before. Yes.
例如，我现在成了 Instacart 的新订阅用户，而以前根本不会这么做。
So, I do think that there's this question, and then there's this question, which I think is these sort of interesting— like, this is maybe the more interesting economic question, is that Instacart account really for me, or is it for the Muse?
所以，我确实认为这里存在一个问题。此外，还有另一个很有意思的问题；从经济层面看，这个问题可能更值得关注：这个 Instacart 账号到底是供我使用，还是供 Muse 使用？
Because I think, like, over time, I think that the question is, is the total addressable market for all of commerce just increasing dramatically?
因为我认为，随着时间推移，问题将变成：整个商业领域的潜在市场总规模是否会大幅增长？
And I think that's where this starts to, like, become a really very exciting opportunity where anyone who sits this out, I think, is going to end up feeling sad over time.
我认为，正是从这里开始，这件事成为了一个非常令人兴奋的机会。任何选择置身事外的人，最终都会随着时间推移而感到遗憾。
We have been very excited about agentic commerce for a long time, obviously, because again, felt to us like clearly how people are going to buy things in the future.
我们很早就对智能体商业充满期待，因为在我们看来，这显然就是人们未来购买商品的方式。
And that was both apparent because no one needs all that friction in their life from the actual buying stage, and how early product research with AI had product-market fit, even from kind of the earliest AIs.
这一点之所以显而易见，一方面是因为没有人需要在实际购买阶段承受生活中如此多的麻烦；另一方面，早期利用人工智能进行产品研究的方式已经实现了产品与市场的契合，哪怕使用的只是最早期的人工智能。
It really feels like computer use was the last big unlock that we needed to make it happen, because now you can ask your muse to buy anything, more or less.
计算机操作似乎确实是促成这件事所需的最后一项重大突破，因为现在你几乎可以让 Muse 购买任何东西。
Yeah, it's pretty— this is a set of things that Matt really obsessed over.
是的，这正是 Nat 真正反复琢磨的一系列事情。
I think, I think there was a lot of, like, I think a lot of magic of Muse is the fact that like a lot of this stuff just works.
我认为，Muse 的许多神奇之处在于，其中很多功能直接就能使用。
And so like clearly if everything was just like a little bit less reliable, it would feel like a total pain in the ass and nobody would use it and people just throw away the app.
显然，如果所有功能的可靠性都稍差一些，使用起来就会非常烦琐，根本不会有人使用，大家只会直接删除这个应用。
And so I think the fact that so much of this has gotten to the point of reliability where You know, it just works is like a huge part of the magic.
我认为，其中很多功能已经达到了足够可靠的程度，确实能够正常使用。这正是它如此神奇的一大原因。
But this is, I mean, but yeah, I mean, this getting this experience to be good, I think is something that he and a core group of people on that team really focused on.
不过，我认为，把这种体验做好，正是他和团队中一小群核心成员真正专注的事情。
I think I had a dinner with Patrick, like, I want to say it was 2023, where I was, he was like, what do you think we should be doing now?
我记得大概在 2023 年，我和 Patrick 吃过一次饭。当时，他问我：“你觉得我们现在应该做什么？”
I was Oh, you're like, you're already way behind the 8-ball on this agentic commerce stuff.
我回答：“你们在智能体商业这件事上早已远远落后了。”
So you were saying this to Patrick or Patrick saying this to you?
这是你对 Patrick 说的，还是 Patrick 对你说的？
Well, because Patrick was very politely asking, what do you think we should be doing?
因为 Patrick 当时非常客气地问我：“你觉得我们应该做什么？”
And so, um, no, it's very interesting to look at the agentic commerce stats at Stripe because we started developing all this stuff and we had a whole bunch of announcements at Sessions back in May, uh, and then it only as of September basically started going vertical.
不过，查看 Stripe 的智能体商业数据确实很有意思。我们开始开发所有这些东西，并在 5 月的 Sessions 上发布了大量公告，但直到 9 月，相关数据才基本开始直线上升。
Meta is building increasingly capable models, including Muse Wonder, and their models are available on OpenRouter, Alongside hundreds of others.
Meta 正在打造能力日益强大的模型，包括 Muse Wonder；他们的模型和另外数百个模型都可以在 OpenRouter 上使用。
Stripe has agreed to acquire OpenRouter, who's the leading model gateway.
Stripe 已同意收购领先的模型网关 OpenRouter。
OpenRouter already routes more than 10 trillion tokens a day and handles much of the complexity of working with different models through a single API, making it easier to experiment, find competitive prices, and automatically route around provider outages.
OpenRouter 每天路由的 token 已超过 10 万亿个，并通过统一的 API 处理使用不同模型时面临的诸多复杂问题，让用户能够更轻松地进行实验、寻找更具竞争力的价格，并在服务提供商宕机时自动切换路由。
People have been trying to make the speak to a computer AI vision happen for a really long time.
长期以来，人们一直希望让“对着电脑说话”这一人工智能愿景成为现实。
Where Apple launched Siri in 2014, 2015, something like that.
Apple 大概是在 2014 年或 2015 年推出 Siri 的。
And obviously there's been Voice Mode and various AI products for a few years now.
当然，语音模式和各种人工智能产品如今也已经出现好几年了。
If you were driving your car and you wanted to check on the baking or move a meeting or something like that, you actually probably wouldn't do it with the voice.
如果你正在开车，想查看食物烤得怎么样，或者调整会议时间等，可能并不会使用语音来完成这些操作。
Siri right now is actually a good example. As we're talking over each other.
Siri 就是一个很好的例子。我们说话时会互相打断。
So is it just technical limitations? That's why we haven't done it?
这只是因为技术限制吗？这就是我们一直没能做到的原因吗？
So there's a question of when you speak with it, even if it understands, how's that going to work?
问题在于，当你与它交谈时，即使它能够理解，接下来又该如何运作？
Then I think that there's a lot of stuff about the way that people have conversations.
此外，我认为还有许多问题与人们的对话方式有关。
It isn't typically, I will give a prompt, and now I will wait, and you will speak.
通常并不是由我先给出一条提示，等待片刻后再轮到你说话。
Um, how many of your sessions when you're talking to an agent in voice mode, do you basically, oh, it starts saying something, but then like, I don't know, it's like, I kind of feel like if I pause for 3 seconds, it's going to think I'm done and want to speak.
你们有多少次在使用语音模式与智能体交谈时遇到这种情况：它开始说话了，但我还在考虑接下来该说什么；我总担心，只要停顿 3 秒，它就会认为我已经说完，并且希望它开口。
And then if it starts speaking, but then I cut back in, is it going to think that it's like a whole new question?
如果它已经开始说话，我又插话，它会不会认为这是一个全新的问题？
Um, I think there's a lot of the stuff about the conversational dynamics that you want to get right.
我认为，对话节奏方面有很多细节需要妥善处理。
But then there's this other issue, which is that For voice overall, it's quite high bandwidth for you to communicate with voice, but in terms of consuming information, having someone speak to you is very inefficient compared to having it be visual.
但还有另一个问题。从语音交互的整体情况来看，使用语音表达时，信息传输效率很高；但在接收信息方面，与视觉方式相比，听别人说话的效率非常低。
So I think like that's why, you know, partially when we do these, these interfaces, I think it's, it's kind of working on its task.
因此，我认为，在设计这些界面时，可以让它自行处理任务。
Most of the researchers just kind of like whisper into their keyboard because it is like, as Mark's mentioning, it's like the highest bandwidth way to get your like thoughts out, or usually the lowest friction way to get your thoughts out.
大多数研究人员基本都会对着键盘轻声念叨。正如 Mark 刚才提到的，这是表达自己想法时信息传输效率最高，或者通常来说最省力的方式。
I think the most extreme version of this, by the way, is I think how college students now use AI is maybe the most interesting, where they just— I think a very common modality of how to use AI is that they just will completely word vomit for 10, 15 minutes, being full stream of consciousness, chain of thought of all the things.
我认为，最极端、也可能最有意思的例子，是现在大学生使用人工智能的方式。一种非常常见的方式是，他们会连续讲上 10, 15 分钟，以完全意识流的方式说出整件事的全部思维链。
And have AI effectively do all of this organization of it and pulling things together and structuring everything.
人工智能会完成所有整理工作，将各种内容汇总起来，再梳理出清晰的结构。
I do think agents is like, that's a big step change, because I think the mental model is just very different.
我确实认为，智能体带来了巨大的跃迁，因为人们对它的认知模式已经完全不同。
Like, if it's just like an informational thing that you're talking to, that like, you will clearly get into situations where it just feels useless.
例如，如果与你对话的只是一个提供信息的工具，那么你显然会遇到认为它完全无用的情况。
Whereas now that it can do things for you, it's much closer to the experience of talking to a human. Totally.
但现在，它能够替你做事，这种体验就更接近与真人交谈。我完全同意这一点。
And so I think there's all these interesting uncanny valleys in interacting with AIs that until they've been able to take actions have felt kind of strange.
因此，我认为，与人工智能互动时会遇到各种有意思的恐怖谷现象。在人工智能尚不能采取行动时，这种互动总会让人感到有些奇怪。
And now that we have agents, I think one of the cool things about Muse is this confluence of a number of things.
现在有了智能体，我认为 Muse 很酷的一点就在于，它是多项事物的交汇。
I think The fact that Muse is embodied and it like is this little guy that like you feel like is working on your behalf is also only possible and a coherent like representation and embodiment of the product because it has persistent memory and it has the ability to do things for you and the ability to take actions and has like a little bit of a personality.
Muse 有一个实体形象，就像这样一个小家伙，让人觉得它正在替自己做事。它之所以能够成为一个连贯统一的产品形象与实体，是因为它拥有持久记忆，能够为你做事、采取行动，同时还有一些自己的个性。
It's like, okay, a lot of people want it even with just like no display or a small display. Yes.
很多人都想要它，即使它根本没有显示屏，或是只有一块小显示屏。
And I— and then we started shipping the display glasses that aren't the kind of wide field of view full augmented reality, which, I mean, I thought would be awesome too, but even the small display is pretty powerful.
随后，我们开始交付带显示屏的眼镜。它并不是那种拥有宽广视野的全增强现实眼镜。我认为全增强现实眼镜也会很棒，但即使只配备一块小显示屏，它的功能也相当强大。
So I think you'll start getting some interesting interactions there that are kind of fully multimodal, right?
因此，我认为人们将开始看到一些很有意思的交互方式，它们可以说是完全多模态的。
So it's voice in and out and It'll be able to see what you see, and it'll be able to kind of paint the pixels.
它能够通过语音进行输入和输出，也能看到你所看到的内容，还能以某种方式绘制像素。
And one of the— I mean, the other exciting thing that we've done here is Muse Gadgets, which is really, I think, unleashing the ecosystem to try cool device formats and form factors that could be really exciting.
我们在这里所做的另一件令人兴奋的事，就是 Muse Gadgets。我认为，它真正释放了整个生态系统的潜力，让人们可以尝试各种很酷的设备样式和形态，其中可能会出现一些非常令人兴奋的产品。
Like, since releasing that, we've gotten actually even a number of phone manufacturers, you know, talk to us about excited You know, we're like, we're just going to integrate Muse Gadgets into our phone and see how well it works. And like, we're really excited about that.
自发布以来，已经有不少手机厂商联系我们，表达了他们的兴奋之情。他们表示，打算将 Muse Gadgets 集成到自己的手机中，看看效果如何。我们对此也感到非常兴奋。
Like, I think, um, I think part of what is cool about the current moment of technology is that like, it's very clear agents are really exciting.
我认为，当下技术的魅力，部分在于大家已经清楚地认识到，智能体确实令人兴奋。
One of the gadgets that we just shipped was the little HomeLink.
我们刚刚推出的小型设备中，有一款小巧的 HomeLink。
It's like the size of a little, like, nickel or something, but with a USB, right? Yeah.
它差不多只有一枚 5 美分硬币那么大，但配有 USB。
Um, uh, so you can basically connect that, and then your Muse can connect to everything in your home.
把它连接好之后，Muse 就可以连接家里的所有设备。
This is actually one of the most fun things too is just having it connect to your Sonos and connect to all your lights.
这也是最好玩的功能之一：让它连接你的 Sonos 和家里的所有灯。
I mean, I have it basically like when it's time for me to start winding down in the evening, I just have it like dim all the lights across the house and it's nice.
我把它设置为每天晚上到了该放松的时候，就自动调暗全屋的灯光，体验很不错。
Everyone's worried about, you know, Google went free cash flow negative.
大家都很担心，Google 的自由现金流已经转为负数。
You know, you were talking about 1 gigawatt and then 5 gigawatts and everything like this.
你之前提到过 1 吉瓦，之后又提到 5 吉瓦，以及诸如此类的规划。
Meta's not free cash flow negative yet, but there's always more CapEx to fund.
Meta 目前的自由现金流尚未转负，但总有更多资本支出需要投入。
No, I mean, look, I think that there are going to be a lot of ways to create value from this.
我认为，我们会有许多方式从中创造价值。
So what we're trying to do with Muse specifically is make it so that it can be widely used, because I think that there are going to be some kind of valuable network effects at scale as people start being able to interact with each other, and that's going to be great.
我们在 Muse 上的具体目标，是让它得到广泛使用。我认为，随着人们开始借助它彼此互动，规模扩大后将产生某种有价值的网络效应，这会非常棒。
So I think there are all these reasons why you want it to get to large scale.
出于上述所有原因，我们都希望它能够达到很大的规模。
Partners are going to want to integrate more the more people are using it.
使用它的人越多，合作伙伴就越愿意进行更深入的集成。
The best way to do that is to basically have a model where it is affordable and free for people to use.
要实现这一点，最好的办法就是采用一种所有人都负担得起且可以免费使用的模式。
But there are going to be a lot of ways that I think we can scale revenue.
但我认为，我们会有许多扩大营收的方式。
You know, basically, just because the base version of it is free and we want the free version to be excellent doesn't mean that there are things that just naturally aren't going to be economical to include in that.
基础版本是免费的，我们也希望免费版本足够出色，但这并不意味着将某些功能纳入免费版本在经济上必然可行。
So, you know, if you want the kind of— you know, we talked about this earlier.
因此，如果你想使用我们之前讨论过的那类功能，就需要付费。
If you want to start using Muse Wonder in Muse tomorrow, then, you know, the way that you'll do that is you'll pay for a subscription.
如果你明天就想在 Muse 中开始使用 Muse Wonder，就需要付费订阅。
There are a bunch of other things around commerce that we're quite excited about too, where we basically think— we believe that Muse is actually going to make people money and save people money by using it, and that over time there are lots of different ways to effectively take a small cut of what you are saving or making for people, and that will end up being a very large revenue stream.
在商业方面，还有许多令我们非常兴奋的机会。我们相信，使用 Muse 能够切实帮助人们赚钱和省钱。随着时间的推移，我们将通过多种方式，从为用户节省或赚取的钱中获得一小部分，并最终形成非常庞大的收入来源。
Well, there are already all these sites that I think are interesting where people are like, go look through my emails, find a bunch of subscriptions I don't actually use, cancel them, whatever.
现在已经有不少这类网站。人们会让它们检查自己的电子邮件，找出那些根本没有使用的订阅并将其取消，诸如此类。对此，我觉得很有意思。
Um, yeah, so yeah, I mean, that, um, I don't know if that's the thing that we would monetize, but, but like, but you get, you get, you kind of get the point and where it helps you make money.
是的。我不知道我们是否会通过这种功能变现，但你大概能够理解我的意思，也就是它可以如何帮助你赚钱。
I mean, one of the ideas that we kind of talk about, I mean, this is, this is something that I think you've been kind of obsessed with, is making it like a business in a box for small businesses.
我们讨论过一个想法，我认为你一直对此颇为着迷：把它打造成一套面向小企业的“开箱即用”经营方案。
So someone can just kind of— it can help them administer all these parts of what they're doing.
这样一来，它就能帮助人们管理业务中的方方面面。
And you can imagine that being very valuable and being very profitable, both for the business owner and for Meta.
可以想象，这将非常有价值，也能带来很高的利润，无论对企业主还是对 Meta 都是如此。
But I think one of the things that's interesting is, like, OK, if we're building the models to build the personal agents and the business agent, there is this large surplus of value because now the model, it's very general, right?
不过，有一点很有意思：如果我们打造的是用于构建个人智能体的模型，那么商业智能体领域将存在大量价值盈余，因为这个模型本身具有很强的通用性。
A big part of why we want to start the Meta Enterprise Platform and have CJ Desai join us to build this is we have this view, which is like, all right, yeah, sure.
我们之所以想启动 Meta Enterprise Platform，并邀请 CJ Desai 加入我们、负责打造这一平台，很大程度上源于我们的一个判断。
So traditionally Meta hasn't been an enterprise company, but we believe that a huge part of the challenge is building the leading models in the first place. And they're very general.
一直以来，Meta 都不是一家企业服务公司，但我们认为，首要挑战在于打造领先的模型，而这些模型具有很强的通用性。
And once we do that, if we make a significant investment in building an enterprise group that can provide that in the way that enterprises need to consume it, then I mean, I think a lot of enterprises are going to want to use Muse Wonder.
一旦实现这一点，如果我们投入大量资源组建企业服务团队，以企业需要的方式向它们提供这些模型，那么许多企业都会希望使用 Muse Wonder。
And like, it's so, um, so I think that that is going to end up being a very big business too.
因此，我认为它最终也会成为一项规模非常大的业务。
And all of these things put together, coupled with the fact that we were already starting off as a very profitable business that can, that has good credit and can kind of underwrite these large investments, I think puts us in a pretty good position as a lab to be able to build out the compute that we need to build leading products, build up the network around this, and position the company to really lead in the area for decades to come.
将这些因素结合起来，再考虑到我们从一开始就是一家利润丰厚、信用良好，足以支撑这些大规模投资的企业，我认为，这使我们作为一家实验室处于相当有利的位置：既能建设打造领先产品所需的算力，也能围绕这项事业构建网络，并使公司真正占据有利位置，在未来几十年引领这一领域。
Well, the history of AI as an industry, I think, to some degree is like, you know, people have always sort of been hand-wringing about, you know, how are these businesses going to make money?
在我看来，从某种程度上说，人工智能作为一个行业，其发展史始终伴随着一种忧虑：这些企业究竟要如何赚钱？
And I think, you know, every year people are sort of generally surprised at the upside.
但我认为，几乎每一年，人们都会对它超出预期的增长感到惊讶。
Yeah, exactly. This is what— this is kind of Meta's whole playbook, right?
是的，正是如此。这其实就是 Meta 一贯的做法。
Is we build something, we get it to a billion people, then we scale the revenue while growing our revenue primarily in the meantime by continuing to scale the things that are already kind of at scale and growing. Yeah.
我们打造一款产品，使其触达十亿人，然后扩大其收入规模。在此期间，我们的收入增长主要依靠继续扩大那些已经形成规模且仍在增长的业务。
I think this is also one of the dynamics, like, I think, you know, I think Dario's talked about this as well, but like, just in general, because it requires so much advanced planning to build these models.
我认为，这也是其中一条规律。Dario 应该也谈到过这一点。总体而言，打造这些模型需要进行大量前瞻性规划。
Like, if you look at the R&D expense at any given time, that is like over— an overlapping R&D expense of multiple model generations at once. Yes.
例如，观察任一时间点的研发支出，就会发现其中同时包含多个相互重叠的模型世代的研发费用。
The other thing is that we use the models to accelerate our core business, right?
此外，我们还会利用这些模型来加速核心业务。
So I mean, you have this core business that's hundreds of billions of dollars and still growing extremely quickly.
这项核心业务的规模达数千亿美元，而且仍在飞速增长。
And now, like, if you can grow it a few percent faster by using these models to accelerate these things, then that ends up funding a huge amount of this too as you're scaling it out.
如果利用这些模型能使其增速再提高几个百分点，最终也会为这方面的大规模扩张提供巨额资金。毕竟，这项业务的规模仍在不断扩大。
Yeah, I think you're also getting at something, Alex, which is like there's an interesting accounting artifact where if you go build a data center, you have CapEx and you get to depreciate that over a useful life of many years.
Alex，我认为你也指出了一个问题：这里存在一种有趣的会计现象。如果建造一座数据中心，这笔成本会被列为资本支出，并可在多年的使用寿命内计提折旧。
And people debate what that useful life is, but the point is you do get to depreciate it, whereas a lot of the AI costs are also just in researcher salaries, and those salaries are being spent on, you know, model training that also has a useful life of many years, but you do not get to capitalize generally.
人们会争论其使用寿命究竟有多长，但关键在于，这笔支出确实可以计提折旧。人工智能的许多成本只是研究人员的薪资，而这些研究人员从事的是模型训练。此类模型同样具有多年的使用寿命，但相关支出通常无法资本化。
And that's also a financial advantage for us building out the data centers as opposed to a bunch of the other companies that are paying clouds, whereas they're paying the annual rate, but then at some point the GPUs will be fully depreciated, but we're probably still going to be using them for a while because they I mean, the value of even like an H100 at this point is higher than it was last year, right?
与向云厂商付费的公司相比，我们自建数据中心在财务上也更有优势，因为这些公司需要按年付费。到了某个时候，这些 GPU 会完成全部折旧，而我们可能仍会继续使用一段时间。即使是 H100，目前的价值也比去年更高。
Yeah, I think part of the exciting— part of what we're starting to see in terms of the models, especially the frontier models that are being developed, like Muse Wonder versus, let's say, a Fable or an Astra, is that we actually are starting to focus in quite different directions from the other models.
我认为，令人兴奋的是，我们已经开始看到，对于这些模型，尤其是目前正在开发的 Muse Wonder 等前沿模型，与 Fable 或 Astra 相比，我们实际上已经开始聚焦于截然不同的方向。
And I think there's a lot of intentionality that goes into the model development process that we've discussed as part of this, you know, in this conversation, that is genuinely different from how other models are focusing.
我认为，模型开发过程中融入了许多有意识的考量，其中包括我们在这场讨论中提到的内容。这些考量确实不同于其他模型所关注的方向。
And so, you know, when these models kind of get collapsed to the same set of benchmarks, that elides all of the—
因此，当这些模型被压缩成同一套基准测试结果时，所有这些差异都会被掩盖——
The interesting, um, sort of like, uh, texture that differs between the various models.
有意思的是，各个模型之间具有不同的质感。
And so I think one thing that is an exciting opportunity for OpenRouter, generally speaking, is like, you know, um, helping to Helping developers understand what these sort of subtle differences are so that they can make the most informed decisions about what models ultimately they select.
因此，我认为 OpenRouter 面临着一个非常令人兴奋的机会：帮助开发者理解这些细微差异，使他们在最终选择模型时能够做出最明智的决定。
And then also you can imagine the flip side of that, which is, you have all these developers using these models.
也可以反过来思考：目前有如此多的开发者正在使用这些模型。
And so those developers are great people to ask about what are the interesting textural differences between these models that are more than just tweet length.
因此，我们需要了解这些模型之间存在哪些有趣的文本差异，而不只是质感上的差异；这些差异也远不只是推文长度不同。
Like, this model's good, this model's bad, more than just that level of insight.
相关见解不能只停留在这个模型好、那个模型差的层面。
And I think this is like getting to the wine sommelier version or form of understanding the models and their sort of relative strengths and weaknesses, I think would be quite exciting.
这就像达到品酒师的水平，以相应的方式理解模型及其相对优势与劣势。我认为这会非常令人兴奋。
Okay, so based on your experience, the OpenRouter view that there will be a diversity of models in the future and they're optimized for different tasks and—
那么，根据你们的经验，OpenRouter 的观点是，未来将出现多种多样的模型，而且它们会针对不同任务进行优化，而——
We care a lot more about personal and professional agents, my observation, than the other major labs, which is good.
据我观察，与其他主要实验室相比，我们更加重视个人智能体和专业智能体，这是一件好事。
I mean, I think that part of what I love about you guys working on OpenRouter and being able to scale it is just the more infrastructure that can get developed, that supports a diversity of players in the ecosystem, I think is good overall.
我很欣赏你们开发 OpenRouter，并且能够将它发展壮大。能够开发的基础设施越多，就越能支持生态系统中的多元化参与者。总体而言，我认为这是一件好事。
I also, for the same reason, am quite a fan of some of the recent NVIDIA moves to support the open source ecosystem.
出于同样的原因，我也非常支持 NVIDIA 最近为支持开源生态系统而采取的一些举措。
I don't know, but this just kind of goes with our whole theory and philosophy about, like, what does a good future look like?
这也符合我们一直以来的整体理论和理念，即一个美好的未来应该是什么样的。
We think that a good future is one where you achieve a, like, you basically get prosperity by empowering people.
我们认为，在美好的未来中，繁荣主要通过赋能人们来实现。
That's kind of the personal superintelligence thing that we're doing, but there's the whole business version. Yeah.
这有些类似于我们正在开发的个人超级智能，但它还包括一套面向企业的完整版本。
But for a lot of the important safety topics, like, our view is that the way to get there is having a diversity and having checks and balances by kind of having a lot of different things out there and accessible to people, not like just having a couple of good models that are only in a couple companies' hands.
对于许多重要的安全议题，我们认为，实现安全的方式是保持多样性，让许多不同的事物广泛存在并可供所有人使用，从而形成制衡，而不是让少数几家公司掌握少数几个优秀模型。
I think that that would be— from my perspective, I think that is probably a riskier configuration for the future than lots of people having access to different things.
在我看来，后一种格局可能会给未来带来更高的风险。相比之下，让更多人接触到不同的事物会更好。
Where I feel like maybe people predicted more of a winner-takes-all dynamic in AI, or one person clearly— like, a story that people told back in 2023 is that one AI lab will start to develop an advantage, and that will allow them to get recursive self-improvement going, And then they'll just pull further and further away from everyone else.
此前，人们可能预测人工智能领域将呈现赢家通吃的局面，或者由某一方明显胜出。例如，2023 年时有一种常见说法：某一家人工智能实验室会开始形成优势，由此启动递归式自我改进，进而将其他参与者远远甩在身后。
Whereas it has remained extremely competitive where there's kind of multiple people at the frontier.
但实际情况却是，竞争始终极其激烈，前沿位置上同时存在多方参与者。
You guys with Wonder today are coming along with your frontier-class model.
如今，你们也凭借 Wonder 和自己的前沿级模型迎头赶上。
Open source models have done an impressive job of keeping up, all told.
总体而言，开源模型在紧跟前沿水平方面表现得十分出色。
And so we had this story a few years ago that someone was just going to run away with the prize.
因此，几年前我们曾设想，可能会有一方直接独占鳌头。
Well, I think if all players, when they were in the lead, executed perfectly, they would probably be able to stay in the lead.
我认为，如果所有参与者都能在领先时做到完美执行，他们很可能就能一直保持领先。
I think the power and strength of distillation is certainly another factor that's certainly benefiting the open source ecosystem quite a bit. And I think this is just like a—
我认为，蒸馏的威力和效能无疑是另一个因素。它显然让开源生态受益良多，同时也体现出一种力量。
It's not like anyone planned for distillation to work super duper well, but it just turns out that you can get a lot of capabilities naturally through some form of distillation.
并不是有人事先设计好，要让蒸馏产生如此惊人的效果。事实证明，通过某种形式的蒸馏，自然而然就能获得大量能力。
And this is, broadly speaking, distillation is quite a good force for the entire ecosystem because it means that you can pack very strong intelligence capabilities into smaller and smaller models, which enables us to serve them at much wider scale.
总的来说，蒸馏对整个生态而言是一股非常积极的力量。这意味着，可以把非常强大的智能能力压缩到越来越小的模型中，从而让我们能够以更广泛的规模提供这些模型。
And that's certainly one of the very exciting stories is just like really small models today are like, you know, much more powerful than the biggest models a year ago or 2 years ago.
这无疑也是最令人兴奋的进展之一。如今非常小的模型，都比1年前或2年前最大的模型强大得多。
And then I think the last is that there's just sort of this— I think there's like, I actually think a lot of people in the ecosystem want this future where it is not just this like, you know, definitely.
最后，我认为，生态中的许多人其实都不希望未来变成那样。
And I think, like, you know, this is actually one of the reasons why a lot of people were excited to join Meta and Meta Superintelligence Labs in the first place, was that they were just sort of like, you know, I want to work at a place that is focused on giving this technology to billions of people.
我认为，这也是很多人一开始期待加入 Meta 和 Meta Superintelligence Labs 的原因。他们希望在一个致力于把这项技术带给数十亿人的地方工作。
That is, like, not the stated goal necessarily of all the other frontier labs.
这未必是其他所有前沿实验室公开宣称的目标。
And so, I think there's, like, a lot of sort of just, like, natural forces where many people in the industry are sort of, like, grappling with, okay, what is the actual world state that I want to live in in the future?
因此，我认为其中存在很多自然而然的力量。业内很多人其实都在思考，自己未来真正想生活在一个什么样的世界里。
And there's lots of sort of, like, decisions that people make at all levels in terms of where they work or who they do business with or what business deals they do that sort of, like, drive towards a more equalized outcome.
人们在各个层面都会做出许多决定，例如去哪里工作、与谁做生意，或者达成什么样的商业交易。这些决定都会推动结果变得更加均衡。
You get into the story of the MSL startup, which I also wanted to ask about, which is it just seems fundamentally very hard still to— again, we talked about what goes into developing a frontier model, but also from a standing start, like when was the standing start for MSL?
你讲到了 MSL 创业公司的故事，这也是我想问的。从根本上看，这件事似乎依然非常困难。我们之前讨论过开发前沿模型需要投入什么，但还有一个问题：如果从零起步算，MSL 是什么时候起步的？
Well, we launched Llama 4 on April 4th, I think it was, 2025.
我们是在 2025 年 4 月 4 日发布 Llama 4的，我记得是这个日期。
And then I think, what was it, April 20th or 21st was, I think, when I was like, hey, when I started calling people for advice on what to do.
我记得，大概在 4 月 20 日或 21 日，我意识到自己该开始打电话联系别人，请教接下来该怎么做了。
Well, I mean, I think that there was this whole thing early on which is just like, first we had to pull together this core group.
我认为，早期的首要任务是组建核心团队。
So, um, so Nat and Alex and I, we were like talking a bit up front, um, and, you know, at first we were, you know, brainstorming.
Nat、Alex 和我前期进行了一些讨论，最初主要是共同开展头脑风暴。
Yeah, like I was talking to Nat about, could we— is there a way that we can like structure this so that way it's like a win for Alex's shareholders and like we can like get him over here?
例如，我当时和 Nat 讨论，能否设计出一种方案，既让 Alex 的股东受益，又能邀请他加入我们。
And, and then basically I was just, I was in this mode where I was just like asking a lot of different people what we should do.
当时，我基本上一直在询问许多不同的人，了解我们应该如何推进。
And like every time that, that we would talk, I'd come away with like 3 more things of like, oh yeah, this is what we should do.
每次讨论结束后，我都会获得大约 3 个新想法，并认为这些正是我们应该做的事。
It's like very like high signal, very like a lot of strategic clarity.
这些讨论的信息密度很高，战略思路也非常清晰。
And then Nat and Daniel joined and we pulled together a few of the the initial researchers.
后来，Nat 和 Daniel 加入了，我们又召集了几位最初的研究员。
But like, in the beginning, it was sort of like the Supernova thing where we just kind of like jammed a bunch of people together.
不过，最开始的情形有些像 Supernova 时期：我们只是把一群人硬凑到了一起。
And I don't know, it was like, I don't think either of us slept that much for the first couple of months.
我觉得，我们俩最初几个月几乎都没有睡觉。
I like, don't even, I kind of have this mental block almost where I like, I kind of almost cleared my calendar for like a month or 2 and was just meeting researchers and talking to them.
我甚至不愿回想那段经历，仿佛对此产生了某种心理障碍。当时，我几乎空出了大约 1 个月或 2 个月的日程，专门与研究人员见面交流。
And I remember even before we announced the deal where you were going to come over, we were basically like, you were recruiting people and people were like, what is going on?
我记得，在我们宣布你要过来之前，你基本上就已经开始招人了，大家都在问究竟发生了什么。
But, um, but then there was the whole question, which was a really hands-on data labeler.
但后来还有一个大问题，确实涉及一位非常亲力亲为的数据标注员。
But, but that was, that was this whole question too, where it's like, all right, well, like, okay, so if we're not happy with the trajectory of the current program, Like, who were the right set of people to lead the program going forward?
但这也是当时面临的一个重大问题：如果我们不满意当前项目的发展轨迹，未来应该由哪些合适的人来领导这个项目？
And I was always just very impressed by how he thought about things strategically.
他思考问题时展现出的战略眼光一直令我印象深刻。
And I kind of had this view of like, all right, yeah, currently he's doing this very operational data business.
我当时大概是这样想的：他现在经营的是一门非常偏重运营的数据业务。
But it always felt to me like he was the sharpest entrepreneur of his generation, of the ones who I had had a chance to meet.
但我一直认为，在我有机会见到的同代人中，他是最敏锐的创业者。
And like, it always just struck me, it's like, all right, like, this is a guy who, yeah, he's like running this like kind of very grindy operational business, but like—
而且我一直有一种强烈的感觉：这个人经营的虽然是一门非常辛苦、偏重运营的业务，但是——
It's not a— it's a— there's more Commentary on the difficulty of running the business.
这里更多是在说经营这门生意有多么困难。
When you run the 8th wonder of the world, Alex, a lot of businesses look like a grindy operational business.
Alex，当你经营着世界第8大奇迹时，很多生意看起来都会像是辛苦、偏重运营的业务。
I'm like, all right, like, even though this is sort of I think maybe when we got started, a lot of people were like, okay, like, how is this going to work?
我当时想，尽管这多少有些……但在刚开始时，很多人可能都在思考：双方要如何合作？
It's like the kind of like the social media guy is working with the like data guy.
就像是那个从事社交媒体业务的人，要和那个从事数据业务的人一起工作。
Like, is that like, are we going to build a frontier lab here or what?
这真的可行吗？我们是要建立一个前沿实验室，还是别的什么机构？
But, but I mean, but my, but I mean, my like really strong conviction, because I talked to like all the leaders of all the labs, like all the heads of research, like all this stuff.
但我当时对此有着非常强烈的确信，因为我已经与所有实验室负责人、研究主管等人都谈过。
And like, I was like, no, Alex is the guy. Like, he's the fucking guy.
随后我便认定，不，就是Alex。就是他妈的他。
And like, and we're going to, and basically, like, we're going to basically be able to, to, to go pull together a team and recruit the people that we need to go do this.
而且我们基本上能够组建一支团队，把完成这件事所需要的人才招进来。
And like, we need this kind of very, this kind of like strategic clarity that he has around the different ways that we want to approach this, the kind of extreme focus on talent density, on kind of like the clarity of communicating the different vision, which we kind of fashioned together over this period.
我们需要他那种非常明确的战略思路，以厘清我们希望通过哪些不同方式推进这件事；也需要他对人才密度的极致专注，以及清晰传达不同愿景的能力。这些都是我们在那段时间里共同逐渐摸索出来的。
And a bunch of things that I think were kind of extreme but have basically worked, like having this, this like very flat culture was something that you were kind of really obsessed with, right?
还有很多事情在当时显得有些极端，但最终基本都奏效了。例如，打造一种非常扁平的文化，就是你当时特别执着的一件事。
That's like, we have, it's not, it doesn't, it's not quite that extreme in practice.
我们确实如此——但实际上并没有极端到那种程度。
I think that there's like, like team leads and all, and all this stuff.
我认为公司里仍然会有团队负责人等角色，也会有各种不同的安排。
Yeah, well, that's— But like, but there were these series of things that I think quite early we, you kind of articulated, and I was like, okay, this is like kind of extreme, but it seems like directionally right, and like we should go do this.
不过，当时确实发生了一系列事情。我记得你很早就提出了这个想法，而我当时认为，这虽然有些极端，但大方向似乎是正确的，我们应该这样做。
And, you know, I went and like talked to other people who were more like the people you'd conventionally think of, and it's like, it was just very much like, I don't know what you'd expect expect a big company to operate.
后来，我也与其他更符合人们通常想象的那类人交谈过，但他们采用的都是非常典型的运作方式，也就是人们预期中大公司会采用的方式。
As you describe this, I'll note 4 things that I think even other opinionated, effective founder CEOs would not have done.
听你讲述这些事情时，我注意到其中有 4 点，可能连其他那些很有主见、能力也很强的创始人 CEO 都未必会做到。
One, you just called the recent prior effort by a bunch of people at the company a failure and not good enough.
第一，你刚才明确表示，公司里一群人此前的努力失败了，而且做得不够好。
I mean, you were more diplomatic than that, but the act of stepping off the trajectory— Yeah, but the act of standing up So, you know,
你当时的表达比这更委婉，但决定退出这项行动原有的方向，并公开站出来作出这一决定，本身已经传达了明确的信息。
Thing 2 is you are willing to pay ungodly amounts of money for effectively talent.
第二，你愿意支付高得惊人的费用，而目的实际上就是获得人才。
Like, you're paying a lot of money to bring Alex and some other people, but, you know, a lot of money for Alex in particular over.
例如，你花了很多钱邀请 Alex 和其他一些人加入，尤其在 Alex 身上投入了大量资金。
Thing 3 is, I'll note, Alex had never trained a frontier model before.
第三，我要指出的是，Alex 此前从未训练过前沿模型。
And so you were taking some amount of a bet on Alex's potential to do something where he had done stuff in AI, but no offense, he just had not done the specific thing that you were being hired to do.
因此，从某种程度上说，你是在押注 Alex 有潜力完成一件他从未做过的事。他确实从事过一些人工智能方面的工作，但无意冒犯，他确实没有做过你准备聘请他完成的这项具体工作。
And then 4th, you were saying you cleared your calendar and just only did this for 2 months.
第四，你说自己清空了日程，整整 2 个月只专注于这一件事。
I think a lot of CEOs would've had the business review and would have had the board thing and they would have had the customer stuff and things like that.
我认为，很多 CEO 仍然会照常参加业务评审，并处理董事会事务、客户事务等工作。
Those 4 seem to me pretty interesting where I think a lot of CEOs would not have done those. Yeah, that's interesting.
在我看来，这 4 点都很有意思，因为很多 CEO 都不会这样做。这一点确实值得关注。
I definitely think— I mean, one thing you may not appreciate this about yourself, but I think one thing that is quite special about Mark is the degree to which, like, once he concludes that something is the right thing or something makes sense, he just like fully shifts everything to be consistent with that belief set.
我确实认为，有一点或许连你自己都没有意识到。就你本人而言，我认为 Mark 有一个非常特别的地方：一旦他认定某件事是正确的，或者认为某件事合乎情理，他就会彻底调整一切，使所有事情都与这套信念保持一致。
And it is quite like, to your point, I think like the inertia of most, even very successful founders who like are opinionated and edgy or whatnot, would just be to like, oh, I still have like my whole calendar.
而且，正如你所说，我认为大多数人都会受到这种惯性的影响。即使是那些非常成功、很有主见、也很有锋芒的创始人，也会认为自己的日程仍然排得满满的。
And like, so I'll just like talk about doing this thing and hope that somebody does it kind of situation.
因此，我只能谈论应该做这件事，然后期待有人付诸行动，情况大致如此。
Whereas I think, No, I think when Marc concludes that something is the thing he has to do or makes a lot of sense, then he really does, like, he has, like, maybe very high learning rate in that sense.
但我并不这样认为。当 Marc 认定某件事是他必须做的，或者确实很有道理时，他就会真正付诸行动。他在这方面的学习率可能非常高。
Like, he's very willing to, like, pivot entire things to be consistent with it.
他非常愿意彻底调整整体方向，让一切都与之保持一致。
And I think that, like, you know, one very underappreciated thing is the degree to which so much of the company had to pivot to support MSL as well.
还有一点被严重低估了，那就是公司有如此多的部门都必须转向支持 MSL。
Like, I think that while it was a kind of like this more startup-y effort within a larger company, you know, the needs of MSL were still quite large.
尽管 MSL 有些像大公司内部一个更具创业公司风格的项目，但它的需求仍然非常庞大。
And especially as Muse has been successful, we've also needed lots of—
尤其是随着 Muse 取得成功，我们也越来越需要——
And this is one thing that I think kind of only is possible because Mark is the founder CEO, and he can ensure that we have kind of like a global optimization towards this sort of like AI maximalist version of Meta versus the natural inertia of an organization towards just local optimization ad infinitum.
这件事之所以能够实现，某种程度上正是因为 Mark 是创始人兼 CEO。他可以确保我们以全局优化的方式，推动 Meta 朝着人工智能最大化的形态发展，而不是任由组织的自然惯性驱使，只顾无休止地进行局部优化。
But it's also a lot of how we work together is it's important to me to be in the details of this work so that we can get everything around the company to go do what we need to do. Right.
这也很能体现我们的合作方式。对我而言，深入了解这项工作的细节非常重要，只有这样，我们才能推动整个公司去做我们需要完成的事情。
And so it's like, I mean, yeah, I'll have opinions about this stuff, but I'm not like, I'm not the lead researcher. Right.
我当然会对这些事情有自己的看法，但我并不是首席研究员。
It's like, it's, you know, it's, but I think that's kind of a lot of, I think how we work together well is that, yeah, it's, I mean, we're, we're kind of like, quite in the details of all of this stuff, but running the lab is a big thing.
我们之所以能够合作得如此顺利，很大程度上是因为我们都相当深入地参与了所有这些细节，但管理实验室本身是一项艰巨的工作。
And one thing that I think was quite important about when we set the whole thing up is Meta, the company, invested a huge amount to stand up MSL, and then we very clearly told everyone like, hey, there's no deadlines.
当初建立整个体系时，有一点非常重要：Meta 投入巨资组建了 MSL，随后我们又非常明确地告诉所有人，这项工作没有截止日期。
We are going to take things at the right pace, which I think on the one hand could be like, you know, absolutely terrifying.
我们会按照合适的节奏推进。一方面，这可能会让人感到无比恐慌。
It's like you've just like stood up this extremely expensive unit and you're telling everyone like, hey—
因为你刚刚成立了一个耗资极其巨大的部门，却又告诉所有人——
Yeah, and you're worried inside a larger company that like, okay, 7 years from now maybe something will ship, right?
是的，而且身处一家规模更大的公司，你会担心，也许要到 7 年后才会有成果发布。
And so, but I think that like, first, I think one of the cool things about, you know, focusing on talent density and hiring all these very brilliant people who have done so much already, that's very impressive, is, you know, everybody had like a lot of motivation to ship and everybody wanted to sort of like leave their mark, so to speak, on the field.
不过，首先值得肯定的是，我们专注于人才密度，招揽了许多已经成就斐然、聪明绝顶的人。这非常了不起。每个人都非常有动力将产品做出来，也都希望在这个领域留下自己的印记。
And so everybody was like heavily, highly self-motivated to move as quickly as possible.
因此，每个人都有极强的自驱力，希望尽可能快速地推进工作。
And one of the things that's been funny and I've had to continue reminding the team of is like, you know, it is a team of a lot of insecure overachievers.
有一件事很有意思：我必须不断提醒团队，这个团队里有很多缺乏安全感的高成就者。
So like at any one point, if I ask them like, hey, how good do you think— where do you think we'll be in 3 months?
因此，每当我问他们：“你们认为我们能做到多好？你们认为 3 个月后会达到什么水平？”
I think it's, it's not, it's not that they're pessimistic about— I think that the average person seems quite optimistic about MSL and very self-critical about their own project. Yeah.
我认为，这并不是说他们对——持悲观态度。整体而言，大家似乎都对MSL相当乐观，但对自己的项目却非常挑剔。
Like, I think, I think, and, and this is, I mean, in many ways Muse Wonder is like this incredible milestone for everyone because it is a— The team was very, very critical. Yeah.
从很多方面来看，Muse Wonder都是所有人共同取得的一个了不起的里程碑，因为参与其中的团队非常、非常挑剔。
You're peeking in the oven and it's like, I don't know, it seems kind of gooey still.
这就像往烤箱里看一眼，却无法确定结果，只觉得看起来似乎还有些黏糊。
And yeah, yeah, going back to it, I mean, there are all these intermediate intermediate signals that you can get along the way that things are good or bad.
回到刚才的话题，在整个过程中，你会得到各种中间信号，以此判断进展是好是坏。
So you have a different regime of evals that can help you in pre-training.
因此，在预训练阶段，会有另一套评测体系为你提供帮助。
And then you have the more traditional benchmarks that can help you towards RL.
进入强化学习阶段后，则可以借助更传统的基准测试。
But no, I mean, this model, kind of as you mentioned, has been cooking for quite a while.
不过，正如你所提到的，这个模型已经经过了相当长时间的打磨。
And You know, a lot of— I think one of the funniest things about AI research is that, you know, there's a lot of times where people just sort of like are kind of squinting at a bunch of dots and being like, yeah, looks like it's good, or, uh, looks a little bit off.
我认为，人工智能研究最有意思的一点是，很多时候，大家只是眯着眼睛盯着一堆数据点，然后判断：“看起来不错”，或者“看起来有些不太对劲”。
Like, one of the funniest things is the people have protractors, like analog protractors that they put against you know, a set of lines on their monitor to like measure an angle.
比如，最好笑的一件事是，有人会拿着实体量角器，贴在显示器上的几条线上测量角度。
So I think there's, I think there's like a huge amount of like, uh, like it's very, um, there's more vibes, um, driving like, okay.
所以我认为，其中很多判断更多是凭感觉作出的。
But when did we have this conversation where, when did you know it was good?
但我们是什么时候谈到这个问题的？你什么时候知道它的效果不错？
Um, well, we had this debate after it was done pre-training where it was okay.
预训练完成后，我们讨论过这个问题，当时觉得效果还可以。
How long will it take in RL before it is better than Muse Spark?
它需要经过多久的强化学习，才能超越 Muse Spark？
And I remember the folks on the pre-training team were just trying to set expectations.
我记得，预训练团队的人当时只是想让大家形成合理的预期。
But going back to the beginning, I think one of the things that was very interesting and challenging in standing the thing up was like, you really just wanted this like very high talent density effort.
不过，回到最开始，我觉得在推动这件事时，有一点非常有意思，也很有挑战性：你真正需要的是组建一支人才密度极高的团队。
And we were just kind of pulling some of the best people from Meta in and recruiting a lot of the best people.
我们当时一边从 Meta 内部抽调一些最优秀的人才，一边招募许多顶尖人才。
And I mean, you were asking, it's like, all right, well, like, okay, for people who, like, if you were asking before, it's like, well, Alex hadn't necessarily trained models like this.
你刚才问到了这一点。例如，正如你之前问到的，Alex 此前未必训练过这样的模型。
But to me, one of the big things was like all of the best people were very excited to work with him.
但对我来说，一个很重要的因素是，所有最优秀的人都非常期待与他共事。
And I think like, it's, I mean, you, I'm sure see this in all of, in a lot of the things that you do, which is like, you know, it's like you, most people are not kind of one-dimensional, right?
我相信，你在自己从事的许多工作中也会看到这一点：大多数人都不是单一维度的。
Yeah, maybe you do a specific thing, but your ability to grow and expand is kind of largely based on how effectively can you just get people who are excellent in different domains want to come work with you on different things.
也许你从事的是某项特定工作，但你成长和拓展的能力，在很大程度上取决于你能否有效地吸引那些在不同领域都很出色的人，让他们愿意与你一起开展不同的工作。
And so I think that there was a little bit of that, but also just It's even for the people that don't know you yet, or like, or aren't, you know, don't know what you're about.
所以我认为，其中确实存在这一因素。但除此之外，即使面对那些还不认识你，或者不太了解你为人的人，
Like, are you like an impressive person and a compelling person that people want to come work with?
你是否是一个令人印象深刻、富有吸引力，并且能让大家愿意与你共事的人？
We were kind of— Some of the stuff I think you kind of had fully formed as beliefs of how you thought we should do this.
我觉得，当时有些事情你已经形成了非常成熟的理念，清楚我们应该如何推进。
But I also remember when we were having the initial recruiting conversations, we were sort of in real time shaping— it's like, okay, well, this construct that we're talking about with this person isn't quite going to work, but like, they seem good, and like, they have this thought.
但我也记得，在最初开展招聘沟通时，我们仍在边做边实时调整。例如，我们会发现，正在讨论的构想可能不适合某个人，但对方看起来很优秀，而且也有自己的想法。
And it was just like, like, really kind of coalescing this vision about what the shape of it should be in real time.
当时，我们确实是在逐渐凝聚这件事应当以何种形态呈现的愿景，而且这一愿景是在实践中实时形成的。
this, where it was like there were some of the initial concepts of like, I mean, I think probably the most extreme thing was I think you had the concept of the talent density and all and that kind of piece.
当时已经有了一些初步构想。我觉得，其中可能最极端的一点，就是你提出的人才密度等相关构想。
But I think we weren't sure upfront that it was going to have to be a full reboot, right?
但我觉得，我们一开始并不确定是否需要彻底重启。
It's like that basically started when we were out talking to a lot of the kind of best people.
这一过程基本上始于我们与许多最优秀的人交流。
And we realized that, hey, like, in order to make this as attractive as possible, it should feel like, okay, we're starting this thing again from scratch.
随后我们意识到，要让这件事尽可能具有吸引力，就应当让人感到我们准备从头再来一遍。
And yeah, we'll pull in a lot of the great people from the Llama effort and FAIR and all of that too.
当然，我们也会吸纳许多参与 Llama 项目的优秀人才，以及来自 FAIR 等各方的人才。
But that was a very intentional thing that we basically figured out and that you figured out as we were out talking to the initial set of people of like what was going to make this compelling.
不过，这个决定是经过深思熟虑才作出的。基本上，这是我们在与最初那批人交流时发现的，也是你在这一过程中摸索出来的：怎样才能让这件事真正具有吸引力。
And it ended up being really smart and like making a huge difference in the culture of the lab and how it got and how it got started.
事实证明，这个决定非常明智，对实验室的文化及其建立和起步方式产生了巨大影响。
But I don't know, it's like being able to be that dynamic about figuring out, like, what does the shape of this need to be?
不过，关键在于能够灵活判断这件事究竟需要采用何种形态。
So, when I was thinking about, like, all right, like, this field is going to keep on evolving, right?
所以，我当时就在想，这个领域还会不断演变。
So, what is the thing that's going to be important for someone who's going to run a lab well?
那么，对于一个要管理好实验室的人来说，什么才是重要的？
It's not like, are they an expert in whatever the latest modality is of, like, or kind of of training technique for a lab.
重点并不在于这个人是否精通最新的模态或实验室最新的训练技术。
It's, is the person a clear thinker and able to kind of dynamically update and move in, like, as kind of an entrepreneur would?
真正重要的是，这个人能否清晰地思考，并像创业者一样灵活地更新认知、采取行动。
Has this experience changed your view on the value of going external, where both your reboot, a bunch of the people you brought in were external, like Alex and Nat and Daniel and folks.
这段经历是否改变了你对从外部引入人才和观点的价值的看法？在这次重启中，你引入的许多人都来自外部，例如 Alex、Nat、Daniel 等人。
But then also you're describing deciding whether you needed a reboot based on going to the outside intelligence of the aggregate views of the researchers in the industry and getting a sense for the— But just Do you feel like you'll carry forward that approach for future unrelated projects?
你刚才还提到，你们在判断是否需要重启时，借助了外部的集体智慧，也就是通过汇总业内研究人员的观点来了解情况。你认为自己今后在其他不相关的项目中也会沿用这种做法吗？
I mean, obviously you can't— a lot of what we had to do here was rotate a lot of the rest of the company.
显然，我们不可能轻易做到这一点。这里要做的许多工作，就是推动公司的其他许多部门转向。
And like for things like the infrastructure team that now, you know, it's like building these huge clusters to support this.
例如，基础设施团队需要搭建这些超大规模集群，为此提供支持。
I mean, we brought in this guy Kunal Shah, who, you know, it's like, I think he's like one of the top entrepreneurs in India, building this great company CRED, and And now he's running WhatsApp, which I always thought should do more in commerce, and we were sort of underserved there, and India is the biggest country that we're in.
我们请来了 Kunal Shah。我认为他是印度最顶尖的企业家之一，创办了 CRED 这家出色的公司。现在，他开始负责 WhatsApp。我一直认为 WhatsApp 应该在商业领域开展更多业务，而我们在这方面做得还不够。印度又是我们业务覆盖的最大国家。
But, I mean, that's— it's a bit of a different situation, because we're not, like, pivoting the whole company towards WhatsApp.
但这种情况仍然有所不同，因为我们并不是要让整个公司都转向 WhatsApp。
I don't want to talk about it in the past tense, as if it's done, because it's still a level 10 out of 10 difficulty thing. Yeah, yeah, sure.
我不想用过去时谈论这件事，仿佛它已经完成，因为它现在仍然是一件难度达到 10 分满分中 10 分的事。当然如此。
But I think the number of pieces that kind of had to work, it was quite all-encompassing.
但我认为，需要协同运作的环节非常多，几乎涵盖了方方面面。
And I think that it's both difficult for the people doing it, but I do think that that also puts some level of strain on the rest of the company, on the financial model of the company.
我认为，这对参与其中的人来说很困难；但我也确实认为，它在某种程度上给公司的其他部门以及公司的财务模式带来了压力。
It's like I think there's yeah, it's good to have new blood, and it's good to have people who understand the context.
我认为，注入新鲜血液固然是好事，但拥有了解背景情况的人也很重要。
And you know, you were talking about people who've been around Meta for a while and understand the ups and downs.
你刚才提到了一些已经在 Meta 工作一段时间、了解公司一路上起伏经历的人。
And it's like I don't know if you've worked at Meta for 10 years, you've seen some shit, right?
如果一个人在 Meta 工作了 10 年，肯定也见过不少糟糕的事情。
It's like it's like we've gone through a lot of hard times, and and like I think that that's kind of important context too.
我们经历过许多艰难时期，而我认为这些背景也相当重要。
trying to, you know, you're— whenever you're building a team, you're— you inherently, you know, it's like Alex had a theory for what the team shape needed to be, and like, I guess I had a theory for what like the shape of the leaders needed to be for it.
尝试去……每当组建团队时，都需要采用某种架构。Alex 对团队需要采用怎样的架构有一套设想，而我对领导层需要采用怎样的架构也有一套设想。
Like, it's like, first of all, it was just good to try something different, right?
首先，尝试一些不同的做法本身就是一件好事。
Yes. And yeah, I mean, that's— I think there's some other dimensions here that are worth mentioning too.
是的。我认为这里还有几个方面值得一提。
Like, one of the— like, when deciding we had to do the full reboot, like, you know, a lot of it was, you know, one of the main inputs was external, but also one of the inputs— I don't think I've ever— actually, I can't remember if I told you this story, but one of the first meetings I had When I joined, I was in this meeting with some of the leads of the existing effort.
例如，在决定必须进行全面重启时，外部因素在很大程度上是主要考量之一。另一个决策依据来自我刚加入时参加的一场会议。我不记得是否向你讲过这件事：那是我最早参加的几场会议之一，与会者包括当时负责现有工作的几位负责人。
And I was like, okay, I think the first thing we need to do is train a really good reasoning model by some timeframe.
我提出，我们首先要做的，是在某个时间点之前训练出一个非常优秀的推理模型。
I don't know when, but this is clearly the most important milestone.
我当时不知道具体应该定在什么时候，但这显然是最重要的里程碑。
And one of the key leads in the group who no longer works at Meta, but was like, oh, we don't need a reasoning model.
团队里的一位核心负责人当时表示，我们不需要推理模型。他现在已经不在 Meta 工作了。
But I think it was specifically an indicator of something that I think, like, Mark, when he talked to people, like, certainly also started to realize, which is like, we needed leadership in the company to be much more AI-pilled and much more sort of like forward-leaning and front-leaning on like where the— how quickly the technology's moving, where it's going to go to, what the possibilities are.
这件事尤其反映出一个问题。Mark 在与大家交流后，显然也开始意识到，公司领导层需要更加笃信人工智能，也需要更积极主动地走在前面，真正跟上技术发展的速度，看清它将走向何方，以及存在哪些可能性。
And I think as Mark started pivoting and shifting the whole company towards this direction, and it was very clear that this is like, I think Mark got AI-pilled, and then everyone saw that he was AI-pilled and everyone realized, like, oh, this is where the company's going.
随着 Mark 开始转向，并推动整个公司朝这个方向转型，大家都清楚地看到，他已经完全笃信人工智能，也由此意识到，这就是公司接下来的发展方向。
Then I think you had a lot of the existing leaders like very quickly pick it up and change.
此后，许多原有的领导者也很快接受了这一点，并做出了改变。
And like Adam Mosseri, for example, who runs Instagram, has one of the most like aesthetic but also like effective kind of like personal vibe code and personal mini app setup of anyone that I've seen.
例如，负责 Instagram 的 Adam Mosseri 所采用的个人 vibe coding 和个人迷你应用配置，是我见过最具美感、同时也最实用的方案之一。
Now, AI is, like, maybe, you know, 5 or 10 times more turbulent and faster moving than those past waves.
当然，人工智能带来的动荡程度和发展速度，可能是过去几波浪潮的 5 倍或 10 倍。
And so I think there's, like, I've been very impressed, actually, by seeing how quickly a lot of Meta leadership If I compare the kind of conversations I would have with them when I first joined versus the kind of conversations I have now— They're now super AI-pilled. Yeah.
坦率地说，这令我印象非常深刻，因为我看到 Meta 的许多领导者迅速适应了这一变化。对比我刚加入时与他们交流的内容和现在的交流内容，可以发现他们如今也完全成了人工智能的拥趸。
And it's very impressive, honestly, for that to happen at the large scale of Meta, which is one of the largest companies in the world.
坦率地说，Meta 能在如此大的规模上实现这种转变，确实非常了不起，而 Meta 是全球最大的公司之一。
It was, yeah, but I mean, like you're saying, it's like if you start coding, then it's helpful.
不过，正如你所说，如果开始写代码，就会很有帮助。
But I think a big part of what, a big part of the culture that we were trying to create is also, you know, when we were talking about how you have all these direct reports, the point of it isn't that you have the direct reports, it's that there aren't, it's the absence of non-technical managers, right?
不过我认为，我们当时想要打造的文化还有一个非常重要的部分。刚才谈到你拥有这么多直属下属时，重点并不在于直属下属的数量，而在于这里没有不懂技术的管理者。
Alex, well known for— Career chats and emotional, how are you doing, check-ins.
Alex 最为人熟知的就是与人讨论职业发展、关心他人的情绪，以及询问“你最近怎么样”之类的问题。
But I think it's just well known among researchers that, you know, there's this common pattern.
不过我认为，研究人员都很清楚有这样一种常见模式。
It's like you go from being very technical to you go to a role where you're, you're not like in the details anymore, and then you very quickly become not technical or don't kind of understand it much faster than you probably realize.
一个人起初技术能力很强，后来转到另一个岗位，不再深入研究技术细节，很快就会变得不再那么懂技术。换言之，他对技术的理解会以远超自己意识的速度退化。
I think that, like, this whole thing of just, like, you have to stay technical, people have to use the agents, people should be writing code. Yeah.
我认为，整件事的关键在于必须始终保持技术能力。每个人都必须使用智能体，也都应该写代码。
I think that that's— that has been a really important part of this whole thing.
我认为，这始终是整件事中非常重要的一部分。
It's like one of his— I mean, what this is like, I think it goes back to your your philosophy, you're talking about this for recruiting back at Scale, is like hire people who really care, right?
这算是他的一个特点。我认为，这可以追溯到你的理念，也就是你之前提到的：Scale 在招聘时，要招那些真正上心的人。
I think that's like one of the things that I thought was, you know, I think that that's like really important, right?
我当时认为，这一点确实非常重要。
It's, so I think part of what you kind of get when you like have the supernova of like all these different talented but very strong-willed people together is like, you're going to have like fun fights, um, and not fun fights, but like, but it's like to some degree I'd rather have people push for things that they believe in and like argue over things than, you know, just have people think that they're like blindly
所以，当你把这么多各有才华、同时又非常有主见的人聚在一起，引发一场超新星爆发时，难免会出现一些有意思的争吵，也会出现一些不那么有意思的争吵。但在某种程度上，我宁愿大家为自己相信的事情据理力争，为各种问题展开争论，也不愿让大家觉得自己只是在盲目地
Yeah, no, it's No, I think that like the— I think it has been very— one of the very cool things about like working with Mark over the past year has just been watching like the true degree of his boldness and like, and this trait that we were talking about before of like, you know, once he internalizes certain beliefs, like the— speed and magnitude of willingness to change prior actions and pivot towards the worldview that is
不是。我觉得，过去一年与 Mark 共事，有一点尤其令人赞叹：我得以亲眼看到他究竟有多么大胆，以及我们之前谈到的那种特质。一旦他真正接受了某些信念，他就会以惊人的速度和力度采取行动，改变此前的做法，转向一种
I think these are pretty unusual traits, generally speaking, even among this base of founders.
总体而言，我认为这些特质相当罕见，即使在这群创始人当中也是如此。
Mark actually ingests new information and fully updates his worldview and then acts in accordance with the newly installed worldview.
Mark 确实会吸收新信息，彻底更新自己的世界观，然后按照这个重新构建的世界观采取行动。
You're not allowed to talk about posteriors anymore after the whole incident. Yeah.
出了那档子事之后，你就不许再谈后验了。
And I think that, like, the— this is— it was, like, actually quite puzzling to me at first, the degree to which you were, like, willing to make these massive bets.
其实我一开始很困惑：你居然愿意下这么大的赌注。
And so I think that this, I think it's like, I mean, that is quite a notable trait and something I really admire.
我认为，这确实是一个非常突出的特质，也是我尤其钦佩的一点。
And I think is very, I actually, if I look back at the history of Meta and I imagine what the scenario would've been in all these pivotal moments, I can see how like this is kind of the movie, so to speak.
事实上，回顾 Meta 的历史，想象所有这些关键时刻当时可能的情形，我能够看出，打个比方，这大概就是那部电影的剧情。
Or like this is kind of how a lot of those situations go where it's like there's— I feel like probably the way I imagine is like there clearly is moments where like something very, very big is happening.
很多情况大概都是这样发展的。在我的设想中，显然会有一些时刻，某件极其重大的事情正在发生。
And it may not even really be showing in the metrics or it may not be impacting user counts or impacting revenue or whatever it might be.
这件事甚至可能完全没有体现在指标上，也可能没有影响用户数量、收入或其他任何方面。
But clearly, if you believe that thing is going to happen, then it has massive implications for the company and the product.
但如果你相信那件事将会发生，它显然会对公司和产品产生极其重大的影响。
And I think my post hoc retelling of Meta's history is that I think Meta is a company where every time that happens, Meta takes it extremely, extremely seriously.
事后复盘 Meta 的历史可以发现，Meta 是这样一家公司：每当这种事情发生时，Meta 都会极其认真地对待。
And that has enabled the company to innovate and continue evolving with a very fast-moving and fast-changing tech landscape.
这也使公司能够不断创新、持续演进，跟上瞬息万变、快速发展的科技格局。
And I think AI is, I mean, I think AI is maybe that, like, you know, times 100 or whatnot.
人工智能可能正是这样的事情，只不过其影响被放大了 100 倍。
But, like, I think the, you know, a lot of CEOs would've told you last, you know, last summer, oh yeah, I think AI is the most important thing for my business.
去年夏天，很多 CEO 可能都会告诉你：“人工智能对我的业务而言是最重要的事情。”
Like, I think maybe every CEO would say that, But then you look at, like, what they're actually doing and how consistent are their actions with that belief.
可能每位 CEO 都会这样说，但还要看他们实际上采取了什么行动，以及这些行动与其信念究竟有多一致。
And I think this is, like, a, like, quite, quite special thing about Meta and Mark.
这正是 Meta 和 Mark 极其特别的一点。
What do you make of all of Alex's deranged posting on Twitter?
你如何看待 Alex 在 Twitter 上那些疯疯癫癫的发帖？
I mean, I think most of it is great, but, but I, I guess my main view— But it doesn't matter how How much you fail, it matters how win— how big you win when you win.
我认为大部分内容都很棒。我的主要看法是：失败多少次并不重要，重要的是获胜时能赢得多大。
I think with— so I think with communication, I think it's most important to just be authentic.
因此，我认为沟通中最重要的是真实地做自己。
And I think when you're trying to be authentic, some of the things don't come out the way you want.
当你努力真实地做自己时，有些话说出来的效果并不会像你预想的那样。
And I think that that is a— that it is better to make the trade in that direction than it is to be overly scripted and not be authentic.
但我认为，选择这个方向，总比一切都照本宣科、失去真实感要好。
I mean, it's like a bunch of the researchers are there and we want to be over there, but everything I post there, I also post on Threads and it's all good.
很多研究人员都在 X 上，我们也想去那里。不过，我在 X 上发布的所有内容也会同步到 Threads，这样很好。
But every time I load X, I have like 4 Alex posts and it just like cracks me up.
但我每次打开 X，都会刷到差不多 4 条 Alex 的帖子，每次都把我逗得不行。
Well, you're now a clearinghouse, I presume, where people send you good stuff.
我想，你现在已经成了一个信息集散地，大家都会把好东西发给你。
Um, most of the— like, uh, we were, we were talking about this with my partner yesterday, and he was, you know, I think one observation was just, um, there was actually probably still to this moment, I— my, my, uh, thought pattern oftentimes is, oh, that's funny.
大部分情况是这样的：昨天我和我的合伙人讨论过这个问题。我认为，其中一个观察是，即使到了现在，我的思路往往仍然是：“这个真好笑。”
That being said, I have experienced the sort of like brutal metabolism of the internet and how quickly it like metabolizes memes.
话虽如此，我也体验过互联网近乎残酷的消化速度，以及它消化梗图的速度有多快。
And so the sort of like, you know, you really have to stay on the cutting edge of memeing to—
所以，你必须始终走在玩梗的最前沿，才能——