章节 1：AI发展的现状与质疑

📝 本节摘要：

本节作为访谈的开篇，主持人Alex回顾了过去一年中业界关于“大语言模型（LLMs）是否已触及天花板”的质疑。Google DeepMind CEO Demis Hassabis对此作出回应，表示DeepMind内部从未有过此类担忧。他反驳了“数据枯竭”将导致停滞的观点，指出通过合成数据以及优化现有架构，AI技术仍有巨大的提升空间（Headroom），目前的进展远未停止。

[原文] [Alex Kantrowitz]: google DeepMind CEO Dennis Asabis joins us to talk about the path from here to AGI when Google's AI glasses are coming and whether the pace of AI progress can keep up at this rate that's coming up right after this welcome to a special edition of Big Technology Podcast from Davos i'm Alex Caneritz and I'm joined today by a special guest Dennis Estabis the CEO of Google Deep Mind demis welcome back to the show

[译文] [Alex Kantrowitz]: Google DeepMind CEO Demis Hassabis（德米斯·哈萨比斯）做客本期节目，我们将探讨从现状通往AGI（通用人工智能）的路径、Google智能眼镜的发布时间，以及AI的发展速度能否保持当前的频率。精彩内容稍后即呈。欢迎来到在大达沃斯（Davos）录制的《Big Technology Podcast》特别版，我是Alex Kantrowitz。今天我邀请到的特别嘉宾是Google DeepMind的CEO Demis Hassabis。Demis，欢迎回到节目。

[原文] [Demis Hassabis]: it's great to be here

[译文] [Demis Hassabis]: 很高兴来到这里。

[原文] [Alex Kantrowitz]: a year ago there were real questions about whether AI progress was tailing off it was in fashion to ask whether LLMs were going to hit a wall and those questions seem like they've been settled there's been a tremendous amount of progress over the past year uh could you tell us what specifically has happened that's gotten the AI industry from that moment of question last year to the point that it is today

[译文] [Alex Kantrowitz]: 一年前，人们曾真正质疑AI的进步是否正在放缓；当时很流行讨论大语言模型（LLMs）是否会遭遇瓶颈（hit a wall）。而如今这些问题似乎已经有了定论，过去一年里取得了巨大的进步。你能否具体告诉我们，发生了什么让AI行业从去年的质疑时刻发展到了今天的地步？

[原文] [Demis Hassabis]: well I'm for us internally we were never questioning that just to be clear I think we we've always been seeing um great improvements um so we were a bit puzzled by by why there was this question in in the air i mean some of it was to do we people were worried about data running out um and there is you know some some truth in that is all the data had been used can we create synthetic data that's going to be useful to learn from but actually it turns out you can ring more uh more juice out of the existing architectures and data so there was plenty of room I think and we're still seeing that in both the pre-training the post-training and the thinking paradigms and also the way that they all kind of uh uh fit together so I think there's still plenty of headroom there just uh with the techniques we already know about and and tweaking and and and kind of innovating on top of that,

[译文] [Demis Hassabis]: 嗯，对于我们内部来说，为了澄清一下，我们从未怀疑过这一点。我认为我们一直都在目睹巨大的改进，所以我们对当时空气中弥漫的这种质疑感到有些困惑。我的意思是，这部分原因在于人们担心数据会耗尽。这确实有一定道理，比如是否所有数据都已被使用？我们要不要创造有用的合成数据（synthetic data）来供模型学习？但实际上，结果证明你可以从现有的架构和数据中榨取更多的“汁水”（价值），所以其实还有很大的空间。我认为无论是在预训练（pre-training）、后训练（post-training）还是思维范式（thinking paradigms）上，以及它们如何相互结合的方式上，我们仍然看到了这种进步。所以我认为，仅凭我们已经掌握的技术，通过微调和在此基础上的创新，仍然有着充足的上升空间（headroom）。

章节 2：通往AGI的技术路径

📝 本节摘要：

在本节中，讨论转向了技术深水区。Alex提出了怀疑论者的观点，质疑现有的大语言模型（LLMs）是否仅仅依赖“脚手架”等技巧，而缺乏真正的记忆和持续学习能力。Demis坦诚地回应，认为要实现AGI，仅仅扩大现有模型规模可能不够，或许还需要一两个重大的突破，特别是在持续学习（Continual Learning）、记忆（Memory）和规划（Planning）方面。他重点介绍了Google DeepMind在神经符号系统（Neurosymbolics）和混合系统方面的探索（如AlphaFold），并强调“学习”是智能的同义词，而目前模型仍处于“金鱼脑”阶段，尚未解决在线持续更新模型的难题。

[原文] [Alex Kantrowitz]: all right here's what a skeptic would say that there have been a lot of tricks that have been put on top of LLMs um I hear often about scaffolding and orchestration and AI that can use a tool to search the web but it won't remember what it learns as soon as you close that session it forgets is that just a limitation of the large language model paradigm

[译文] [Alex Kantrowitz]: 好吧，怀疑论者可能会这样说：在大语言模型（LLMs）之上已经堆砌了太多的把戏（tricks）。我经常听到关于“脚手架”（scaffolding）和“编排”（orchestration）的说法，以及AI可以使用工具搜索网络，但它记不住自己学到了什么。一旦你关闭会话，它就全忘了。这是大语言模型范式本身的局限性吗？

[原文] [Demis Hassabis]: well look I think there is and I'm definitely a subscriber to the idea that maybe we need one or two more big breakthroughs before we'll get to AGI and I think they're along the lines of things like continual learning better memory longer context windows or or perhaps more efficient context windows would be the right way to say it so don't store everything just store the important things that would be a lot more efficient that's what the brain does um and better long-term reasoning and planning

[译文] [Demis Hassabis]: 嗯，听着，我认为确实存在这种情况。我绝对赞同这样一种观点，即在实现AGI（通用人工智能）之前，我们可能还需要一两个大的突破。我认为这些突破将体现在诸如持续学习（continual learning）、更好的记忆力、更长的上下文窗口——或者也许更准确的说法是更高效的上下文窗口，即不需要存储所有信息，只存储重要的信息，这样会高效得多，这也是人脑的运作方式——以及更好的长期推理和规划能力。

[原文] [Demis Hassabis]: now it remains to be seen whether just sort of scaling up existing ideas and technologies will be enough to do that uh or we need one or two more uh uh really big insightful innovations i'm probably if you were to push me I would I would be in the latter camp um but I think um no matter what camp you're in we're going to need large foundation models as the key component of the final AGI systems of that I'm sure so I don't I'm not subscriber to someone like Yan Lun who thinks you know that they're just sort of some kind of dead end i think the only debate in my mind is are they a a key component or the only component

[译文] [Demis Hassabis]: 目前还有待观察的是，仅仅扩大现有的理念和技术的规模（scaling up）是否足以实现目标，还是说我们需要一两个真正具有深刻洞察力的重大创新。如果你非要逼问我，我可能会属于后者（需要创新）那个阵营。但我认为，无论你属于哪个阵营，我们都将需要大型基础模型（foundation models）作为最终AGI系统的关键组件，这一点我是确信的。所以我并不赞同像Yann LeCun（杨立昆）那样认为这只是某种“死胡同”的观点。我认为我脑海中唯一的争论在于：它们是关键组件，还是唯一的组件？

[原文] [Demis Hassabis]: so I think it's between those two two options and and for me we this is one advantage we have of having such a deep and rich research bench we can go after both of those things at maximum with maximum uh force both you know scaling up uh the current paradigms and ideas and and and when I say scaling up that also involves innovation by the way um pre-training especially I think we're very strong on um and then really uh new blue sky ideas for new architectures and things you know the kinds of things we've invented over the last 10 years as Google and Deep Mind you know of course including transformers

[译文] [Demis Hassabis]: 所以我认为答案介于这两个选项之间。对我们来说，这是我们拥有如此深厚和丰富研究储备的一个优势，我们可以全力以赴地同时追求这两个方向：既扩大当前的范式和理念——顺便说一句，当我说“扩大规模”时，这也包含了创新，特别是在预训练方面，我认为我们要强得多——同时也真正追求新的、天马行空的（blue sky）新架构理念。你知道，就是那种Google和DeepMind在过去10年里发明的东西，当然也包括Transformer模型。

[原文] [Alex Kantrowitz]: can something with a lot of hard-coded stuff uh ever be be considered AGI

[译文] [Alex Kantrowitz]: 如果一个系统包含大量硬编码（hard-coded）的东西，它还能被视为AGI吗？

[原文] [Demis Hassabis]: no I think um well depends what you mean by a lot i think that uh I'm very interested in hybrid systems is what I would call them or neurosymbolics sometimes people call them you know AlphaFold Alph Go are examples of that so some of our uh most important work um combines neural networks and deep learning with things like Monte Carlo research so I think uh that could be possible

[译文] [Demis Hassabis]: 不，我觉得……这取决于你说的“大量”是指多少。我对“混合系统”（hybrid systems）非常感兴趣，或者有时人们称之为“神经符号系统”（neurosymbolics）。你知道，AlphaFold和AlphaGo就是这样的例子。我们一些最重要的工作结合了神经网络、深度学习与蒙特卡洛搜索（Monte Carlo search，原文口误为research）之类的技术。所以我认为这是可能的。

[原文] [Demis Hassabis]: and there's some very interesting work we're doing building using the LLMs with things like evolutionary uh uh methods alpha evolve uh to actually go and discover new uh knowledge um you may need something beyond what the existing methods do but I think learning is a critical part of uh a gen of AGI it's actually almost the defining feature uh when we say general we mean general learning can it can it learn uh new knowledge and can it learn across any domain that's the general part so for me learning is synonymous with intelligence and always has been

[译文] [Demis Hassabis]: 我们正在做一些非常有趣的工作，利用LLMs结合进化方法（evolutionary methods）——比如AlphaEvolve——去真正发现新知识。你可能需要一些超越现有方法的东西。但我认为“学习”是AGI的关键部分，实际上它几乎是定义性的特征。当我们说“通用”（General）时，我们指的是“通用学习”：它能否学习新知识？它能否跨领域学习？这就是“通用”的部分。所以对我来说，学习就是智能的同义词，一直如此。

[原文] [Alex Kantrowitz]: okay so if learning is synonymous with intelligence and these models still don't have the ability to continually learn no uh like I said earlier it has goldfish brain it can search the internet and it can be like I figured this out but it doesn't change the model it's just it will forget it after the session um do you have a theory as to how continual the continual learning problem can be solved and do you want to share it with us

[译文] [Alex Kantrowitz]: 好吧，如果学习是智能的同义词，而这些模型仍然不具备持续学习的能力——不，就像我之前说的，它只有“金鱼般的脑子”（记忆短暂），它可以搜索互联网，表现得像“我搞懂了这个”，但它并没有改变模型本身，只是在会话结束后就忘了——那么你对于如何解决这个持续学习问题有一套理论吗？你想和我们要分享一下吗？

[原文] [Demis Hassabis]: all i can give you some clues we are working very hard on it um we've done some work on you know I think the best work on this in the past with things like Alpha Zero you know that learned from scratch um versions of Alph Go alph Go Zero also learned on top of the the knowledge it already had so we've done it in much narrower domains you know games are obviously a lot easier than the messy real world so it remains to be seen if that those kinds of techniques will really scale and generalize to the to the real world and and actual real world problems

[译文] [Demis Hassabis]: 我只能给你一些线索，我们正在为此付出巨大努力。你知道，我认为过去在这方面最好的工作是像AlphaZero这样的项目，它从零开始学习；AlphaGo的某些版本（如AlphaGo Zero）也能在已有知识的基础上继续学习。所以我们在更狭窄的领域里已经做到了这一点。你知道，游戏显然比混乱的现实世界要容易得多。所以，这些技术是否真的能扩展并泛化到现实世界以及实际的现实问题中，还有待观察。

[原文] [Demis Hassabis]: um but at least the methods we know uh can do some pretty impressive things and so now the question is can we blend that at least in my mind with the uh these big foundation models um and so of course the foundation models are learning during training but we would love them to learn you know out in the wild and including things like personalization i think that's going to happen and I feel like that's a critical part of of of building a great assistant is that it it understands you and it works for you as technology that works for you... but I think to have it uh uh you want to do it more than just having your data in the context window that's uh you want to have something a bit deeper than that which is as you say actually changes the model over time that's what ideally you would have um and that technique has not been cracked yet

[译文] [Demis Hassabis]: 不过，至少我们知道这些方法能做到一些相当令人印象深刻的事情。现在的关键问题是——至少在我看来——我们能否将这些方法与大型基础模型融合在一起。当然，基础模型在（预）训练期间是在学习的，但我们非常希望它们能在“野外”（实际应用中）学习，包括个性化之类的东西。我认为这是会发生的，这也是构建出色助手的关键部分，即它理解你并为你服务……但我认为，要做到这一点，你不能仅仅把数据放在上下文窗口里，你需要比这更深层的东西，也就是正如你所说的，随着时间的推移真正改变模型。这是理想情况下应该有的，但这项技术目前尚未被攻克。

章节 3：定义AGI与超级智能

📝 本节摘要：

在本节中，Alex 引用了 OpenAI CEO Sam Altman 的观点，即 AGI（通用人工智能）定义模糊，人类可能在未察觉的情况下直接“掠过”它进入“超级智能”阶段。Demis 强烈反对将 AGI 变成模糊的营销术语，他提出了严格的科学定义：系统必须具备人类所有的认知能力，包括顶级的创造力（如提出广义相对论或开创艺术流派）以及物理智能（如机器人的身体控制）。他认为目前的系统距离这一目标还有 5 到 10 年，而“超级智能”则是指超越人类生理极限（如 14 维思考）的更高阶概念。

[原文] [Alex Kantrowitz]: we've brought up AGI a couple times um so let me let me put this to you because I was speaking with Sam Alman towards the end of the year and I asked him I was like you know you seem to be saying two things we're not at AGI yet but every time he talks about what GPT models can do it seems like it fits his definition and he said uh a that AGI is underdefined and what he wishes everybody could agree to was that we've sort of whooshed by AGI and we move towards super intelligence do you agree with that i'm sure he does wish that but it's um No absolutely not,

[译文] [Alex Kantrowitz]: 我们已经几次提到 AGI（通用人工智能）了，所以让我把这个问题抛给你。去年底我和 Sam Altman（山姆·奥特曼）聊过，我问他：“你看，你似乎在说两件事：我们要么还没达到 AGI，但每次他谈论 GPT 模型能做什么时，似乎又符合他的定义。”他说 AGI 的定义尚不明确，他希望大家能达成共识的是，我们已经某种程度上“呼啸而过”（whooshed by）了 AGI，正朝着超级智能（super intelligence）迈进。你同意这个观点吗？我相信他确实希望如此，但是……不，绝对不同意。

[原文] [Demis Hassabis]: i don't think AGI should be sort of turned into a marketing term or for commercial gain i think there is always been a scientific uh definition of that my definition of that is a system that um can exhibit all the cognitive capabilities humans can and I mean all so that means you know the the the the the kind of highest levels of human creativity that we always celebrate the scientists and the artists that we admire

[译文] [Demis Hassabis]: 我认为不应该把 AGI 变成某种营销术语或为了商业利益而模糊化。我认为它一直有一个科学的定义。我的定义是：一个能够展现人类所具备的所有认知能力的系统。我是指“所有”，这意味着包括我们一直推崇的科学家和艺术家所具备的那种最高水平的人类创造力。

[原文] [Demis Hassabis]: so it means you know not just solving a maths equation or a conjecture but coming up with a breakthrough conjecture that's much harder you know not solving something in physics or some bit of chemistry some problem even like alpha fold you know protein folding but actually coming up with a new theory of physics something like um you know like Einstein did with general relativity right can a system come up with that because of course we can do that the smartest uh humans with their brain archite brain architectures have been able to do that in history,

[译文] [Demis Hassabis]: 所以这意味着，不仅仅是解一个数学方程或猜想，而是提出一个突破性的猜想，这要难得多。你知道，不是解决物理学或化学中的某个小问题——甚至不像 AlphaFold 解决蛋白质折叠那样——而是真正提出一个新的物理学理论，就像爱因斯坦提出广义相对论那样，对吧？一个系统能做到这一点吗？因为我们当然能做到，历史上最聪明的人类凭借他们的大脑架构已经能够做到这一点。

[原文] [Demis Hassabis]: and the same on the art side you know not just create a pastiche of what's known but actually be Picasso or Mozart and create a completely new genre of art that we had never seen before right and today's systems in my opinion are nowhere near that um doesn't matter how many you know erdos problems you solve which for some reason you know I mean you know that it's good that we're doing those things but I think it's far far from what uh you know a true invention or someone like a ramen would have been able to do

[译文] [Demis Hassabis]: 在艺术方面也是如此。不仅仅是创作出已知事物的模仿作（pastiche），而是真正成为毕加索或莫扎特，创造出我们从未见过的全新艺术流派，对吧？在我看来，今天的系统离那个境界还差得很远。不管你解决了多少个厄多斯（Erdős）问题——当然，我们在做这些事情是好的——但我认为这与真正的发明，或者像拉马努金（Ramanujan，推测原文Ramen为口误）那样的人所能做到的事情相比，还差得很远。

[原文] [Demis Hassabis]: and you need to And you need to have a system that can potentially do that across all these domains and then on top of that I'd add in physical intelligence because of course you know we can play sports and control our bodies and to amazing levels the elite sports people that are walking around you know here today in Davos and um and we're still way off of that on robotics as another example so I think an AGI system would have to be able to do all of those things to to really fulfill uh the the the original sort of goal of of the AI field and I think you know we're 5 to 10 years away from that,

[译文] [Demis Hassabis]: 而且你需要一个能够跨所有这些领域做到这一点的系统。在此之上，我还要加上物理智能（physical intelligence）。因为你知道，我们可以进行体育运动，并以惊人的水平控制我们的身体，就像今天在达沃斯这里走动的那些精英运动员一样。而在机器人技术方面，我们离那个水平还差得很远。所以我认为一个 AGI 系统必须能够做到所有这些事情，才能真正实现 AI 领域的最初目标。我认为我们距离那个目标还有 5 到 10 年的时间。

[原文] [Alex Kantrowitz]: i think the argument would be that if something can do all those things it would be considered super intelligence but you think AGI is a good No of course not because the individual humans could we can come up with new theories einstein did Fineman did all all the all the greats that all my scientific heroes they were able to do that it's rare but it's possible with the human brain architecture

[译文] [Alex Kantrowitz]: 我想人们的争论点在于，如果某种东西能做完所有这些事，它就会被视为“超级智能”，但你认为这只是 AGI？

[Demis Hassabis]: 不，当然不是。因为个体人类是可以做到的。我们可以提出新理论，爱因斯坦做到了，费曼（Feynman）做到了，所有那些伟人，我的科学偶像们，他们都能做到。这虽然罕见，但凭人类的大脑架构是可能的。

[原文] [Demis Hassabis]: so super intelligence is another concept that's worth talking about but that would be things that can really go beyond what human intelligence can do we can't think in 14 dimensions or you know plug in weather satellites into our brains uh not yet anyway but um and so that that those are truly beyond human or superhuman and uh that you know that's a whole another debate to have but once we get to AGI,

[译文] [Demis Hassabis]: “超级智能”是另一个值得讨论的概念，但那指的是真正超越人类智能所能做到的事情。例如，我们无法在 14 个维度中思考，或者把气象卫星直接接入我们的大脑——至少目前还不行。这些才是真正超越人类或“超人”的能力。那是完全另一场辩论，但那是发生在我们达到 AGI 之后的事。

章节 4：世界模型与多模态理解

📝 本节摘要：

在本节中，Alex 提到了 Demis 曾在一个播客中令人惊讶地将一个图像生成器（代号 Nano Banana）视为接近 AGI 的系统。Demis 澄清并深入解释了为何视频生成模型（如 Veo）对 AGI 至关重要。他提出，视频生成实际上是对物理世界的模拟，要求模型具备“直觉物理学”（Intuitive Physics）和“世界模型”（World Models）的能力，即理解因果关系和物体行为。这种能力是实现长期规划（Long-term Planning）的基础，而规划能力正是目前 AI 与人类（如为了十年后的职业目标而学习四年）之间的巨大差距。最后，他强调了 多模态（Multimodal） 融合对于构建通用助手的重要性。

[原文] [Alex Kantrowitz]: once we get to AGI I was listening to you recently and something you said really surprised me you were asked um on the Google Deep Mind podcast which is a great listen if you have a system today that is close to AGI I thought it might be Gemini 3 you named Nano Banana yes the image generator yes what well you know sometimes you have to have these fun names and have fun with those and and you know but how is the image generator close to AGI

[译文] [Alex Kantrowitz]: 既然说到 AGI，我最近在听你的节目时，你说的一句话真的让我很惊讶。在 Google DeepMind 的播客上——顺便说一句，那个播客很好听——有人问你，如果你现在有一个接近 AGI 的系统，那会是什么？我原以为你会说是 Gemini 3，结果你提到了“Nano Banana”（纳米香蕉），是的，那个图像生成器。这是怎么回事？好吧，我知道有时候你们得取些有趣的名字来自娱自乐，但是，一个图像生成器怎么会接近 AGI 呢？

[原文] [Demis Hassabis]: oh well of course look let's take image generators but also uh let's talk about our video generator VO which is the state-of-the-art in video generation i think that's even more interesting in from an AGI perspective you know you can think of a video model that can generate you 10 seconds 20 seconds of a realistic scene it's sort of a model of the physical world intuitive physics we'd sometimes call it in physics land

[译文] [Demis Hassabis]: 噢，当然，听着，我们先不说图像生成器，咱们来谈谈我们的视频生成器 Veo，那是目前视频生成领域的最先进技术。我认为从 AGI 的角度来看，这其实更有趣。你可以把一个能为你生成 10 秒、20 秒逼真场景的视频模型，看作是某种物理世界的模型，在物理学领域我们有时称之为“直觉物理学”（intuitive physics）。

[原文] [Demis Hassabis]: and it sort of intuitively understood how uh liquids and and and and and objects behave in the world and that's um and obviously one way to exhibit understanding is to be able to generate it at least to the to the to the human eye being accurate enough to to be satisfying to the human eye obviously it's not completely accurate from a physics point of view and we're getting it we're going to improve that but it's it's it's steps towards having uh this idea of a world model a system that can understand the world and the mechanics and the causality of the world

[译文] [Demis Hassabis]: 它某种程度上直觉性地理解了液体、物体在世界中是如何表现的。显然，展示理解能力的一种方式就是能够生成它，至少在人眼看来足够准确、令人满意。当然，从严格的物理学角度来看它还不是完全准确的，我们在努力改进这一点。但这迈向了拥有“世界模型”（world model）这一概念的一步，即一个能够理解世界、理解其力学机制以及因果关系的系统。

[原文] [Demis Hassabis]: and then of course that would be I think essential for AGI because that would allow these systems to plan long-term plan in the real world um over perhaps very long time horizons which of course we as humans can do you know um I'll spend four years getting a degree so that I have more qualifications so that in 10 years I'll have a better job you know these are very long-term plans that we we all do quite effortlessly and at the moment without these systems we still don't know how to do we can do short-term plans over one time scale

[译文] [Demis Hassabis]: 当然，我认为这对于 AGI 来说是必不可少的。因为这将允许这些系统在现实世界中进行规划，进行长期规划，甚至可能是跨越非常长时间跨度的规划。这当然是我们人类能做到的。你知道，比如“我会花四年时间拿个学位，这样我就有更多的资质，以便在 10 年后能有一份更好的工作”。这些都是我们大家都能毫不费力地制定的长期计划，而目前如果没有这些（世界模型）系统，我们还不知道该让 AI 如何做到这一点。我们目前只能在一个时间尺度上做短期规划。

[原文] [Demis Hassabis]: um but I think you need these kind of world models and I think you imagine robotics that's exactly what you want for robotics is robots planning in the real world being able to imagine many trajectories from the current situation they're in in order to complete some task uh that's exactly what you'd want

[译文] [Demis Hassabis]: 但我认为你需要这种世界模型。你可以想象一下机器人技术，这正是你在机器人领域所需要的：机器人在现实世界中进行规划，能够从当前的处境出发，想象出许多种（行动）轨迹，以完成某项任务。这正是你想要的能力。

[原文] [Demis Hassabis]: uh and then finally from um our point of view and why this is why we've worked with Gemini as being multimodal from the beginning able to deal with you know video image uh and eventually converge that all into one model that's our plan is that uh it'll be very useful for a universal assistant as well

[译文] [Demis Hassabis]: 最后，从我们的角度来看——这也是为什么我们从一开始就把 Gemini 打造成多模态（multimodal）的原因，让它能够处理视频、图像，并最终将所有这些融合进同一个模型中——那是我们的计划。这对于构建一个通用助手来说也将是非常有用的。

章节 5：AI眼镜与下一代硬件形态

📝 本节摘要：

本节话题从软件转向硬件。Alex 提到在观看纪录片《The Thinking Game》时，看到 DeepMind 团队举着手机测试 AI 视觉功能，这显得非常笨拙，他认为智能眼镜才是正确的形态。Demis 对此完全认同，并指出手机并不是理想的载体（Form Factor）。他回顾了 Google Glass 过去的教训，认为当时除了硬件（电池、体积）不够成熟外，更关键的是缺少“杀手级应用”。如今，随着 Gemini 3 等模型的强大能力，通用数字助手（Universal Digital Assistant） 将成为智能眼镜的灵魂。Demis 透露他正亲自参与该项目，并提到了与 Samsung、Warby Parker 等品牌的合作，暗示新产品可能最早在今年夏天面世。

[原文] [Alex Kantrowitz]: so let's talk product a little bit uh I watched the documentary the thinking game along with 300 million other people um there was something kind of interesting that happened there uh throughout the documentary yourself and some colleagues uh kept pointing your phone at things and asking an assistant alpha what was going on and I was yelling at the computer as I usually do uh and said this guy needs glasses like he needs smart glasses to be able to do it the phone is the wrong form factor um what is your vision for AI glasses and when is the roll out happening

[译文] [Alex Kantrowitz]: 那么我们来谈谈产品吧。我和其他3亿人一样观看了纪录片《思考的游戏》（The Thinking Game）。里面发生了一些有趣的事情：在整部纪录片中，你和一些同事不断地用手机对着物体，问助手“Alpha”发生了什么。我当时对着电脑大喊——就像我通常做的那样——我说：“这家伙需要眼镜！”就像他需要智能眼镜来做这件事一样，手机是错误的产品形态（form factor）。你对 AI 眼镜的愿景是什么？什么时候会推出？

[原文] [Demis Hassabis]: yeah I think you're exactly right and that that was our conclusion it's very obvious when you sort of dog food these things and internally that as you saw from the film we were holding up you know you're holding up your phone to to uh get it to tell you about the real world and it's it's it's it's amazing it works but it's not the it's clearly not the right form factor for a lot of things you want to do you know cooking or you want roaming around the city and asking for directions or recommendations um or even helping the you know partially cited there's a huge I think use case there to help uh uh uh with those types of uh situations

[译文] [Demis Hassabis]: 是的，我觉得你说得完全正确，这也是我们的结论。当你“吃自家的狗粮”（dog food，意为内部测试自家产品）时，这一点非常明显。正如你在影片中看到的，我们举着手机，让它告诉你现实世界的情况。这很神奇，它确实能工作，但对于你想做的很多事情来说，这显然不是正确的产品形态。比如做饭，或者你在城市里漫游并询问方向或推荐，甚至帮助视力受损的人（partially sighted）——我认为在帮助这类人群方面存在巨大的应用场景。

[原文] [Demis Hassabis]: and um and for that I think you need something that's hands-free and the obvious thing is for those of us anyway that wear glasses like me is is to put it on glasses but there may well be other devices too i'm not sure that glasses is the final form factor but it's definitely it's obviously a clear next form factor and of course at Google and Alphabet we have a long history with glasses and uh maybe we're a bit too early in the past but I think the my analysis of it and talking to the people working on that project was a couple of things the the form factor was a bit too chunky and clunky and the battery life and these kind of things which are now more or less solved

[译文] [Demis Hassabis]: 对于这些场景，我认为你需要某种解放双手（hands-free）的设备。对于像我这样本来就戴眼镜的人来说，显而易见的方案就是把它做在眼镜上。但也可能有其他设备，我不确定眼镜是否是最终的形态，但它绝对是、显然是下一个明确的形态。当然，在 Google 和 Alphabet，我们在眼镜方面有着悠久的历史。也许过去我们做得太早了。但我对此的分析，以及与该项目工作人员的交流得出的结论是两点：一是之前的形态有点太厚重笨拙（chunky and clunky），还有电池续航之类的问题——这些问题现在或多或少已经解决了。

[原文] [Demis Hassabis]: um but I think the thing it was missing was a killer app and I think the killer app is universal digital assistant that's with you helping you in your everyday life um and they're available to you on any surface on your computer on your browser on your phone but also on you know devices like glasses when you're uh uh you know walking around the the city and I think it needs to be kind of seamless um and kind of knows uh each of those contexts and understands each of those contexts around you

[译文] [Demis Hassabis]: 但我认为当时缺失的是一个“杀手级应用”（killer app）。我认为这个杀手级应用就是通用的数字助手，它伴随你并在日常生活中为你提供帮助。它们可以在任何界面上为你服务——在你的电脑上、浏览器上、手机上，也可以在像眼镜这样的设备上，当你漫步在城市中时。我认为这种体验需要是无缝衔接的，并且能够知道并理解你周围的每一个情境（context）。

[原文] [Demis Hassabis]: and I think we're close now especially with Gemini 3 i feel we finally got AI that is maybe powerful enough to make that a reality and we're you know it's one of the most exciting projects we're working on I would say and it's one of the things I'm personally working on is is making smart glasses really work uh and we hope to we've we've done some great partnerships with WBY Parker and Gentle Monster and Samsung to build uh these next generation glasses and you should start seeing that uh you know maybe by the summer

[译文] [Demis Hassabis]: 我认为我们现在已经很接近了，特别是有了 Gemini 3。我觉得我们终于有了足够强大的 AI 来实现这一点。这可以说我们正在进行的最激动人心的项目之一，也是我个人正在致力于的事情之一，就是让智能眼镜真正运作起来。我们已经与 Warby Parker（原文转录为WBY Parker）、Gentle Monster 和三星（Samsung）建立了很好的合作伙伴关系，以打造下一代眼镜。你应该会开始看到成果……也许在今年夏天。

[原文] [Alex Kantrowitz]: yeah wby Parker did have a filing that said that these glasses are coming out pretty pretty soon this year yeah and the the prototype design it depends how you know we're in prototype phase depends how quickly that advances but um I think it's going to happen very soon and I think it'll be you know a category a new category defining uh technology

[译文] [Alex Kantrowitz]: 是的，Warby Parker 确实有一份文件称这些眼镜今年很快就会推出。

[Demis Hassabis]: 是的，至于原型设计……这取决于——你知道我们处于原型阶段——取决于进展有多快。但我认为这很快就会发生，而且我认为这将是一种定义新类别的技术。

[原文] [Alex Kantrowitz]: given your personal involvement is it safe to say that this is a pretty important initiative for yeah it's one well yes but it's I mean I you know I like to it's it's not just as important obviously I like spending my own time on important things but I like to be at the push the most cutting edge thing and and and and that's often the hardest thing and picking interim goals and giving confidence to the and and also just sort of understanding if the timing is right

[译文] [Alex Kantrowitz]: 鉴于你的个人参与，可以说这是一项相当重要的举措吗？

[Demis Hassabis]: 是的，这是一个……嗯，是的。我的意思是，你知道，显然我喜欢把时间花在重要的事情上，但我更喜欢推动最前沿的事物。那通常也是最困难的事情——选择阶段性目标，给团队信心，以及判断时机是否成熟。

[原文] [Demis Hassabis]: and over the years I've been doing this the many you know the decades now um you know I've got quite good at doing that so I try to uh be at the the most cutting edge parts of um I feel I can make the most difference there so things like glasses robotics I'm spending time on and and world models right

[译文] [Demis Hassabis]: 这么多年来，几十年来我一直在做这个，我在判断方面已经做得相当不错了。所以我试着让自己处于那些最前沿的领域，我觉得我在那里能发挥最大的作用。所以像眼镜、机器人技术，以及世界模型，这些都是我正在花时间研究的领域。

章节 6：商业模式与信任危机

📝 本节摘要：

本节讨论了AI产品的商业模式与信任之间的张力。Alex 引用社交媒体上的犀利评论，质疑如果一家公司依赖广告盈利，是否意味着其离AGI（通用人工智能）还很远。Demis 回应称，目前对于 Gemini App 没有引入广告的计划。他强调了“信任”在AI助手中的核心地位：如果用户要与助手分享生活，必须确信助手是代表用户利益工作的。他警告广告模式可能会渗透并混淆推荐的动机，导致用户不信任。面对关于“2026年引入广告”的传闻，Demis 给予了明确否认。

[原文] [Alex Kantrowitz]: sure is the timing right for for ads um let let me set it okay Um there's been some news that Gemini might include ads there's been some news that some of your competitors might include ads um the funniest thing I saw about that on social media was someone who said "These people are nowhere close to AGI it's not going to be this world disrupting uh technology if the business model is advertising." Do you agree

[译文] [Alex Kantrowitz]: 当然，现在是引入广告的好时机吗？嗯，让我设定一下背景，好的。有些新闻说 Gemini 可能会包含广告，也有新闻说你们的一些竞争对手可能会包含广告。我在社交媒体上看到的最有趣的一条评论是有人说：“这些人离 AGI 还差得远呢。如果商业模式是广告，那它就不可能是那种颠覆世界的技术。”你同意吗？

[原文] [Demis Hassabis]: well it's interesting i think those are tells on you know I think actions speak louder than words going back to the original conversation we were having with you know Sam and others claiming AGI is around the corner um why would you bother with ads then so that is I think a reasonable question to ask

[译文] [Demis Hassabis]: 嗯，这很有趣。我认为这些都是迹象（tells）。你知道，我认为行动胜于雄辩。回到我们最初关于 Sam（Altman）和其他人声称 AGI 指日可待的对话……既然那样，那你为什么要费心做广告呢？所以我认为这是一个值得问的合理问题。

[原文] [Demis Hassabis]: um but I I I think uh look from our point of view we have no plans at the moment to uh to do ads if if you're talking about the Gemini app right uh specifically i think uh we are going obviously we're going to watch very carefully what uh you know the outcome of what uh Chhatri PT are saying they're going to do

[译文] [Demis Hassabis]: 但我认为……看，从我们的角度来看，我们目前没有计划做广告，如果你具体是指 Gemini App 的话。我想……显然我们会非常仔细地观察……你知道，ChatGPT（原文口误为 Chhatri PT）声称他们要做的结果。

[原文] [Demis Hassabis]: um I think it has to be handled very carefully because the dichotomy I see is that um uh uh uh if you if you want an assistant that works for you what is the most important thing trust okay so trust and security and privacy um because you want to share potentially your life with that assistant and you want to be confident that it's it's working on your you know behalf and uh and and with your best interests

[译文] [Demis Hassabis]: 我认为这必须非常小心地处理，因为我看到的二分法（dichotomy）是……如果你想要一个为你工作的助手，什么是最重要的？是信任。好的，所以是信任、安全和隐私。因为你可能想和那个助手分享你的生活，你想确信它是代表你在工作，并且符合你的最大利益。

[原文] [Demis Hassabis]: and so you know you got to be careful I think there are ways one could do it but you got to be careful that it doesn't the advertising model doesn't bleed into that and confuse the the user uh as to what you know what is this assistant recommending you and I think um you know that's going to be an interesting challenge uh in that space and that's what not to do

[译文] [Demis Hassabis]: 所以你知道，你必须小心。我认为有些方法可以做到这一点，但你必须小心，不要让广告模式渗透进去，从而让用户对这个助手推荐给你的东西到底是什么感到困惑。我认为……这将是该领域一个有趣的挑战，那就是“什么不该做”。

[原文] [Alex Kantrowitz]: and Sudar in a recent earnings call said there are some ideas within Google of the right way to approach this sure how do you approach advertising well you know that's still we're still brainstorming that but I think it's I think it's um there are also you know very interesting ways when if you think about glasses devices there are other revenue models out there okay

[译文] [Alex Kantrowitz]: Sundar（Google CEO）在最近的财报电话会议上说，Google 内部有一些关于如何正确处理此事的想法。

[Demis Hassabis]: 当然。

[Alex Kantrowitz]: 你们会如何处理广告？

[Demis Hassabis]: 嗯，你知道我们还在头脑风暴阶段。但我认为……如果你考虑到眼镜设备，其实还有其他的收入模式。

[Alex Kantrowitz]: 好的。

[原文] [Demis Hassabis]: um so you know it's going to be interesting to see i don't think we've made any strong conclusions on that but it's an area that uh needs very careful thought

[译文] [Demis Hassabis]: 所以看结果会很有趣。我认为我们还没有对此得出任何强有力的结论，但这是一个需要非常仔细思考的领域。

[原文] [Alex Kantrowitz]: just to get a definitive answer from you I think you've given it but I'm just going to do it one more time i read before we met Google has told advertisers in recent days is from last year uh that it plans to bring ads to its AI chatbot Gemini in 2026

[译文] [Alex Kantrowitz]: 为了从你这得到一个明确的答案——我想你已经给出了，但我还是想再问一次。我在我们见面前读到，Google 最近告诉广告商——这是去年的消息——计划在 2026 年将广告引入其 AI 聊天机器人 Gemini。

[原文] [Demis Hassabis]: nope we have no current plans that's all I can say that's pretty so y uh all right

[译文] [Demis Hassabis]: 不（Nope），我们目前没有计划。我只能说这么多。

[Alex Kantrowitz]: 这很清楚了。好的。

章节 7：行业竞争与编程能力

📝 本节摘要：

在本节中，Alex 将话题引向了 Google 的竞争对手 Anthropic 及其近期引发轰动的产品 Claude Code，特别是关于非程序员通过 AI 快速构建软件（如 CRM 系统）的案例。Demis 对此展现出开放的态度，向 Anthropic 的成果致敬，并分享了自己圣诞期间利用 Gemini 3 进行游戏原型开发的经历，表达了对 "Vibe Coding"（氛围编程） 浪潮的喜爱。他透露 Google 刚刚发布了内部极其抢手的 IDE "Anti-gravity"，并指出虽然 Anthropic 专注于单一的语言与编程模型，而 Google DeepMind 覆盖 多模态 和 世界模型，但这种专注确实值得赞赏，也推动了 Google 在该领域的持续进步。

[原文] [Alex Kantrowitz]: let's let's just you know keep going through some of your competitors uh anthropic um Claude Code Claude Code and Claude Co-work have caused a tremendous amount of buzz um it it is amazing to see what some people have done mhm um I saw a post from an ex Amazon executive who said that he built a custom CRM in a weekend or actually a day and a half mhm let's call it a weekend um what do you think about it and do you do you plan to have an answer to it

[译文] [Alex Kantrowitz]: 让我们继续聊聊你们的一些竞争对手……比如 Anthropic。Claude Code 和 Claude Co-work 引发了巨大的轰动。看到一些人做出的成果真是令人惊叹。我看到一位前亚马逊高管发的帖子，说他在一个周末——或者确切说是一天半——内构建了一个定制的 CRM（客户关系管理系统）。我们就当是一个周末吧。你怎么看这件事？你们计划对此做出回应吗？

[原文] [Demis Hassabis]: it's very exciting um and I think you know kudos to to to Anthropic i think they built a very good model there with claw code we're very happy with the the current coding capabilities of Gemini 3 it's very good at certain things like front-end work i've been using it over the Christmas to prototype games so it's amazing it's getting me back into programming i I love the whole vibe coding wave that's happening

[译文] [Demis Hassabis]: 这非常令人兴奋。我认为要向 Anthropic 致敬（kudos），我认为他们用 Claude Code（原文录音误识别为 claw code）构建了一个非常好的模型。我们对 Gemini 3 目前的编程能力也非常满意，它在某些方面非常出色，比如前端工作。圣诞期间我一直在用它来制作游戏原型，这太棒了，它让我重新回到了编程中。我喜欢现在正在发生的整个“氛围编程”（vibe coding）浪潮。

[原文] [Demis Hassabis]: i think it will open up uh the whole productivity space to designers creatives artists that maybe would have had to work with teams access to teams of programmers now they can probably do you know a lot more just on their own i think that's going to be amazing uh once that's sort of um out in the world in a more general way to you know create lots of new creative opportunities

[译文] [Demis Hassabis]: 我认为这将为设计师、创意人员和艺术家打开整个生产力空间。以前他们可能必须与团队合作，或者需要程序员团队的支持，而现在他们可能仅凭自己就能做更多事情。我认为这将会非常了不起，一旦这种能力以更普遍的方式推向世界，将会创造大量新的创意机会。

[原文] [Demis Hassabis]: um we're we're working on we're very happy with our work on code we got a lot you know we got more to do there we've just released anti-gravity our own uh IDE um which is very very popular we can't actually serve all the demand that we're seeing there um and we're pushing very hard on coding and tool use performance of Gemini

[译文] [Demis Hassabis]: 我们正在努力，我们对自己在这个代码方面的工作很满意。你知道，我们在那里还有更多工作要做。我们刚刚发布了 Anti-gravity，这是我们自己的 IDE（集成开发环境），它非常非常受欢迎，我们实际上甚至无法满足所有的需求。我们正在大力推进 Gemini 的编程和工具使用性能。

[原文] [Demis Hassabis]: but it's one thing that I think Anthropic have fully focused on you know they don't make image models multimodal models world models they just do you know coding and language models and um they're very very good at that and um you know we're pleased to be partnering on that on on the one hand and also it gives us something to push for uh to improve with our own models

[译文] [Demis Hassabis]: 但我认为这是 Anthropic 全力专注于的一件事。你知道，他们不做图像模型、多模态模型或世界模型，他们只做编程和语言模型。他们在这一块做得非常非常好。一方面我们很高兴能在这一领域与之并肩（即存在这种竞争关系），另一方面这也给了我们动力，推动我们去改进我们自己的模型。

章节 8：AI泡沫与经济崩盘论

📝 本节摘要：

在本节中，主持人 Alex 抛出了一个悲观的“三步崩盘论”，质疑如果大模型回报递减、且出现廉价模型（如Flash），会导致巨额基础设施投资变得毫无价值，最终引发经济连锁崩盘。Demis 承认这是一种“看似合理但不太可能”的情景。他坚定认为AI是人类史上最具变革性的技术，目前仍处于“能力过剩”（Capability Overhang）阶段，即现有模型能力远未被产品化（如AI收件箱）完全消化。关于“AI泡沫”，他坦承某些没有产品却融资数十亿的初创公司确实存在泡沫，但对于拥有实体业务的Google而言，无论市场是否存在泡沫，都处于必胜的有利位置。

[原文] [Alex Kantrowitz]: let's just talk broadly about the AI industry business um I have a theory for how this could all fall apart um and I want to run it by you so it's three-step a three-step process uh the first is that large language model training runs produce limited returns uh the second is that there are flash models like Gemini flash that run AI computing as cheap as search and then step three is that the massive infrastructure commitments uh that have been made become somewhat useless given those two factors and there is a cascading collapse that happens is that a legitimate worry

[译文] [Alex Kantrowitz]: 让我们以此宽泛地谈谈 AI 行业的生意。我有一套关于这一切可能如何崩溃的理论，我想让你听听。这是一个三步走的过程：第一步，大语言模型的训练运行产生的回报有限（递减）；第二步，出现了像 Gemini Flash 这样的“快闪模型”，使得 AI 计算成本像搜索一样便宜；第三步，鉴于前两个因素，已经做出的巨额基础设施承诺变得某种程度上毫无用处，从而发生连锁崩盘。这是一种合理的担忧吗？

[原文] [Demis Hassabis]: um I think it's a plausible possible scenario i don't think it's the I don't think it's the likely one in my opinion um I mean in my mind there's no doubt AI's gone already proved out enough I would say and and our work I think in things like science and and alpha fold and drug discovery that it's here to stay it's not like tomorrow oh like oh we found out AI doesn't work we've gone way we've blasted way past that so I think that's it's clearly going to be the most transformous technology in human history there's maybe a question mark about timelines is it 2 years or 5 years i mean either way it's very soon for something this transformative

[译文] [Demis Hassabis]: 嗯，我认为这是一个看似合理（plausible）的可能情景，但在我看来，这并不是最可能发生的情况。我的意思是，在我心中毫无疑问，AI 已经充分证明了自己。我想说，我们在科学、AlphaFold 和药物发现等方面的工作证明了它是会长期存在的。这不像明天我们突然发现：“噢，原来 AI 不管用。”我们已经远远超越了那个阶段。所以我认为这显然将成为人类历史上最具变革性的技术。也许关于时间线还有一个问号——是 2 年还是 5 年？但无论哪种情况，对于如此具有变革性的事物来说，这都非常快了。

[原文] [Demis Hassabis]: and I think we're still in the nent era of actually figuring out how to make use of it and deploy it because the technology is improving so fast i think there's a huge capability overhang actually of what even today's models can do that maybe even us as building those things don't fully know um so I think there's just a vast amount of product opportunities that we see and I think we're as Google only just started to scratch the surface now of actually uh natively sort of plugging these things in to our amazing existing products let alone building the new ones

[译文] [Demis Hassabis]: 而且我认为我们仍处于弄清楚如何利用和部署它的萌芽时代。因为技术进步太快了，我认为实际上存在巨大的“能力过剩”（capability overhang），即便是今天的模型能做的事情，也许连我们这些构建它们的人都还没完全搞清楚。所以我认为我们看到了海量的产品机会。而且我认为，作为 Google，我们在将这些东西原生植入我们需要惊人的现有产品中这方面，才刚刚触及皮毛，更不用说构建新产品了。

[原文] [Demis Hassabis]: you know AI inbox we've just started triing I mean who wants to do email admin I mean wouldn't we all love that to just go away that's my number one pain point for my work king day and um there's so many examples like that just just waiting to addressed i think you know agents in browsers um helping out with YouTube obviously we're now powering search with it so I think there's enormous opportunities

[译文] [Demis Hassabis]: 你知道，AI 收件箱（AI inbox）我们才刚刚开始试用。我的意思是，谁想做邮件管理这种琐事？难道我们不希望它直接消失吗？那是我工作日里的头号痛点。还有很多类似的例子等待解决。我认为，比如浏览器里的智能体（agents）、协助处理 YouTube 内容，显然我们现在也在用它驱动搜索。所以我认为机会巨大。

[原文] [Alex Kantrowitz]: and if you're talking about the AI bubble if that's the question you know as the AI bubble I think it's fine I mean seems like that's the question I'm very happy to answer it because I think um look my view is it's not binary when are we in a bubble not in a bubble I think parts of the AI industry probably are and other parts I think it remains to be seen

[译文] [Demis Hassabis]: 如果你在谈论 AI 泡沫，如果那是你的问题——即所谓的 AI 泡沫——我觉得没问题。我的意思是，既然这似乎是你的问题，我很乐意回答。因为我认为，看，我的观点是，这并不是非黑即白的——我们是在泡沫中还是不在泡沫中。我认为 AI 行业的某些部分可能确实处于泡沫中，而其他部分还有待观察。

[原文] [Demis Hassabis]: so I think some of the things are you know when you see seed rounds of tens of billions of dollars with of companies that basically have no product or research it's just some people coming together that seems a bit unsustainable to me in a normal market a bit frothy um on the other hand you know we're businesses like us we have massive underlying businesses that and and products that it's very obvious how AI would uh increase the efficiency or the productivity of using those products

[译文] [Demis Hassabis]: 所以我认为有些事情，比如当你看到某些基本上没有产品或研究的公司进行数百亿美元的种子轮融资，只是一群人聚在一起，这在我看来在正常市场中有点不可持续，有点泡沫（frothy）。另一方面，你知道，像我们这样的企业，我们拥有巨大的底层业务和产品，显而易见 AI 将如何提高使用这些产品的效率或生产力。

[原文] [Demis Hassabis]: and then it remains to be seen how popular the monetization of these new AI native products like chat bots glasses all of these things um we we'll have to see i think there will be enormous markets but they're yet to be proven out but from my my perspective you know running Google Deep Mind is um my job is to make sure that whatever happens with an AR bubble if it if it if it bursts or if there isn't one and it continues we win either way

[译文] [Demis Hassabis]: 至于那些新的 AI 原生产品——比如聊天机器人、眼镜等——其变现模式会有多受欢迎，还有待观察。我们得走着瞧。我认为会有巨大的市场，但它们尚未被完全证实。但从我的角度来看，作为 Google DeepMind 的管理者，我的工作是确保无论 AI 泡沫发生什么——如果它破裂了，或者如果没有泡沫并继续发展——无论哪种情况，我们都能赢。

[原文] [Demis Hassabis]: and I think we're incredibly well positioned as Alphabet uh in either case um and you know doubling down on existing businesses in the one case or being at the forefront and the frontier uh in in the in the bull case

[译文] [Demis Hassabis]: 我认为作为 Alphabet，无论在哪种情况下，我们的定位都极好。在一种情况下我们可以加倍投入现有业务；而在另一种“牛市”（bull case，指行情好）的情况下，我们则处于最前沿和先锋地位。

章节 9：人类的意义与适应性

📝 本节摘要：

随着AI逐渐渗透进知识工作领域，Alex 担忧人类会像被AI击败的围棋和星际争霸选手一样失去价值。Demis 用游戏界的现状反驳了这一悲观论点：国际象棋在Deep Blue击败人类后反而更受欢迎，新一代围棋选手通过向AI学习变得更强。他用“百米赛跑”做类比——尽管汽车跑得更快，我们依然为此喝彩。Demis 坚信人类作为“通用智能”和“工具制造者”，拥有无限的适应能力。但他同时指出，除了经济层面的工作问题，更深层的挑战在于如何在很多工作被自动化后找到新的“目的与意义”，这可能需要新时代的哲学家来共同探讨。

[原文] [Alex Kantrowitz]: going back to thinking game speaking of the way that this will impact the economy I started to feel bad for the opponents of your technology um Li Doll okay uh demoralized sure uh this guy Mana who played Starcraft beat your bot but realized that it's basically over for humans versus machines um now we're all up against this in some way as this stuff makes its way into knowledge work um I thought you were meaning our AI competitors them I'm okay with i don't feel sad about that so relentless progress of AI you mean the gamers gamers yeah you made me feel bad for gamers

[译文] [Alex Kantrowitz]: 回到《思考的游戏》（The Thinking Game）这部片子，说到这将如何影响经济，我开始为你们技术的对手感到难过。李世石（Lee Sedol，原文误听为Li Doll），好吧，他确实很沮丧；还有那个玩《星际争霸》的家伙 Mana，虽然他赢了你们的机器人一局，但也意识到人类对抗机器的时代基本结束了。现在，随着这些东西进入知识工作领域，我们在某种程度上都要面对这种局面。

[Demis Hassabis]: 我以为你是说我们的 AI 竞争对手呢，对他们我倒是无所谓，我不为他们感到难过。

[Alex Kantrowitz]: 不，我是说 AI 的无情进步……我是指那些游戏玩家。是的，你让我为玩家们感到难过。

[原文] [Alex Kantrowitz]: um you know but I I want to ask about this you know we're going to have the same situation uh with knowledge work that these these models that you know performed admirably against the world's best Starcraft and Go players are now starting to do our work and are we going to end up in the same position

[译文] [Alex Kantrowitz]: 嗯，你知道，但我想问的是这个。你知道，我们在知识工作方面也会面临同样的情况：那些在对抗世界顶尖《星际争霸》和围棋选手时表现令人钦佩的模型，现在开始做我们的工作了。我们会落得同样的下场吗？

[原文] [Demis Hassabis]: well look let me let's given given you brought up games as an example let's let's look at what's happened in games so chess we've had chess computers that are better since I was a teenager than Gary Kasparov in the '90s right they weren't general AI systems but they were you know deep blue chess is more popular than ever no one's interested in seeing computers playing computers we're interested in Magnus Carlson playing you know the top the other top chess players in the world

[译文] [Demis Hassabis]: 嗯，听着，既然你提到了游戏作为例子，让我们看看游戏界发生了什么。拿国际象棋来说，从我还是个青少年时起，90年代我们就有了比加里·卡斯帕罗夫（Garry Kasparov）更强的国际象棋电脑，对吧？它们不是通用 AI 系统，但它们是……你知道，Deep Blue（深蓝）。现在国际象棋比以往任何时候都受欢迎。没人有兴趣看电脑跟电脑下棋，我们感兴趣的是看马格努斯·卡尔森（Magnus Carlsen）对战世界上其他顶尖棋手。

[原文] [Demis Hassabis]: uh interestingly in Go um the best South Go player in the world is a South Korean and he was about 15 I think when Alpha Go match happened he's in his mid20s now and he's by far the strongest player there's ever been by the ELO ratings cuz he's learned natively young enough he was you know he's the first generation you could say that's learned with Alph Go knowledge in the knowledge pool and um you know he may actually be stronger than Alph Go was back then

[译文] [Demis Hassabis]: 有趣的是，在围棋界，目前世界上最好的韩国围棋选手（指申真谞），在 AlphaGo 比赛时大概 15 岁。他现在 20 多岁了，从 ELO 评分来看，他是迄今为止最强的选手。因为他学棋时足够年轻，可以原生地（natively）学习，你可以说他是第一代利用知识库中的 AlphaGo 知识进行学习的人。而且你知道，他实际上可能比当年的 AlphaGo 还要强。

[原文] [Demis Hassabis]: so I think and and we all still enjoy Starcraft and all the other all the other um computer games we enjoy Human Endeavor i think it's a bit more a bit similar to like we still love the 100 meters uh uh Olympic race um even though we have vehicles that can go way faster than Usain Bolt but you know we we don't you know that's that's a different thing right

[译文] [Demis Hassabis]: 所以我认为，我们仍然都在享受《星际争霸》和其他所有的电脑游戏。我们享受的是“人类的努力”（Human Endeavor）。我认为这有点类似于，我们仍然热爱奥运会的百米赛跑，即便我们拥有比尤塞恩·博尔特（Usain Bolt）跑得快得多的车辆，但你知道……那完全是另一回事，对吧？

[原文] [Demis Hassabis]: and so I think we have infinite capacity to adapt and um and uh and and sort of evolve uh with our technologies because why is that because we have we are general intelligences um that's the thing about is we are AGI systems we are obviously we're not artificial we're general systems and it's and and we are capable of inventing ing science and uh we're tool making uh uh animals that's what separates us humans from from the other animals is we're able to make tools all around modern civilization including computers and of course AI being the ultimate expression of computers that all has come from our human minds which were evolved for you know hunter gathering lifestyle

[译文] [Demis Hassabis]: 所以我认为我们拥有无限的适应能力，并能够随着我们的技术一起进化。为什么会这样？因为我们要知道：我们是“通用智能”（general intelligences）。这就是关键所在，我们就是 AGI 系统——当然我们不是人造的（Artificial），我们是通用系统。我们有能力发明科学。我们是制造工具的动物，这就是我们人类与其他动物的区别：我们能够制造工具。现代文明周围的一切，包括计算机，当然还有作为计算机终极表现形式的 AI，所有这些都源自我们原本是为了狩猎采集生活方式而进化出来的人类大脑。

[原文] [Demis Hassabis]: so it's kind of amazing we were able and it shows how general we are that we're able to get to the modern civilization we see around us today and we're talking about things like AI and you know science and physics and all these things and I think we'll adapt again

[译文] [Demis Hassabis]: 所以这真的很神奇，这也表明了我们是多么的“通用”，以至于我们能够建立起今天周围所见的现代文明，并且我们在讨论像 AI、科学和物理学这些东西。所以我认为我们会再次适应的。

[原文] [Demis Hassabis]: but there is an important question actually beyond the economics one about jobs and those things is purpose and meaning because we all get a lot of our purpose and meaning from the jobs we do I certainly do from the science I do so how does what happens when a lot of that is automated

[译文] [Demis Hassabis]: 但实际上，除了关于工作之类的经济问题外，还有一个重要的问题，那就是“目的与意义”（purpose and meaning）。因为我们都从我们所做的工作中获得了很多目的感和意义感——我从我所做的科学研究中肯定获得了这些。那么，当很多工作被自动化后，会发生什么呢？

[原文] [Demis Hassabis]: um I think you know that that's why I've been calling for you know I think we knew new new great philosophers actually and it will be a change to the human condition but I don't think it necessarily has to be worse i think we've it's like the industrial revolution maybe 10x of that but we'll have to adapt again and I think we'll find new um uh uh meaning and things and we do a lot of things already today that are not just for economic gain you know art extreme sports ex polaro exploration many of these things um and maybe we'll have much more sophisticated esoteric versions of those things in the future

[译文] [Demis Hassabis]: 嗯，我认为……这也是为什么我一直在呼吁，我认为我们实际上需要新的伟大哲学家。这将是人类生存状况的一次改变，但我不认为这必然会变得更糟。我认为这就像工业革命，也许是它的 10 倍，但我们将不得不再次适应。而且我认为我们会找到新的意义和事物。今天我们已经做了很多不仅仅是为了经济利益的事情，你知道，比如艺术、极限运动、极地探险（原文为 ex polaro exploration，推测为 polar exploration）等许多事情。也许在未来，我们会拥有这些事物更加复杂、深奥的版本。

章节 10：宇宙的本质是信息

📝 本节摘要：

随着访谈接近尾声，Alex 提出了一个极具哲学深度的问题：Demis 曾提出“信息”而非物质或能量，才是宇宙最基本的单元。Demis 在有限的时间内阐述了这一深刻理论：他认为生物系统本质上是抵抗熵（Entropy）的信息系统。甚至像山脉、行星这样的非生物，也是在经历某种物理“选择压力”后保留下来的稳定信息结构。他将这一理论连接到 DeepMind 的工作，解释了 AlphaFold 之所以能解决蛋白质折叠问题，是因为它学会了理解蛋白质结构的信息拓扑（Information Topology）和能量景观。只要掌握了这种视角，那些看似是大海捞针（needle in the haystack）的棘手问题——如治疗疾病、发现新材料和超导体——都会变得有迹可循。

[原文] [Alex Kantrowitz]: okay two minutes left i have two questions i don't know if we're going to get to both of them um let me ask the one that I want to know the answer most about uh in a recent interview you said that you have a theory that information is the most fundamental unit of the universe not energy not matter information yeah how

[译文] [Alex Kantrowitz]: 好的，还剩两分钟，我有两个问题。我不确定我们能不能两个都聊完。让我先问那个我最想知道答案的问题。在最近的一次采访中，你说你有一个理论，认为“信息”是宇宙最基本的单元，而不是能量，也不是物质，是信息。是的，这是怎么回事？

[原文] [Demis Hassabis]: well look I think if you look at energy I mean I don't know if we're going to cover this in 2 minutes but in in in energy energy and and and matter you can definitely I think a lot of people sort of think of them as isomeorphic with uh uh information but I think information is really the right way to understand the universe

[译文] [Demis Hassabis]: 嗯，听着，如果你看能量……我的意思是，我不确定我们在两分钟内能不能讲完这个。但在能量和物质方面，你绝对可以……我认为很多人某种程度上把它们看作是与信息同构（isomorphic）的。但我认为“信息”确实是理解宇宙的正确方式。

[原文] [Demis Hassabis]: so um if you think of biology and living systems we're information systems that are resisting entropy right we're trying to retain our structure retain our information in the face of you know a randomness that's happening around us

[译文] [Demis Hassabis]: 所以，如果你思考生物学和生命系统，我们实际上是正在抵抗熵（entropy）的信息系统，对吧？我们在面对周围发生的随机性时，试图保持我们的结构，保持我们的信息。

[原文] [Demis Hassabis]: and I think you can look at that uh uh you know in a larger physics scale so almost not just biology but things like you know mountains and and planets and asteroids they've all been subject to some kind of selection pressure not not Darwinian evolution but some kind of external pressure and the fact that they've been stable over a long amount of time means that that that information is kind of stable and meaningful so I think one could view the world in terms of its complexity information complexity

[译文] [Demis Hassabis]: 我认为你可以从更大的物理尺度上来看待这一点，不仅仅是生物学，还包括像山脉、行星和小行星这样的东西。它们都受到过某种“选择压力”的影响——不是达尔文式的进化论，而是某种外部压力。事实上，它们在很长一段时间内保持稳定，这意味着那些“信息”是某种程度上稳定且有意义的。所以我认为人们可以从复杂性、信息复杂性的角度来看待这个世界。

[原文] [Demis Hassabis]: and I think a lot of what we're doing with our the reason I'm thinking about all of that is because of things like alpha go and alpha fold especially alpha fold where you know we solved all the protein structures that are kind of known to science and how have we done that well because only a certain number of those in the kind of almost infinite possibilities of protein structures are stable and and those are the ones you've got to find

[译文] [Demis Hassabis]: 我认为我们正在做的很多事情……我之所以思考这一切，是因为像 AlphaGo 和 AlphaFold 这样的项目，尤其是 AlphaFold。你知道，我们解出了科学界已知的所有蛋白质结构。我们是如何做到的呢？因为在近乎无限的蛋白质结构可能性中，只有一定数量的结构是稳定的，而那些正是你需要找到的。

[原文] [Demis Hassabis]: so you've got to understand that topology uh that information topology and follow it and then suddenly these problems that seem to be intractable because you know how can you find the needle in the haststack actually become very tractable if you understand the energy landscape or the information landscape around that

[译文] [Demis Hassabis]: 所以你必须理解那种拓扑结构——那种“信息拓扑”（information topology）——并顺着它走。然后突然之间，这些看似棘手的问题——因为你知道，怎么可能在大海里捞针呢——实际上变得非常容易处理，只要你理解了围绕它的能量景观（energy landscape）或信息景观。

[原文] [Demis Hassabis]: and that's how I think eventually we'll solve u most diseases come up with new drugs new materials new superconductors with the help of AI helping us uh navigate that information landscape

[译文] [Demis Hassabis]: 这就是为什么我认为最终我们将解决大多数疾病，开发出新药、新材料、新超导体，这将通过 AI 帮助我们探索那个信息景观来实现。

章节 11：AlphaFold与科学开源

📝 本节摘要：

在本节中，Alex 提到了纪录片中的一个紧张时刻：Demis 坚持要立即发布 AlphaFold 的成果，而不是陷入繁琐的流程。Demis 解释说，AlphaFold 解决的是生物学中长达50年的“根节点问题”（Root Node Problem）。团队意识到，仅凭DeepMind一己之力无法挖掘其全部潜力，为了最大化对人类健康的影响，必须将其开源给全球科学界。如今已有300万研究人员使用该工具，未来的药物发现几乎都将离不开它。Alex 将其解读为“充满激情的小团队在大公司内部打破官僚主义”的隐喻，Demis 对此回应称，Google 本质上是一家推崇科学方法的公司，给予了他们巨大的支持。

[原文] [Alex Kantrowitz]: uh Dennis before we go I just want to wrap with this uh well Maybe maybe quickly this first one and then a big question at the end in the thinking game speaking of um health and AI there's this moment where there's a discussion in the lab about whether to release the results of AlphaFold and you kind of sit there adamantly and you're like why are we going through a process release it release it now talk a little bit about um the lesson from there

[译文] [Alex Kantrowitz]: 呃，Demis，在节目结束前，我想用这个问题收尾。嗯，也许快速过一下这第一个问题，最后再问一个大问题。在《思考的游戏》（The Thinking Game）中，谈到健康和 AI 时，有这样一个时刻：实验室里正在讨论是否要发布 AlphaFold 的结果。你坚定地坐在那里说：“我们为什么要走流程？发布它，现在就发布。”能谈谈那件事带来的教训吗？

[原文] [Demis Hassabis]: yeah well look we we started Alphafold to crack a unbelievably tough scientific challenge 50 year grand challenge uh of protein folding and protein structure prediction and the reason we worked on that and the reason we've put so much effort into it is we sort of thought it was a root node problem if we could solve it and put that out in the world it could be it could do amazing untold impact on things like human health and understanding of biology

[译文] [Demis Hassabis]: 是的，你看，我们要攻克 AlphaFold，是为了解决一个极其艰难的科学挑战——蛋白质折叠和蛋白质结构预测这一“50年重大挑战”。我们致力于此并投入如此多精力的原因在于，我们认为这是一个“根节点问题”（root node problem）。如果我们能解决它并将其推向世界，它可能会对人类健康和生物学理解产生惊人的、不可估量的影响。

[原文] [Demis Hassabis]: but we as a team no matter how talented or hard hard working we are we would only be able to scratch a surf a small tiny amount of that potential on our own it's clear so um in that case and in this case it was obviously the right thing to do to maximize the benefit to the world here to put it out there to the scientific massive scientific community uh to build on top of and use alpha fold

[译文] [Demis Hassabis]: 但作为一个团队，无论我们多么有才华或多么努力，仅凭我们自己只能挖掘出其潜力的冰山一角，这是显而易见的。所以在这种情况下——显然，为了最大化对世界的益处，正确的做法是把它发布给庞大的科学界，让他们在 AlphaFold 的基础上进行构建和使用。

[原文] [Demis Hassabis]: and it's been incredibly gratifying to see you know 3 million researchers around the world use it in their important research i think in future um all almost every single drug that's discovered from now on will probably have used alpha fold at some point in that process which is you know amazing for us and you know really that's what we do all the work we do for

[译文] [Demis Hassabis]: 看到全世界有 300 万研究人员在他们的重要研究中使用它，这真是令人难以置信地欣慰。我认为在未来，从现在开始发现的几乎每一种药物，在研发过程的某个阶段可能都会用到 AlphaFold。这对我们来说是惊人的，真的，这就是我们做所有这些工作的目的。

[原文] [Alex Kantrowitz]: I also read that moment you tell me if I'm wrong as something of a of a metaphor small uh passionate AI division kind of yelling in a big company get this out cut the red tape

[译文] [Alex Kantrowitz]: 我也把那一刻解读为——如果我错了请纠正我——某种隐喻：一个虽小但充满激情的 AI 部门在一家大公司里大声疾呼：“把这个发出去，砍掉繁文缛节（red tape）。”

[原文] [Demis Hassabis]: yeah potentially but look I mean we've had amazing support from the beginning from Google and they the reason that we you know we we joined forces with Google back in 2014 is Google itself is a scientific research engineering technical company always has been and has that at its core and that's why you know I think that we have the scientific method and the scientific approach that thoughtful approach that rigorous approach in everything we do so of course they're going to love something like Alpha Fold

[译文] [Demis Hassabis]: 是的，这有可能。但你看，我的意思是，我们从一开始就得到了 Google 惊人的支持。我们在 2014 年加入 Google 的原因是，Google 本身就是一家科学研究、工程和技术公司，一直如此，并以此为核心。这就是为什么我认为我们在做的每件事中都秉持着科学方法、那种深思熟虑和严谨的态度。所以，他们当然会喜欢像 AlphaFold 这样的成果。

==================================================

第 12 章

==================================================

这是基于Demis Hassabis访谈内容整理的第12章（最后一章）。

章节 12：未来的发现与AlphaZero时刻

📝 本节摘要：

在访谈的最后，Alex 提出了一个极具想象力的问题，将话题引向了终极未来。他回顾了 DeepMind 的历史：先是 AlphaGo 学习人类棋谱，随后 AlphaZero 摆脱人类知识束缚，自我进化出人类无法想象的新策略。他询问 Demis，当大语言模型（LLM）掌握了人类所有知识并被“松绑”后，是否也会发生类似的飞跃？Demis 充满激情地描绘了那个“AGI时刻”：届时 AI 将不再仅仅是模仿，而是通过类似 AlphaZero 的自我探索机制，进入人类未知的领域（Uncharted Territory），发现室温超导体、新能源或新物理定律。

[原文] [Alex Kantrowitz]: okay here's here's the big question at the end um you built Alph Go uh trained um the computer to play Go on human knowledge and then once it mastered the human level playing you kind of like let it loose yes with a program called Alphazero and it started doing things that you could never even imagine and making new circuits in ways that surprised you

[译文] [Alex Kantrowitz]: 好的，最后这有一个大问题。你构建了 AlphaGo，利用人类知识训练电脑下围棋；而一旦它掌握了人类水平的棋艺，你就像是把它“放开”（let it loose）了——是的，通过一个叫 AlphaZero 的程序——然后它开始做一些你甚至无法想象的事情，并以令你惊讶的方式开辟了新的回路（指策略或思维路径）。

[原文] [Alex Kantrowitz]: um eventually maybe there will come a time where LLMs or some version of them um reach a a mastery of of human knowledge in the same way uh what is going to happen when you then let that loose and it does the same potentially does the same thing as Alpha Zero

[译文] [Alex Kantrowitz]: 最终，也许有一天，大语言模型（LLMs）或者它们的某种版本，会以同样的方式达到对人类知识的精通。当你那时把它“放开”，而它做了同样的事情——潜在地像 AlphaZero 那样做——会发生什么？

[原文] [Demis Hassabis]: i think it' be very exciting i mean that's that's what to me is would be the AGI moment is you know then it will discover a new superconductor room temperature superconductor that's possible in the laws of physics but we just haven't found that needle in the haststack or a new source of energy a new way to build optimal batteries

[译文] [Demis Hassabis]: 我认为那会非常令人兴奋。我是说，对我而言，那才将是真正的“AGI时刻”。你知道，那时它将发现一种新的超导体——室温超导体，这是物理定律允许的，只是我们还没在“大海捞针”中找到这根针；或者一种新的能源，一种构建最优电池的新方法。

[原文] [Demis Hassabis]: i think all of those things will become possible and indeed not just possible I think they will happen uh once we get to a system that's first of all uh got to you know human level knowledge and then there'll be some techniques maybe it will have to help invent some of those techniques but kind of like Alpha Zero that will allow it to go beyond into new uncharted territory

[译文] [Demis Hassabis]: 我认为所有这些事情都将变得可能。实际上不仅仅是可能，我认为它们将会发生。一旦我们拥有了一个系统，首先达到人类水平的知识，然后会有一些技术——也许它必须帮助发明其中一些技术，但就像 AlphaZero 一样——将允许它超越现有知识，进入全新的未知领域（uncharted territory）。

[原文] [Alex Kantrowitz]: that that idea of it like plugging weather system into its brain like it's going to be on that that exactly

[译文] [Alex Kantrowitz]: 就像那种把天气系统直接接入它大脑的想法一样，它将会达到那种境界……

[Demis Hassabis]: 正是（Exactly）。

[原文] [Alex Kantrowitz]: all right so exciting times Dennis thanks for coming on the show

[译文] [Alex Kantrowitz]: 好的，真是激动人心的时刻。Demis，谢谢你来参加节目。

[原文] [Demis Hassabis]: thank you thanks everybody

[译文] [Demis Hassabis]: 谢谢。谢谢大家。

[原文] [Alex Kantrowitz]: Thank you so much

[译文] [Alex Kantrowitz]: 非常感谢。