Google this week announced a series of impressive AI updates at its IO developer conference, including improved video generation, expanded AI Mode in search, and an advanced reasoning architecture called Deep Think.

谷歌本周在其 IO 开发者大会上宣布了一系列令人印象深刻的 AI 更新，包括改进的视频生成、在搜索中扩展的 AI 模式，以及名为 Deep Think 的高级推理架构。

But the news came as the AI industry reckons with questions about how much better these models can get, and whether there is a path toward artificial general intelligence (a term some would rather let die due to overuse).

但这则消息发布之际，AI 行业正面临着关于这些模型还能变得多么强大，以及是否存在通往通用人工智能（某些人因该术语被滥用而希望停用）道路的质疑。

To tackle these questions about the frontier of AI, I sat down with Google DeepMind CEO Demis Hassabis for an on-stage interview at Google IO on Tuesday. And just after I got mic’d up, Google co-founder Sergey Brin walked in and joined as a surprise guest. The conversation went deep into the technology and expanded into some interesting directions, including the future of the web, robotics, and simulation theory.

为了解答这些关于 AI 前沿的问题，我在周二的 Google IO 大会上与 Google DeepMind 首席执行官 Demis Hassabis 进行了一场舞台对谈。就在我戴上麦克风后不久，谷歌联合创始人 Sergey Brin 突然走进来，惊喜加入。谈话深入探讨了技术，并扩展到一些有趣的方向，包括网络的未来、机器人技术以及模拟理论。

You can listen to my conversation with Hassabis and Brin on Apple Podcasts, Spotify, or your podcast app of choice. If you prefer reading, here’s our full discussion, edited lightly for length and clarity.

你可以在 Apple Podcasts、Spotify 或你喜欢的播客应用上收听我与 Hassabis 和 Brin 的对话。如果你更喜欢阅读，这里是经过轻度删减和润色的完整讨论稿。

Alex Kantrowitz:

Demis, given what we know today about AI frontier models, how much improvement is there left to be unlocked?

Demis，鉴于我们今天对 AI 前沿模型的了解，还有多少提升空间尚待发掘？

Demis Hassabis:

We're seeing incredible gains with the existing techniques, pushing them to the limit. But we're also inventing new things all the time as well. And I think to get all the way to something like AGI may require one or two more new breakthroughs. We have lots of promising ideas that we're cooking up, that we hope to bring into the main branch of the Gemini branch.

我们正凭借现有技术取得惊人的进步，并将其推至极限。但我们也在不断发明新方法。我认为，要完全实现类似 AGI 的能力，可能还需要一两项新的突破。我们正在酝酿许多有前景的想法，希望将其纳入 Gemini 主分支。

Alex Kantrowitz:

There's been this debate about whether scaling up data centers solves all problems when it comes to building better AI models. When you’re working on today’s models, is scale still the star, or is it a supporting actor?

关于扩建数据中心是否能解决构建更好 AI 模型的所有问题一直存在争论。当你在开发今天的模型时，规模仍然是主角，还是只是配角？

Hassabis:

I've always been of the opinion you need both. You need to scale to the maximum the techniques that you know about. You want to exploit them to the limit, whether that's data or compute or scale. And at the same time, you want to spend a bunch of effort on what's coming next, maybe six months or a year down the line so you have the next innovation that might do a 10x leap in some way to intersect with the scale. So you want both. Sergey, what do you think?

我一直认为两者都需要。你需要将已知技术扩展到极限——无论是数据、算力还是规模。同时，你还需要投入大量精力研究下一步，也许是六个月或一年后的新创新，以实现某种 10 倍的飞跃并与规模相交汇。所以两者都要。Sergey，你怎么看？

Sergey Brin:

I agree it takes both. You can have algorithmic improvements and simple compute improvements, better chips, more chips, more power, and bigger data centers. Historically, if you look at things like the N-body problem and simulating gravitational bodies and things like that—as you plot it, the algorithmic advances have actually beaten out the computational advances, even with Moore's law. If I had to guess, I would say the algorithmic advances are probably going to be even more significant than the computational advances, but both of them are coming up now, so we're getting the benefits of both.

我同意需要两者兼备。你可以拥有算法改进，也可以拥有简单的计算改进——更好的芯片、更多的芯片、更高的功率以及更大的数据中心。从历史上看，如果你看看 N 体问题、模拟引力系统等领域——当你绘制曲线时，即便有摩尔定律，算法进步实际上也超过了计算硬件的进步。如果让我猜，我会说算法进步可能比计算硬件进步更为重要，但两者现在都在提升，所以我们能同时受益。

Alex Kantrowitz:

Okay, but is the majority of your improvement coming from scale? There's talk about how the world will be just wallpapered with data centers. Is that your vision?

好的，但你们的大部分进步是否来自规模？有人说未来世界将被数据中心铺天盖地地覆盖。这是你的愿景吗？

Hassabis:

We're definitely going to need a lot more data centers. It still amazes me—from a scientific point of view—we turn sand into thinking machines. It's pretty incredible. But it's not just for training. Now, we've got these models that everyone wants to use and we're seeing incredible demand for Gemini 2.5 Pro, and Gemini Flash, we're really excited about how performant that is for the incredible low cost.

我们绝对需要更多的数据中心。这仍然让我惊叹——从科学角度来看，我们把沙子变成了会思考的机器，这实在令人难以置信。但这不仅仅是为了训练。现在，我们拥有人人都想用的模型，我们看到了对 Gemini 2.5 Pro 和 Gemini Flash 的惊人需求，我们对它们在极低成本下的卓越性能感到非常兴奋。

The whole world is going to want to use these things. We're going to need a lot of data centers for serving and also for inference-time compute. You saw, today, 2.5 Pro deep-thinking. The more time you give it, the better it will be. And certain tasks, very high value, very difficult tasks, it will be worth letting it think for a very long time. We're thinking about how to push that even further. Again, that's going to require a lot of chips and runtime.

全世界都会想使用这些技术。为了提供服务并满足推理阶段的计算需求，我们将需要大量数据中心。今天你已经见到 2.5 Pro 的深度思考：给它的时间越长，效果越好。对于某些高价值、极其困难的任务，让它长时间思考是值得的。我们正在思考如何把这一点推得更远。当然，这将需要大量芯片和运行时间。

Alex Kantrowitz:

We've been about a year into this reasoning paradigm. Demis, can you help us contextualize the magnitude of improvement we're seeing from reasoning?

我们在这种推理范式上已经探索了一年左右。Demis，你能帮我们说明一下推理带来的改进幅度到底有多大吗？

Hassabis:

We've always been big believers in what we're now calling the thinking paradigm. If you go back to our very early work on things like AlphaGo, AlphaZero, and our agent work on playing games, they will all have this attribute of a thinking system on top of a model. And you can quantify how much difference that makes.

我们一直坚信如今所说的“思考范式”。如果回到我们早期的研究，例如 AlphaGo、AlphaZero 以及各种博弈代理，它们都具有在模型之上再加一层思考系统的特征，而这种差异是可以量化的。

If you look at a game like Chess or Go, we had versions of AlphaGo and AlphaZero with the thinking turned off – so it was just the model telling you its first idea. And it's not bad. It's maybe like master level, something like that. But then, if you turn the thinking on, it's way beyond World Champion level. It's like a 600+ ELO difference between the two versions. You can see that in games, let alone for the real world, which is way more complicated.

以国际象棋或围棋为例，我们曾经运行过关闭思考功能的 AlphaGo 和 AlphaZero 版本——也就是模型只给出它的第一个想法。那也不错，大概是大师水平。可一旦开启思考功能，它就远超世界冠军水平，两种版本差了 600 多分 ELO。你在游戏中能看到这种差距，更不用提复杂得多的现实世界了。

Of course, the challenge is that your models need to be a world model, and that's much harder than building a model of a simple game. It has errors in it, and those can compound over longer term plans. But I think we're making really good progress on all that, all those fronts.

当然，挑战在于你的模型必须成为“世界模型”，这比构建一个简单游戏的模型困难得多。它内部会有误差，这些误差会在长期计划中累积。但我认为我们在所有这些方面都取得了非常好的进展。

Brin:

As Demis said, DeepMind really pioneered a lot of this reinforcement learning work. And what they did with AlphaGo and AlphaZero, as you mentioned, showed something you would take 5,000 times as much training to match what you were able to do with still a lot of training and the inference time compute, that you were doing with Go. So it's obviously a huge advantage. And obviously, like most of us, we get some benefit by thinking before we speak. I was reminded to do that. The AIs, obviously, are much stronger once you add that capability. We're just at the tip of the iceberg right now. In that sense, it's been less than a year that these models have really been around.

正如 Demis 所说，DeepMind 在强化学习方面确实开创了许多工作。你提到的 AlphaGo 和 AlphaZero 证明，如果没有推理时的计算，要达到他们在围棋中通过大量训练和推理计算取得的能力，需要 5,000 倍的训练量。这显然是一大优势。而且，像我们大多数人一样，开口前先思考会带来好处——我也被提醒要做到这一点。为 AI 加入这种能力后，它们显然会强大得多。现在我们只是触及了冰山一角。从这个意义上说，这些模型真正出现还不到一年。

Hassabis:

An AI during its thinking process can also use a bunch of tools, or even other AIs, to improve what the final output is. It's going to be an incredibly powerful paradigm.

AI 在思考过程中还可以使用各种工具，甚至调用其他 AI，以改进最终输出。这将是一种极其强大的范式。

DeepThink is a new feature that basically runs a bunch of parallel reasoning processes and then checks them against each other and optimizes. It's like reasoning on steroids. Demis, is this one of those few advances you mentioned that might get the industry closer to AGI?

DeepThink 是一项新功能，它基本上会并行运行一系列推理过程，然后相互校验并加以优化。这就像给推理打了“类固醇”。Demis，这是否就是你提到的那少数几个可能让行业更接近 AGI 的突破之一？

Hassabis:

It's maybe part of one, shall we say? There are others too that we need to be part of improving reasoning. Where does true invention come from, where you're not just solving a mass conjecture, you're actually proposing one, or hypothesizing a new theory in physics? We don't have systems yet that can do that type of creativity. They're coming.

可以说，它也许算是其中的一部分吧？要提升推理能力，我们还需要其他突破。真正的发明从何而来？不是仅仅去证明一个庞大的猜想，而是提出新的猜想，或者在物理学中假设新理论？我们目前还没有能够做到这种创造力的系统，但它们正在到来。

We need a lot of advances on the accuracy of the world models that we're building. You saw that with Veo. The potential Veo 3 has amazes me, like how it can intuit the physics of the light and the gravity. I used to work on computer games, not just the AI, but also graphics engines in my early career. I remember having to do all of this by hand and program all of the lighting, and the shaders, and all of these things. It’s incredibly complicated stuff we used to do in early games. And now it's just intuiting it within the model. It's pretty astounding.

我们在构建的世界模型准确性方面需要取得大量进展。你在 Veo 中已经看到了这一点。Veo 3 的潜力让我惊叹，它能直觉地理解光线和重力的物理规律。我早期做过电子游戏，不只是 AI，还写过图形引擎。那时必须手动编程所有光照、着色器等，非常复杂。而现在模型自己就能“领悟”这些，实在令人震撼。

Alex Kantrowitz:

Demis, I saw you shared an image of a frying pan with some onions, some oil. There was no subliminal messaging about that?

Demis，我看到你分享了一张煎锅里有洋葱和油的图片，这里面没有什么潜在暗示吧？

Hassabis:

No, not really. Just a maybe a subtle, subtle message.

没有啦，不是真的。也许只是一个非常、非常微妙的暗示而已。

Alex Kantrowitz:

There's a movement within the AI world right now to stop using the term AGI. Some see it as so overused as to be meaningless. But Demis, you think it's important. Why?

AI 领域现在出现了一股呼声，想要停止使用“AGI”这一术语。有些人认为它被过度使用，已经失去了意义。但 Demis，你认为它很重要。为什么？

Hassabis:

It's very important. There's two things that are getting a little bit conflated. One is, what can a typical person do? And we're all very capable, but, however capable we are, there's only a certain slice of things that one can be an expert in. You could say, what can you do that 90% of humans can’t do? That's obviously going to be economically, very important, and from a product perspective, also very important. So it's a very important milestone. We should say that's typical human intelligence.

这非常重要。有两件事往往被混为一谈。第一是，一个普通人能做什么？每个人都很能干，但一个人只能在有限领域成为专家。你可以问：你能做哪些 90% 的人做不到的事？这在经济和产品层面都至关重要，因此是一个重要的里程碑。我们可以把这称为典型人类智能。

But what I'm interested in, and what I would call AGI, is a more theoretical construct, which is: what is the human brain, as an architecture, able to do? And the human brain is an important reference point, because it's the only evidence we have, maybe in the universe, that general intelligence is possible. You'd have to show your system was capable of doing the range of things even the best humans in history were able to do with the same brain architecture. It's not one brain but the same brain architecture: what Einstein did, what Mozart was able to do, what Marie Curie did and so on. It's clear to me today’s systems don't have that.

但我所关心、并称之为 AGI 的，是一个更为理论化的概念：以人类大脑这套架构而言，它究竟能做什么？人脑是一个重要参照，因为它也许是宇宙中唯一证明泛化智能可行的证据。要称得上 AGI，你得证明你的系统能用同样的大脑架构完成历史上最杰出的人类所能做的广泛事务。这不是特指某个大脑，而是同一套大脑架构：爱因斯坦、莫扎特、居里夫人的成就等等。显然，今天的系统还未达到这一点。

The hype today on AGI is sort of overblown because our systems are not consistent enough to be considered to be fully “general.” Yet they're quite general. They can do thousands of things. You've seen many impressive things today, but every one of us has experience with today's chatbots and assistants. You can easily, within a few minutes, find some obvious flaw with them: some high school math thing that it doesn't solve or some basic game it can't play. It's not very difficult to find those holes in the system, and for something to be called AGI, it would need to be consistent, much more consistent across the board than it is today. It should take a couple of months for a team of experts to find a hole in it, an obvious hole in it, whereas today, it takes an individual only minutes to find one.

如今关于 AGI 的炒作有些言过其实，因为我们的系统还不够一致，无法被视为真正“通用”。它们确实很通用，能做成千上万件事，你也见识了许多令人惊叹的效果。但我们每个人都试用过当今的聊天机器人与助手，只需几分钟就能发现明显缺陷：某道高中数学题解决不了，某个简单游戏不会玩。要称为 AGI，系统必须在各方面都更一致。理想情况下，专家团队得花几个月才能找到明显漏洞，而今天，一个人几分钟就能发现。

Alex Kantrowitz:

Sergey, this is a good one for you. Do you think that AGI is going to be reached by one company and it's game over? Or could you see Google having AGI OpenAI having AGI. Anthropic having AGI, China having AGI?

Sergey，这个问题很适合你。你认为 AGI 会被某家公司率先实现然后尘埃落定，还是会出现谷歌有 AGI、OpenAI 有 AGI、Anthropic 有 AGI、中国也有 AGI 的局面？

Brin:

One company or country or entity will reach AGI first. Now, it is a little bit of a spectrum. It's not a completely precise thing. It's conceivable that there will be more than one roughly in that range at the same time after that. What happens? It's very hard to foresee, but you could certainly imagine there's going to be multiple entities that come through. In our AI space, when we make a certain advance, other companies are quick to follow, and vice versa. When other companies make certain advances, it's a constant leapfrog. There's an inspiration element that you see, and that would probably encourage more and more entities to cross that threshold.

某家公司、国家或实体会最先达到 AGI。不过这是一条光谱，并非绝对精确。之后完全可能出现多家大致同水平的情况。那会发生什么？很难预见，但完全可以想象会有多个实体接踵而至。在 AI 领域，我们取得某项突破后，其他公司会迅速跟进，反之亦然。大家不断你追我赶，互相激励，这可能促使更多实体跨过那道门槛。

Demis, what do you think?

Demis，你怎么看？

Hassabis:

It is important for the field to agree on a definition of AGI. We should try to coalesce, assuming there is one. There probably will be some organizations that get there first. It's important that those first systems are built reliably and safely and after that, if that's the case we can imagine using them to shard off many systems that have safe architectures built underneath them. You could have personal AGIs and all sorts of things happening. As Sergey says, it's pretty difficult to predict, to see beyond the event horizon, to predict what that's going to be like.

让整个领域就 AGI 的定义达成一致至关重要。如果存在这样一个定义，我们应当努力汇聚于此。很可能会有一些组织最先实现 AGI。关键在于这些首批系统必须可靠、安全地构建；如果做到这一点，我们可以设想在它们之下分化出许多具备安全架构的系统——比如个人 AGI，以及形形色色的应用。正如 Sergey 所言，要越过“事件视界”去预测那将是什么光景，非常困难。

Alex Kantrowitz:

The conventional wisdom is AGI must be knowledge. The intelligence of the brain. What about the intelligence of the heart? Demis, does AI have to have emotion to be considered AGI?

传统观点认为 AGI 必须体现知识，也就是大脑的智能。那么“心灵的智能”又如何？Demis，AI 是否必须具备情感才能被视为 AGI？

Hassabis:

It’ll need to understand emotion. It will be almost a design decision if we wanted to mimic emotions. I didn't see any reason why it couldn't, in theory, but it might be different, or it might be not necessary, or in fact, not desirable for them to have the sort of emotional reactions that we do as humans. Again, it's a bit of an open question, as we get closer to this AGI timeframe, which is on a 5 to 10 year timescale. We have a bit of time, not much time, but some time, to research those kinds of questions.

它需要理解情感。如果我们想让 AI 模拟情感，这几乎是一个设计决策。从理论上说，没有什么理由阻止它具备情感，但它可能呈现出不同的形式，也可能并非必要，甚至并不希望它像人类那样产生情绪反应。随着我们逼近 5 到 10 年的 AGI 时间表，这仍是一个开放问题。我们还有一些时间——不算多，但也还有——去研究这些问题。

Alex Kantrowitz:

When I think about how the time frame might be shrunk, I wonder if it's going to be the creation of self-improving systems. Last week, Google DeepMind announced AlphaEvolve, which is an AI that helps design better algorithms. Demis, are you trying to cause an intelligence explosion?

当我思考如何缩短时间表时，我在想这是否会由自我改进系统的诞生来推动。上周，Google DeepMind 发布了 AlphaEvolve——一种帮助设计更优算法的 AI。Demis，你是在试图引发一次智能大爆炸吗？

Hassabis:

No, not an uncontrolled one. It's an interesting first experiment. It's an amazing system, a great team is working on that. Where it's interesting now, to start pairing other types of techniques, in this case evolutionary programming techniques, with the latest foundation models, which are getting increasingly powerful, and actually to see in our exploratory work a lot more of these combinatorial systems and pairing different approaches together.

不会——至少不会是不受控的爆炸。这只是一次有趣的初步实验。这套系统令人惊叹，一支出色的团队正致力于此。现在的看点是把其他技术流派——此处是进化编程——与日益强大的基础模型结合，在我们的探索中看到越来越多组合式系统与多种方法的配对。

And you're right, someone discovering a self improvement loop would be one way where things might accelerate further than they're even going today. And we've seen it before with our own work with things like AlphaZero, learning Chess and Go and any two player game from scratch within less than 24 hours, starting from random with self improving processes. So we know it's possible, but again, those are in quite limited game domains which are very well described. The real world is far messier and far more complex. It remains to be seen if that type of approach can work in a more general way.

你说得对，若有人真正发现了自我改进的闭环，事态可能会比现在还要加速。我们在 AlphaZero 等项目中已见过类似情形——它从零开始，仅用不到 24 小时就通过自我改进学会国际象棋、围棋等双人博弈。因此我们知道这是可行的，但那仍是规则清晰、范围有限的游戏领域。现实世界杂乱得多、复杂得多，这种方法能否在更通用的场景奏效，还有待观察。

Alex Kantrowitz:

We've talked about some very powerful systems. And it's a race. It's a race to develop these systems. Sergey, is that why you came back to Google?

我们刚讨论了这些极为强大的系统。这是一场竞赛，一场开发这些系统的竞赛。Sergey，这就是你重返谷歌的原因吗？

Brin:

As a computer scientist, it's a very unique time in history. Honestly, anybody who's a computer scientist should not be retired right now. They should be working on AI. That's what I would just say. There's just never been a greater problem and opportunity or a greater cusp of technology. I wouldn't say it's because of the race, although we fully intend that Gemini will be the very first AGI. But to be immersed in this incredible technological revolution—it's unlike the web 1.0 thing, it was very exciting, whatever, we had mobile, we had this, we had that—but this is scientifically far more exciting. Ultimately, the impact on the world is going to be even greater. The web and mobile phones have had a lot of impact, AI is going to be vastly more transformative.

作为一名计算机科学家，现在是历史上空前独特的时刻。说实话，任何计算机科学家都不该此时退休，他们应该投身 AI。我只能这么说。再没有比这更大的难题与机遇、更尖端的技术临界点。我不会说是因为竞赛，尽管我们完全希望 Gemini 成为首个 AGI。但能沉浸在这场非凡的技术革命中——这与当年的 Web 1.0 不同，那固然激动人心，我们有移动互联网，有种种创新——而现在这一波在科学层面更加令人兴奋。最终，AI 对世界的影响会更为深远。网络和手机已带来巨大改变，AI 将带来成倍更大的变革。

Alex Kantrowitz:

So what do you do day to day?

那你日常都在做些什么？

Brin:

I think I torture people like Demis. He is amazing — by the way he tolerated me crashing this fireside chat. I'm across the street pretty much every day, and there’s just people who are working on the key Gemini text models, on the pre-training, on the post-training, mostly those. I periodically delve into some of the multi-modal work, Veo 3 as you've all seen, but I tend to be pretty deep in the technical details. That's a luxury I really enjoy, fortunately, because guys like Demis are minding the shop. That's just where my scientific interest is. It's deep in the algorithms and how they can evolve.

我想我经常折磨像 Demis 这样的人。他非常了不起——顺便说一句，他还容忍我闯进这场炉边谈话。我几乎每天都在街对面的办公室，那儿的人正专注于关键的 Gemini 文本模型，做预训练、后训练，主要就是这些。我也时不时涉猎一些多模态工作，比如大家见过的 Veo 3，但我往往深耕技术细节。幸运的是，像 Demis 这样的伙伴在掌舵，我得以奢侈地沉浸其中。我的科研兴趣正是在于深究算法及其演化方式。

Alex Kantrowitz:

Let's talk about the products a little bit. Demis, Google’s vision of AI agents is often through the camera. It's very visual. There was an announcement about smart glasses today. So talk a little bit about why Google is so interested in having an assistant or a companion that is something that sees the world as you see it.

说到产品，Demis，谷歌对 AI 代理的愿景往往是经由摄像头，非常偏重视觉。今天还发布了智能眼镜。能不能谈谈为什么谷歌如此热衷于打造一个像你一样“看”世界的助手或伙伴？

Hassabis:

Well, it's for several reasons, several threads come together. As we talked earlier, we've always been interested in agents. That's actually the heritage of DeepMind. We started with agent-based systems in games. We are trying to build AGI, which is a full general intelligence, clearly, that would have to understand the physical environment, physical world around you. And two of the massive use cases for that are a truly useful assistant that can come around with you in your daily life, not just stuck on your computer or one device. We want it to be useful in your everyday life, for everything. It needs to come around you and understand your physical context.

原因有好几个，且相互交织。如之前所说，我们一直对“智能体”情有独钟——这正是 DeepMind 的传承。我们最初在游戏中就做基于智能体的系统。而构建 AGI 意味着实现真正的通用智能，它必须理解你周围的物理环境与现实世界。对此有两大应用场景：其一是真正实用的助手，能伴你左右，而不仅是困在电脑或某台设备里；我们希望它在你日常生活的方方面面都派得上用场，因而它必须随行并理解你的物理语境。

The other big thing is, I've always felt for robotics to work, you want what you saw with Project Astra on a robot. The bottleneck in robotics isn't so much the hardware — although there are many, many companies working on fantastic hardware, and we partner with a lot of them — but it's actually the software intelligence that is always what's held robotics back.

另一件大事是，我一直认为若想让机器人真正奏效，你得把 Project Astra 那样的能力装进机器人里。机器人的瓶颈与其说在硬件——虽然确实有很多公司在打造出色的硬件，我们也与不少伙伴合作——不如说在于软件智能，这才是长期制约机器人发展的核心。

We're in a really exciting moment now, where finally, with these latest versions — especially Gemini 2.5 and more things that we're going to bring in the Veo technology and other things — we're going to have really exciting algorithms to make robotics finally work and realize its potential, which could be enormous. In the end, AGI needs to be able to do all of those things. That's why, you can see, we always had this in mind. That's why Gemini was built from the beginning, even the earliest versions, to be multimodal. That made it harder at the start, because it's harder to make things multimodal with just text only. But in the end, we're reaping the benefits of those decisions now, and I see many of the Gemini team here in the front row. Of the correct decisions we made, they were the harder decisions, but we made the right decisions, and now you can see the fruits of that.

如今我们正处在一个令人振奋的时刻：借助最新版本——尤其是 Gemini 2.5，以及即将把 Veo 技术等融入的更多能力——我们终于拥有足够强大的算法，让机器人真正发挥作用并释放巨大潜力。归根结底，AGI 需要胜任这些任务。这也是为什么从一开始，甚至在最早的版本里，Gemini 就被设计为多模态——虽然这让早期开发更困难，仅靠文本很难做到多模态。但如今我们正在收获当初那些艰难却正确决定的成果。我看到前排就坐着不少 Gemini 团队成员：这些正确的抉择虽难，却让我们迎来了丰硕成果。

Alex Kantrowitz:

I've been thinking about whether to ask you a Google Glass question, Sergey

我一直在想要不要问你一个 Google Glass 的问题，Sergey。

Brin:

Fire away

尽管问。

Alex Kantrowitz:

What did you learn from glass that Google might be able to apply today, now that it seems like smart glasses have made a reappearance,

现在智能眼镜似乎再度回归，你从 Google Glass 得到哪些经验，今天谷歌可以用得上？

Brin:

I learned a lot. I made a lot of mistakes with Google Glass. I'll be honest, I am still a big believer in the form factor, so I'm glad that we have it now: it looks like normal glasses, it doesn't have the thing in front. There was a technology gap. Now in the AI world, the things that these glasses can do to help you out without constantly distracting you, that capability is much higher.

我学到了很多，也在 Google Glass 上犯了不少错误。实话说，我仍然坚信这种形态，所以很高兴现在的产品看起来就是普通眼镜，没有挡在面前的那块屏。过去确实存在技术差距。而如今在 AI 时代，这些眼镜能帮你做事而不持续分散注意力，能力已经大幅提升。

Also, I didn't know anything about consumer electronics supply chains, and how hard it would be to build that and have it be at a reasonable price point, managing all the manufacturing and so forth. This time, we have great partners that all are helping us build this. That's another step forward. What else can I say? I miss the airship with the wing suiting skydivers for the demo. It would have been even cooler here at Shoreline Amphitheater than it was up in Moscone back in the day. But maybe we'll have to. We should probably polish the product first this time, be ready and available, and then we'll do a really cool demo. That's probably a smart move.

此外，我当年对消费电子供应链一无所知，也不知道要把产品做到合适的价格并管理好制造流程有多难。这一次，我们拥有出色的合作伙伴共同打造，这是又一大进步。我还能说什么呢？我倒是怀念当年用翼装跳伞和飞艇做的那个演示——如果在 Shoreline 圆形剧场再来一次，肯定比当年在 Moscone 更炫。但或许我们得先把产品打磨好、准备充分，再来一次酷炫的演示——这大概才是明智之举。

Hassabis:

We've got, obviously, an incredible history of Google Glass devices and smart devices, we can bring all those learnings to today, and I am very excited about our new glasses, as you saw. What I am always talking to our team about is, and I don't know if Sergei would agree, but I feel like the universal assistant is the killer application for smart glasses, and that's what's going to make it work, apart from the fact that it's all the tech, the hardware and technology has also moved on and improved a lot. This is the natural killer application for it.

显然，我们在 Google Glass 及各类智能设备方面拥有卓越的积累，我们可以把所有经验带到今天，如你所见，我对我们的新款眼镜感到非常兴奋。我经常和团队讨论的一点——虽然不知道 Sergey 是否同意——就是我认为“通用助手”才是智能眼镜的终极杀手级应用，这正是它成功的关键。当然，硬件和技术也在大幅进步，这无疑是它的天然杀手级应用。

Alex Kantrowitz:

Okay, briefly on video generation, Demis, If the internet fills with video that's been made with artificial intelligence, and then you train on that video, does that lead to a lower model quality if than if you were training just from human generated content?

好，简单谈谈视频生成的问题，Demis：如果互联网充斥着 AI 生成的视频，而你又用这些视频进行训练，相较于纯粹使用人类创作内容进行训练，模型质量是否会下降？

Hassabis:

There's a lot of worries about this so-called “model collapse” — video is just one thing, but in any modality, text as well. First of all, we're very rigorous with our data quality management and curation. Also, at least for all of our generative models, we attach synth ID to them, so there's this invisible AI-made watermark that is very robust as held up now for a year or 18 months since we released it, and all of our images and videos are embedded with these watermarks so we can detect and we're releasing tools to allow anyone to detect these watermarks and know that that was an AI generated image or video.

关于所谓“模型崩塌”存在诸多担忧——视频只是其中一种形态，文本等其他模态也一样。首先，我们对数据质量管理与筛选非常严格。此外，至少在我们所有生成模型中，我们都会附加 SynthID，也就是一种不可见的 AI 水印，发布至今一年多依然十分可靠。我们所有图像和视频都嵌入了这种水印，因此我们能检测它，并正推出工具让任何人都能检测该水印，从而分辨哪些是 AI 生成的图像或视频。

And of course, that's important to combat deep fakes and misinformation, but could use that to filter out, if you wanted to, whatever was in your training data. I don't see that as a big problem. Eventually, we may have video models that are so good you could put them back into the loop as a source of additional data, synthetic data, it's called, and there, you’ve got to be very careful that you're creating from the same distribution that you're going to model, you're not distorting that distribution somehow, and the quality is high enough. We have some experience of this in a completely different domain with things like AlphaFold, where there wasn't enough real experimental data to build the final AlphaFold, so we had to build an earlier version that then predicted about a million protein structures, and then we selected it had a confidence level on that. We selected the top 300,000-400,000 and put them back in the training data. It's very cutting edge research to mix synthetic data with real data. There are also ways of doing that. On the terms of the video generator stuff, you can just exclude it if you want to, at least with our own work, and hopefully other generative media companies follow suit and put robust watermarks in, first and foremost to combat deep fakes and misinformation.

当然，这对于打击深度伪造和虚假信息至关重要，但如果你愿意，也可利用水印来过滤训练数据中的这类内容。我并不认为这是大问题。最终，我们或许会拥有质量极高的视频模型，可以将其生成的视频作为“合成数据”重新投入训练循环。在那种情况下，必须确保这些数据与模型要学习的分布一致、不致失真且质量足够高。我们在完全不同的领域已有相关经验，例如 AlphaFold：由于缺乏足够的实验数据，我们先用早期版本预测约一百万个蛋白质结构，再从中筛选出 30-40 万个高置信度样本回填训练集。将合成数据与真实数据混合训练是前沿研究，也有多种方法可行。就视频生成而言，若你愿意完全排除合成内容至少在我们自己的工作里可以做到，也希望其他生成媒体公司跟进，首先用牢固的水印来打击深度伪造和错误信息。

Alex Kantrowitz:

We now move to the miscellaneous section of my questions. Let's see how many we can get through and as fast as we can get through them. Let's go to Sergey with this one. What does the web look like in 10 years?

现在进入我的“杂项”提问环节。让我们尽量多问几条、快速过完。先请教 Sergey：十年后的网络会是什么样子？

Brin:

The rate of progress in AI is so far beyond anything, not just the web, I don't think we really know what the world will look like in 10 years.

AI 的进步速度远超以往任何事物，不仅是对网络如此。我认为我们很难真正知道十年后的世界会是什么样。

Alex Kantrowitz:

Demis?

Demis？

Hassabis:

That's a good answer. In nearer term, the web is going to change quite a lot if you think about an agent-first web. It doesn't necessarily need to see renders like we do as humans using the web. Things will be pretty different in a few years.

这是个不错的答案。从更近的角度看，若以“智能体优先”视角思考，网络将发生巨大变化。未来的智能体未必需要像人类那样看到网页渲染结果。再过几年，一切都将大不相同。

AGI, before 2030 or after? 2030

AGI，在 2030 年之前还是之后？2030

Brin:

I’m going to say before.

Alex Kantrowitz:

Demis?

Demis？

Hassabis:

I’m just after.

Brin:

No pressure Demis.

Hassabis:

Exactly, but I have to go back and get working harder.

没错，但我得回去更加努力工作。

Brin:

I can ask for it, he needs to deliver it. Stop sandbagging.

我可以提出要求，他必须兑现。别再故意保守了。

Alex Kantrowitz:

Would you hire someone that used AI in their interview?

如果有人在面试中使用了 AI，你会录用他吗？

Hassabis:

It depends how they used it. Using today's models and tools, probably not. It depends how they would use it, actually.

要看他们怎么用。就今天的模型和工具而言，大概不会。关键在于他们具体的使用方式。

Alex Kantrowitz:

Sergey?

Sergey？

Brin:

I never interviewed at all, so I don't know. It would be hypocritical for me to judge people exactly how they interview.

我从来没参加过面试，所以我不清楚。由我来评判别人如何面试有点虚伪。

Hassabis:

Yeah, I haven't either. I've never done a job interview.

是啊，我也没有。我从没做过求职面试。

Alex Kantrowitz:

Demis, I've been reading your tweets. You put a very interesting tweet up where there was a prompt that created some sort of natural scene. Here was the tweet. “Nature to simulation at the press of a button does make you wonder.” And people ran with that and wrote some headlines saying, Demis thinks we're in a simulation. Are we in a simulation?

Demis，我读了你的推文。你发了一条很有趣的推文，用一个提示生成了某种自然场景。内容是：“按下按钮，自然即成模拟，令人深思。” 于是有人写标题说 Demis 认为我们活在模拟中。我们是在模拟里吗？

Hassabis:

Not in the way that Nick Bostrom and people talk about. I don't think this is a game, even though I wrote a lot of games. Ultimately, underlying physics is information theory. We're in a computational universe, but it's not a straightforward simulation. I can't answer you in one minute, but the fact that these systems are able to model real structures in nature is quite interesting and telling. I've been thinking a lot about the work we've done with AlphaGo and AlphaFold in these types of systems. I’ve spoken a little about it. Maybe at some point I'll write up a scientific paper about what I think that really means in terms of what's going on here in reality.

并不是 Nick Bostrom 等人所说的那种。我不认为这是一场游戏，尽管我写过很多游戏。从根本上说，底层物理是信息论。我们处在一个计算宇宙中，但这并非简单的模拟。我无法在一分钟内回答，但这些系统能够建模自然中的真实结构，本身就颇耐人寻味。我一直在思考我们在 AlphaGo、AlphaFold 这类系统中的工作，也谈过一些。或许改天我会写篇科学论文，阐述这对理解现实意味着什么。

Alex Kantrowitz:

Sergey, want to make a headline?

Sergey，想不想来个大标题？

Brin:

Well, that argument applies recursively. If we're in a simulation, then by the same argument, whatever beings are making the simulation are themselves in a simulation for roughly the same reasons, and so on and so forth. You're going to have to either accept that we're in an infinite stack of simulations, or that there's got to be some stopping criteria.

Alex Kantrowitz:

And what's your best guess?

那你最好的猜测是什么？

Brin:

We're taking a very anthropocentric view. When we say simulation in the sense that some conscious being is running a simulation that we are then and they have some semblance of desire and consciousness that's similar to us, that's where it breaks down for me. I don't think that we're really equipped to reason about one level up in the hierarchy.

我们的视角过于以人为中心。当我们说“模拟”指某个有意识的存在在运行一个模拟，而我们就在其中，并且他们具有类似我们愿望与意识的时候，这个设想对我来说就崩溃了。我觉得我们并不具备去推理层级再往上一层的能力。

Demis, Sergey, thank you so much.

Demis，Sergey，非常感谢。