德米斯·哈萨比斯谈双子座3、谷歌反重力、医疗级人工智能及其他

德米斯。

Demis.

嘿，罗恩。

Hey, Ron.

很高兴见到你。

Good to see you.

我也很高兴见到你。

Good to see you too.

谢谢你

Thanks for

抽出时间。

your time.

今天，我们要讨论的是Gemini 3，谷歌最智能的旗舰模型。

Today, we're talking about Gemini three, the most intelligent flagship model from Google.

如果要用一句话解释，这次发布为什么重要？

If you had to explain it in one sentence, why is this launch important?

我认为它之所以重要，是因为它延续了过去几年我们在Gemini项目上的进展。

It's important, I think, because it just continues the progression I think we've been on with Gemini over the last couple of years.

我们对这个模型的整体表现非常满意。

And we're really happy with the overall performance of this model.

我想人们会对它感到非常惊喜。

I think people are going to be very pleasantly surprised by it.

我认为它将继续全面提升整体性能。

I think it just continues the overall performance increase across the board.

你可以从所有基准测试中看到这一点，从推理到工具调用，再到可靠性和创造力。

And you can see that from all the benchmarks, from reasoning to tool calling, reliability and creativity.

我认为在所有这些指标上它都更出色。

I think it's better across all those measures.

如果我们回顾从Gemini 2.5发布到现在Gemini 3.0的历程，Gemini在基准测试中达到这一水平，期间发生了哪些突破性进展？

If we rewind back to when Gemini 2.5 was launched to now with Gemini three, what breakthroughs have happened since where Gemini has gone to this level with the benchmarks?

是的。

Yeah.

实际上我们非常专注于...我的意思是，2.5版本也是个很棒的模型，我们确实对它非常满意。

Well, we've really focused quite hard on I mean, 2.5 was a great model too, we're really we're really pleased with that.

你也看到了它在市场上表现多么出色，无论是开发者群体还是Gemini应用中都广受好评。

And you saw how well it did in the market with developers and also in the Gemini app.

但我们希望改进工具调用和工具使用等功能，以及这些操作的可靠性。

But we wanted to improve things like tool calling and tool use and just sort of the reliability of that.

当然，这对编程非常重要，而编程正是这些模型的主要应用场景之一。

Of course, that's important for coding, which is one of the big use cases of these models.

但对于常规推理和日常使用方式也同样重要。

But it's also important for just general reasoning and generally how you use it.

我们做的另一件事，我认为是大幅提升了风格和人格化表现。

The other thing we did, I think we've done is improve the style and the persona a lot.

我觉得它更简洁、更切中要点、更有帮助性。

I think it's more succinct, more to the point, more helpful.

至少我个人...你知道，内部测试也显示用户更喜欢使用这个模型。

And at least I, you know and and the internal testing shows people enjoy using this model even more.

所以在代码推理方面确实有所提升。

So definitely a step up in coding reasoning.

但对于那些日常普通的技术工作者来说，比如现在已经在使用Gemini的非开发人员，他们明天突然能够做到而之前做不到的具体可感知的事情有哪些？

But for like an everyday normal tech worker that's not a developer who already uses Gemini today, what are the noticeable concrete things they'll suddenly be able to do tomorrow that they couldn't before?

是的。

Yeah.

这取决于你的使用场景，但我认为几乎所有我们尝试过的功能，比如头脑风暴、氛围编程、创意写作或内容总结，你都会发现它在所有这些方面都显著更优秀、更可靠、更智能，而且我认为风格上也更好。

It depends what your use case is, but I think almost everything we've tried, like, if you're brainstorming ideas or you're vibe coding or you're just doing some creative writing or summarizing things, you should find it's meaningfully better at all of those things, more reliable, much smarter and I think stylistically better too.

我认为工具调用这类功能，你会隐约感觉到它在底层更好地利用了搜索功能，基本上更准确了，因为工具调用变得更可靠了。

I think the tool calling and things like this, you'll sort of feel it under the hood that it's using search better and it's basically more accurate on things because the tool calling is better and more reliable.

所以我认为整体而言，作为一个普通Gemini应用用户，你应该能感受到这是一个全面提升——比如你会觉得它整体上能力更强、使用起来更愉快。

So I think across the board, you should just feel this is a much more for using it as a general Gemini app user, for example, you should feel it's just across the board, much more capable and more pleasant to to to work with and and and use.

有一点我在所有公告中都没有注意到的是记忆功能。

One thing that I didn't notice from any of the announcement was memory.

嗯。

Mhmm.

我很想听听你对此的看法。

I'd just love to hear your take on this.

我认为谷歌在所有工具上都有真正的优势，毕竟你们拥有Gmail、YouTube、地图等所有产品的海量用户数据。

I think Google has a real advantage across, you know, all the tools given how much data you guys have across Gmail, YouTube, Maps, everything else for the user.

坦白说，对我来说ChatGPT最吸引人的部分就是他们新增的小型记忆组件。

And for me, like, candidly, the stickiest part of ChatGPT has been the small memory component they've added.

嗯。

Mhmm.

对于如何长期整合到Gemini中，你有什么想法？

How do you think about integrating that long term across Gemini?

是的，我们正在深入探索个性化和记忆与上下文领域。

Yeah, we're really sort of going deep into personalization and memory and context.

我认为这是3.0时代的一部分。

And I think you're this is part of the three point zero era.

要知道，Gemini 3时代就是要加倍投入这些方面。

You know, the Gemini three era is to is to kind of double down on those things.

所以随着我们进入Gemini 3时代，你会看到更多相关的讨论。

So you're going to start seeing a lot of discussing that a lot more as we move into the three the Gemini three era.

显然还会有更多模型问世。

Obviously there's more models to come.

我们需要完善整个模型家族，并且会持续开发更多内置功能和能力。

There's the family to fill out family of models to fill out and we'll be doing that and more features and capabilities that are already built into the model.

但我们会逐步在产品及开发者界面中开放更多功能。

But we'll start exposing more and more in our products and our developer surfaces.

包括与谷歌生态系统的其他部分（如Gmail和日历等）的去个性化连接。

So including kind of depersonalization and connection into the other Google, the rest of the Google ecosystem, Gmail and calendar and so on.

你已经能看到一些相关功能正在实现。

And you're seeing bits of that already happening.

但这只是我们计划中的冰山一角。

But it's just sort of scratching the surface of what we have planned.

Gemini 3确实是一个能够实现这些目标的强大模型。

And Gemini three is a really capable model that's able to do that.

再次强调，工具调用和工具使用对于可靠地连接到这些其他界面将非常重要。

And again, things like tool calling and tool usage are going to be really important for reliably connecting in to these other surfaces.

是啊。

Yeah.

从所有基准测试来看，它的能力确实非常强大。

It does seem very capable with all the benchmarks.

我只是觉得，真希望它能早点问世。

I'm just like, I just wish it came sooner.

就像，对啊。

Like Yeah.

我经常使用Chachibis，而且你知道，Gemini在各项基准测试中都遥遥领先，还拥有所有这些访问权限。

I'm using Chachibis so often, and, you know, Gemini, it's beating everything in benchmarks, and it has all this access.

我知道这类事情很难给出确切时间表，但你们有没有大概的预估，比如3.0版本的真实记忆功能什么时候会开始推出？

I know it's hard to give a timeline on these things, but do you have any, like, rough estimates on when that real memory is gonna start rolling out on three point o?

是的。

Yeah.

我们你知道，我们一直在内部进行测试，可以说是持续'吃自己的狗粮'，围绕这个方向有很多不同的想法。

We you know, we we we're we're in testing in in in sort of dog fooding all the time internally, lots of different ideas around this.

当这些功能足够完善，我们觉得足够可靠时，就会尽快发布。

And when those things are polished enough and we feel are reliable enough, we'll put them out as soon as we can.

我们知道用户需要这个功能。

We know users want it.

我们还在构建更高效的模型版本，比如闪存版本这类改进，这将使我们能够大规模提供服务。

We're also building more efficient versions of the model, flash versions, these kinds of things, which will allow us to serve it at scale.

我们对正在进行的原型设计工作感到非常兴奋，很快你们就能看到成果。

So we're very excited about the sort of prototyping we're doing, and you'll see the fruits of that very soon.

另外我想提到的是，这些新模型的多模态能力让我印象深刻。

The other thing I should mention is, I think that I'm super impressed by with these new model is the kind of multimodal capabilities.

如你所知，Gemini一直表现非常出色。

Gemini, as you know, has always been really strong.

我认为在多模态推理、多模态理解与生成方面是业界顶尖水平，比如nanobanana这类技术。

I think best in class So to on multimodal reasoning, multimodal understanding and generation, things like nanobanana.

我们将通过这个新模型全面提升所有这些能力。

And we're going to like up level all of that with this new model.

我认为普通大众和用户将会从中获得很多益处。

So I think there's gonna be a lot think, you know, general the general public, the general user will see a lot of benefits of that.

我们正开始将其整合到其他平台，比如YouTube、AI工作室等，让这些新多模态能力得以展现。

And we're just gonna start plugging that into other surfaces, you know, with YouTube, AI studio and so on where those those will come through, shine through, I think, those new kind of multimodal capabilities.

我也很期待全面测试这些模型，看看世界会用它们做些什么。

I'm I'm excited to be fully testing out and see what the world does with the models as well.

随着3.0新模型的发布，你们还推出了Antigravity——一个新的代理开发平台。

Alongside the the new model three point o, you're launching antigravity, which is a new agentic development platform.

听起来这个平台能让每个开发者都拥有一个能在编辑器、终端和浏览器中运作的AI同事。

And it sounds like the platform enables every developer to have almost like a AI coworker that can operate across the editor, terminal, and browser now.

在你看来，Antigravity与其他主流代理编程应用的主要区别是什么？

But in your mind, what's the differentiator between anti gravity and the other major agentic coding apps that are out there right now?

我认为它会随时间迭代改进，但我们真正想做的是从'代理优先'的角度重新构想集成开发环境。

You know, I think it's gonna iterate over time, but I feel like we're really trying to reimagine the IDE from an agent first perspective.

我认为我们大致掌握了发展方向，明确了Gemini项目的核心目标和技术路线。

I think we sort of have the roadmap of where that's going, where we want to take Gemini and under the hood.

当然，你也可以结合反重力技术使用其他模型。

Although, of course, you can use other models, too, with antigravity.

我觉得我们正试图重新构想这一领域。

And I think we're trying to sort of reimagine that.

我们合作的那些前帆板运动员，他们显然是这方面的专家。

The windsurf guys that we're working on, the ex windsurf people, they're obviously experts in this.

因此我们对这个领域感到非常兴奋。

So this is we're very excited about this area.

我们已在内部使用这项技术，这是第一步，对吧？

We're using it internally, which is the first step, right?

大家使用起来确实乐在其中。

And people are really enjoying using it.

而且你知道，生产力提升的效果相当显著。

And the productivity you know, gains are impressive there.

但我认为我们仍处于起步阶段。

But I think we're still at the beginning of that.

没错。

Right.

随着系统能力不断增强——我们显然对此抱有期待——彻底重构整个体验究竟意味着什么？

As the as the systems become more capable, which we're obviously expecting them to be, what does it really mean to kind of reimagine that whole experience?

显然，我讨论的已经超越了面向业余编程者的氛围编程范畴。

And obviously, I'm talking beyond vibe coding here, which is more for the amateur coder.

对。

Right.

我们就称之为他们开发环境中的专业程序员版本吧。

Let's call it what is the professional coder one from their dev setup.

我认为反重力是我们尝试解答这个问题并制定相关路线图的第一步。

And I think anti gravity is our first attempt at trying to sort of answer that and build a roadmap towards that.

当然，还有像AI工作室这样的工具，可能更适合休闲开发者、独立开发者或专业消费者群体。

And then, of course, you've got things like AI studio that's more maybe for the casual developer or single developer or prosumer, let's call it.

所以我认为我们会根据不同专业程度以及是否团队协作等因素，提供不同的产品界面。

So I think we're going to have different surfaces depending on the level of professionalism and whether you're working in a team, this kind of thing.

我觉得反重力功能，你知道的，人们会非常喜欢这个。

And I think antigravity is, you know, people are really going to enjoy that.

所以反重力功能更多是针对专业程序员而非氛围型程序员？

So antigravity is more for that professional coder rather than the vibe coder?

我认为这是我们当前的目标定位。

I think that's what we're currently aiming for though.

当然，你知道，我们希望所有类型的开发者都能使用它。

Of course, you know, any developer will will, you know, hopefully many, many types, know, all types of developers will use it.

说到内部工具使用，我有个非常好奇的问题。

And speaking of using tools internally, this is a really curious question I have.

我听说谷歌现在正用AI生成大量新代码。

I heard Google's using AI to generate a lot of new code now.

但你们是否有尚未公开的内部工具或模型，以便你们能优先享受这些产品带来的早期优势？

But are there, like, internal tools or models that you guys have access to that you're not releasing to the public just so you guys can really get the the early benefits of these products?

或者你们是怎么考虑的，比如在发布前内部测试工具，或是为了保持竞争优势而保密？

Or how do you guys think about that in terms of, like, testing tools internally before releasing them and or keeping it to yourself to get a leg up over competition?

是的。

Yeah.

我们会评估。

We look.

我们一直拥有大量实验性模型和工具。

We have we have lots of ex experimental models and tools all the time.

还有些工具目前因成本过高而无法大规模部署。

And we also have tools that are, at this time, too expensive to serve at scale.

比如Genie就是这类模型的典型例子。

You could think of like Genie as being an example of a model like that.

我们很希望能开放使用，但目前运行成本太高。

We would love to give access to that, but it's expensive to serve currently.

当然，我们正在通过未来版本的模型解决这个问题。

Obviously, we're working on that with future versions of the model.

部分深度思考模型仅限Ultra会员使用，因为它们的运行成本同样高昂。

Some of our deep think models are only available in ultra the ultra tier because they're also very expensive to serve.

所以我们持续在优化这些方面。

So we're continually trying to optimize for those things.

本质上这更多是算力的物理限制问题。

And then generally, we're able to, it's more of a physical constraint with the compute.

一旦从服务效率角度可行，我们通常会立即开放这些模型供所有人使用。

When we're able to, we generally put those models for everyone to use as soon as we're actually able to do it from a serving point of view efficiently enough.

所以这更像是主导因素。

So it's more that is the dominating factor.

当然，我们也一直有许多研究想法和模型在进行中。

We also, of course, do have lots of research ideas and research models going on all the time.

可以说，这是作为一个前沿实验室的常规工作——拥有非常深厚且广泛的研究基础，可能比其他任何实验室都更广更深。

That's part of the ordinary course of being a kind of frontier lab with a very deep and broad research bench, I would say, probably broader and deeper than anyone else's.

因此我们始终在努力开创下一个AlphaGo、下一个Transformers，探索未来的发展方向。

And so that we're always trying to pioneer the next AlphaGo, the next Transformers, what's coming down the line.

显然，世界模型就是其中之一。

Obviously, world models is one of those things.

所以我们一直在进行实验。

So we're always experimenting.

其中一些成果在准备就绪后，会向公众发布。

And some of those things, when they're ready, that will put them out into the general public.

此外还有其他方面，比如硬件和软件的开发进展，像眼镜、助手这类产品，我们也在不断迭代改进，在准备好向全世界展示之前会先进行实验性测试。

Then there's also other things, too, like hardware and software developments going on, things like you know, glasses, assistant and stuff like that, that we're also, you know, iterating on and and and starts off experimental before we're ready to to, you know, show the general the the world in general about it.

不过你们这些产品的发布速度是不是在慢慢加快？

Are you guys slowly getting quicker on these releases, though?

因为我注意到比如3.0版本，你们一上来就推出了搜索功能。

Because I noticed with three point o, for example, you guys are launching it in search off the bat.

是的。

Yeah.

所以你们是在逐渐加快节奏吗？

So are you guys just slowly getting faster to it?

你对此有什么想法？

How are thinking about that?

是的。

Yes.

对。

Yeah.

那确实是个绝佳的位置。

That's a great spot, actually.

所以我们工作得非常努力。

So we worked really hard.

我认为2.5版本是第一个真正实现这一点的版本，当时我们拥有了世界一流的模型——SOTA模型，并迅速深度整合到谷歌主要平台中。

I think 2.5 was the first real version of that where we had a world class model, SOTA model, and deep integrations into the main Google surfaces very, very quickly.

对吧？

Right?

我想你在IO大会上看到了这一点，这也是让很多人对IO大会印象深刻的原因。

And I think you saw that at IO, which is what I think a lot of people are impressed at IO.

我认为通过Gemini 3.3版本，我们正将其提升到新高度。

I think with three point Gemini three, we're taking it to the next level.

正如你所说，我们正通过搜索和AI模式等实现SIM功能。

And SIM shipping, as you say, with search and AI mode and so on.

我认为这就是发展方向。

And I think that's the direction.

过去几个月我们非常努力地工作，你可以把Google DeepMind视为谷歌的动力核心，对吧？

We've worked really hard over the last few months, I think, you can think of Google DeepMind as being the engine room of Google, right?

因此我们努力确保接入所有PA系统，并为每个重要产品提供支持。

So we've tried to make sure we're plugged into all the PAs and powering up every big product.

谷歌拥有众多出色的产品，从地图到YouTube再到搜索，当然还有Workspace。

And there's so many amazing products at Google from Maps to YouTube to search and, of course, Workspace.

我们希望所有这些——我们通过Gemini及底层模型所做的一切优秀成果——能够真正为这些产品赋能，创造出数十亿用户每天使用并喜爱的新功能和特性。

And we want all of those, the goodness of everything that we're doing with Gemini and the underlying models to really power amazing new capabilities and features in these products that billions of people use every day and love.

我认为我们现在正看到这个飞轮效应开始真正运转。

And I think we're seeing that flywheel really starting now.

我觉得我们仍处于这一演进过程的中期阶段。

I think we're still only midway in that evolution.

还有更多令人兴奋的进展即将到来。

There's a lot more exciting stuff to come.

而且我认为我们还可以走得更快。

And I think we can go even faster.

我的意思是，搜索功能就是我们理想模式的典范。

I mean, I think search is a poster child for how we want it to be.

现在我们需要将这种模式全面推广。

And then we now need to do that across the board.

说到谷歌整个生态系统中实用的应用，Gemini应用刚刚达到6.5亿月活跃用户。

Speaking of useful apps within Google's whole ecosystem, Gemini, the Gemini app has just hit 650,000,000 monthly active users.

是的。

Yep.

恭喜。

Congrats.

是的。

Yeah.

谢谢。

Thanks.

我们对此感到非常自豪。

We're very proud of that.

你们正在迅速赶上Jatzpreeti。

You guys are catching up to Jatzpreeti really quickly.

但我真的很好奇，以你们现在的规模，除了编程之外，Gemini应用中还有哪些具体用例对用户特别有用？

But I'm really curious, like, at the scale you're at now, is there any, like, specific use cases you're seeing across the Gemini app other than coding that are really useful to your users?

是的。

Yeah.

实际上我们发现，我认为Gemini应用在多模态方面表现非常出色。

We're seeing actually, I think I think the Gemini app is really good for multimodal.

所以我觉得Nano Banana对我们来说是一个重要的使用驱动力。

So I think with Nano Banana, that was a big driver of usage for us.

从策划惊喜生日派对邀请函这类有趣的事情，到在某些地区制作小雕像等等。

From very fun things like planning your surprise birthday party invites, whatever, to like in certain part territories, like making little figurines.

可以做很多有趣的事情，比如漫画创作。

So there's so many fun things one can do, comics.

因此我认为利用多模态能力是Gemini应用的一大独特优势。

So I think using the multimodal capability is something pretty unique that Gemini app is good at.

我认为这已经激发了很多兴趣，而且这种趋势还会持续下去。

And I think that's driven a lot of interest, and I think that will continue.

我认为我们正在加倍努力，深入思考教育和健康等领域，这些都是用户喜欢使用聊天机器人解决的问题，我们立志在这些领域做到业界顶尖。

And I think we're also doubling down and thinking through on things like education and health and other stuff that we know users like to use chatbots for, and we want to be absolute best in class in that.

我认为Gemini三将成为这一愿景的基石。

And I think Gemini three is going to be the foundation stone for that.

但对我而言，多模态功能最棒的是能用来头脑风暴，无论是项目命名还是创意验证。

But I think multimodal and at least for me, love brainstorming with these things, whether it's like naming a project of sense checking an idea.

我觉得这款应用在这方面也表现得非常出色。

And I think that the app is really good for that too.

你刚才提到Gemini可能成为健康咨询的基石，这个观点很有意思。

You said something there that was really interesting that Gemini might be like the cornerstone for the health questions.

能否详细说说这点？考虑到你的健康领域背景。

Is there any like more detail you could into that because obviously your background with health?

是的。

Yeah.

没错。

Yeah.

我们还有一系列配套项目，比如'合作科学家'这样的计划。

We've got all these sort of other projects, if you like, like co scientists.

我们在这方面做了大量工作。

And we've done a lot of work on this.

科学团队开发了一套名为AMI的医疗诊断系统。

We have a system called AMI, a medical diagnostic kind of system that are more in the science team.

我们希望将这些功能都整合到主Gemini平台中。

And what we'd like to do is bring all of those capabilities into the main Gemini.

这就是我们正在关注的地方。

So that's where we're looking at that.

我非常希望它能成为所有科学家用来激发灵感或进行研究工作的工具。

I would love it to be what all scientists use to kind of riff ideas on or to do some research.

我认为Gemini三号是奠定这一基础的绝佳选择。

And I think three, Gemini three is a good foundation stone for that.

你将开始看到这些能力的逐步展现，包括基于Gemini三号的各种形式，如深度研究和深度思考等功能的构建。

And you'll start seeing rolling out those capabilities, the various forms of Gemini three, including things like deep research and deep think that built on top of it.

但现在由于Gemini三号在推理和工具调用等方面带来的额外可靠性，这应该会在文献引用和理解中体现出来。

But now with the extra reliability that Gemini three has due to the reasoning and tool calling and so on, that should come through in citations and understanding literature.

再次强调，Gemini在这方面应该会表现惊人，因为它在多模态方面非常出色。

And again, Gemini should be amazing for that because it's so good multimodally.

许多健康和教育领域的问题，以及用户想用它做的事情，都是多模态的。

And a lot of health and education questions and what users want to do with it are multimodal.

这是一张诊断图像。

Here's a diagnostic image.

它意味着什么？

What does it what does it mean?

这是一篇论文。

Here's a paper.

这是图表和数据。

Here's the here's the figure and the tables.

结合文本来看这意味着什么？

What does that mean with the text?

反之亦然。

Or vice versa.

你知道，在教育领域，我得为这个主题制作一张海报。

You know, in education, I've got to make a poster about this subject.

你知道，帮我设计一下版面。

You know, help me lay that out.

对吧？

Right?

还要为它生成视觉效果。

And generate the visuals for that.

我想这就是我所期待的。

I think that is what I'm hoping.

我们期待人们使用Gemini三系统，当然主要包括Gemini应用。

And we're expecting people to use the Gemini three systems for and including, of course, the Gemini app primarily.

我对此非常期待，特别是在AI、医疗保健和教育领域。

I'm I'm very excited for that, especially on the AI and health care and education.

这两个领域都让我很感兴趣。

Both are very interesting to me.

是啊。

Yeah.

展望未来，再说一次，你可能没有答案。

Looking ahead, and again, you might not have an answer to this.

这算是比较长远的事了。

This is kind of further down the road.

但你们是否考虑过将AI用于主动预防性医疗保健？

But are you open to and or looking at using AI for proactive, preventative health care at all?

科学团队和医疗团队正在研究这个方向，类似于某种医疗级应用。

We are looking at that in the science team and health teams, like some kind of medical grade thing.

但这显然需要获得额外的审批，与监管机构进行制衡等等。

But that would need, obviously, additional approvals and checks and balances with the regulators and so on.

必须非常谨慎。

You have to be careful.

显然，Gemini应用并非医疗级工具。

Obviously, Gemini app is not a medical grade tool.

它仅供个人使用，你仍然需要咨询医生等专业意见。

It's for personal use and you still need to consult a doctor and all of those things.

但在世界上医疗条件或教育水平较差的地区，它可能会非常有用。

But it could be really useful in places where poor parts of the world where there isn't very good primary health care or education.

我们对此感到非常兴奋。

And we're very excited about that.

而且我知道你对此也很感兴趣，Rowan。

And I know you're very interested in this, Rowan, as well.

考虑到谷歌的覆盖范围、分发渠道以及Android等平台——这些在部分地区已是生命线般的存在——我们认为至少能为这类人群提供基础医疗知识和护理，这可能会改变他们的处境。

And so we think and because of Google's reach and distribution and Android and things like that, which already are kind of lifelines in some of those places, I think that we can could be very good to at least get a basic level of care and knowledge to those to those kinds of people and which could make a difference to them.

没错。

Right.

我认为我们可以持续优化这些功能，然后着眼于更专业的医疗级应用，看看何时能开发出类似医生助手、伴侣或研究助理这样的工具。

And I think we can continue to kind of improve that with our and then look to these more medical grade applications and see when are we ready for something to be like a doctor's assistant or companion or research assistant or something like that.

我认为我们还需要更多层次的可靠性。

I think we still need some more levels of reliability.

而且我认为我们在Gemini三号上方向是正确的。

And I think we're on the right direction with Gemini three.

但我得说，还有很多需要改进的地方。

But there's still a lot more, I would say, that's needed.

我们正在大力研究这方面。

And we're researching that heavily.

如你所知，这是我的核心热情所在。

As you know, it's a core passion of mine.

没错。

Right.

科学和医学领域正在使用我们的系统来实现这一目标。

Science and medicine are using our systems for that.

当然，我们希望Gemini能成为主要的基石。

And of course, Gemini, we'd love that to be the main foundation stone.

这就是我们为这些额外功能制定的建设计划。

And that is the plan for those additional capabilities to be built on.

所以我们对此感到非常兴奋。

So we're very excited about that.

如你所知，我对此充满热情，包括我们与isomorphic等机构的合作。

I'm very passionate about that, as you know, and also with our work with isomorphic and so on.

我对我们在Gemini三号上取得的进展感到满意，但这只是个开始。

And I'm pleased with the progress we've made with Gemini three, but it's just the beginning of that.

要想在这些应用场景中真正可靠，我们还需要更多努力。

We need a lot more if we want to be really reliable for those types of use cases.

明白了。

Got it.

是的。

Yeah.

这将帮助数十亿人。

It's gonna help billions of people.

我对此非常期待。

I'm very excited for it.

稍微转换下话题，谈谈Gemini在实际工作中的现状。

So switching gears here a little bit to just the real world work, what we currently have with Gemini.

发布时另一个亮点是应用内的Gemini智能体功能，这是个新特性，可以让你连接到Gmail等已有服务。

Another thing that stood out with the launch was Gemini agents within the app, which is new, which allows you to connect to things like Gmail, which was already there before.

但现在它能提供定制化步骤，并允许你直接在Gemini里执行发送邮件等任务。

But now it gives you tailored steps and actually allows you to execute tasks like sending emails directly within Gemini.

随着我们逐步实现这个几乎能成为你生活助手的AI，直接整合在Gmail里，你对这个数字同事的理想愿景是什么？

As we kinda get towards this AI assistant that is almost your life assistant directly in Gmail, what's your, like, dream vision for this, like, digital coworker?

你希望Gemini成为人们日常工作（比如Slack中）每天都会使用的独立助手平台吗？

Do you want Gemini to be, like, this standalone assistant type platform, like, that people use every single day in work, Slack?

还是说它只是个独立工具？

Or is it just, a separate tool?

没错。

Yeah.

我非常希望，你知道，我们有一个通用助手的构想，就是未来版的Gemini，它将在你日常生活的每时每刻都发挥作用，无论是工作事务还是休闲时光，都能为你推荐有趣内容、提供创意灵感、与你即兴互动，并且可能跨多种设备使用。

I I would love it to be you know, we have this idea of a universal assistant, which is, a future version of Gemini, right, that's useful in your everyday life, every moment of your life, where it's great assistant for anything you might be doing productivity wise, but also in your leisure time, recommending you cool things, giving you ideas about things, riffing with you on things and maybe also comes across multiple devices.

所以它会出现在你的电脑和浏览器上，既适用于工作场景也适用于家庭环境，同时还能随身携带在你的手机上，或许还包括智能眼镜等其他设备。

So it's on your computer and your browser, it's also it's a work, it's a home, but it's also comes with you on your phone and maybe some other devices like a smart glass.

我认为这就是未来。

And I think that is the future.

我对此深信不疑。

I feel very strongly about that.

这就是未来。

That's the future.

要实现这一点，我认为需要像Gemini这样真正强大的多模态基础模型，因为它必须理解你周围的物理世界和你所处的环境。

And I think you need a really capable base multimodal model like Gemini to be able to do that because you've got to understand the physical world around you and the context that you're in.

当然，还要能调用和使用所有其他应用程序，从谷歌那些出色的应用开始，比如地图、Workspace和邮件等，但最终会发展成完全通用的形态。

And, of course, to call and use all these other applications, starting with all the amazing Google ones, like Maps and Workspace and email and so on, but then eventually becoming fully general.

这样它就能调用任何工具。

So it can call any tools.

届时我们将进入一个新时代，就像现实生活中拥有一位出色的私人助理那样——我们中那些有幸拥有优秀私人助理的人都知道——把这种便利带给每个人的生活。

And then I think we're into a new era where you have just like if you have a really good personal assistant in real life, those of us lucky enough to have really good personal assistants, bringing that helpfulness to everybody's lives.

我认为这将为你如何安排时间带来真正的福音。

And I think that's going to be a real boon for what you want to do with your time.

所以最终，我希望通过这个方式我们能重新获得时间，更重要的是重新获得注意力空间，这样我们就能把精力真正花在自己热爱和想做的事情上，而不是那些不得不做的事情。

So eventually, I hope eventually we'll get time back through this, right, and our attention space back as well, even more importantly, so we can actually spend it on the things we love doing and we want to be doing rather than the things we just have to be doing.

我对此非常期待。

I'm very excited for it.

我想我们的时间就到这里了，非常感谢您抽出时间。

I think that's all the time we have here, but thank you so much for your time.

很好。

Great.

谢谢。

Thanks.