2025-05-24 Sergey Brin.CatGPT

Refer To：《Why AI is more important than the Internet (Interview with Google Co-Founder, Sergey Brin)》。

Speaker 1:

I would love if you could compare the moment that we are in right now in 2025 to the early days of the Internet. And I'll ask it as a sort of question where it's like, AI is to 2025 as the Internet is to what year?

发言者 1：

我想请您把我们现在（2025 年）所处的时刻与互联网的早期阶段作一番比较。我想用提问的方式来表达：如果说 “AI 之于 2025”，那 “互联网之于哪一年” 才合适？

Sergey Brin:

Okay, Internet history, I guess it was ARPANET and whatnot in the 70s. But the web kind of I think of it as born in Mosaic was launched in 93 or so if I remember correctly and then Netscape however many years after that.

谢尔盖·布林：

好吧，谈到互联网历史，大概 20 世纪 70 年代就有 ARPANET 之类的东西。但我认为真正的 “网络” 诞生于 Mosaic 浏览器——如果我没记错，是在 1993 年左右发布，随后几年又有了 Netscape。

I think in some ways you can draw parallels. I don't know. You could say with sort of, I don't know, Transformers in 2017, the first inklings of our new kind of language models. But I think it's really different in many ways.

我觉得在某些方面你可以找到一些相似之处。我不确定。你可以说 2017 年的 Transformer 算是我们新型语言模型的初露端倪。但在许多方面，两者又确实截然不同。

I mean, for one thing, the Internet was brilliant and enabled many things, but it wasn't, like, technically revolutionary, in the sense, you know, like, with the web, Tim Berners-Lee at CERN,

首先，互联网当然非常出色，催生了许多事物，但从技术层面来说它并不是那种“颠覆式革命”。比如说，CERN 的蒂姆·伯纳斯-李发明的万维网，

they were just organizing the scientists' data and stuff and sharing it, and they did a great job, and it took off as a viral organizational thing, and it's been phenomenal, don't get me wrong,

当时他们只是把科学家们的数据整理并共享出去，这件事做得非常好，于是这种架构以类似病毒式的方式迅速扩散，并取得了惊人的成就，这一点毋庸置疑，

but it wasn't, like, Nobody would have questioned whether that was physically possible five years before. There was no real limitations. In this case, we don't really know what intelligence is. We don't know how far we can take it.

但那种做法并不会让人质疑五年前它在物理上是否可行，并没有真正的技术极限。而在 AI 这里，我们对 “智能” 的本质尚不清楚，也不知道它最终能发展到什么程度。

I think a lot of people, myself included, are just surprised how quickly and how far it has ramped. So that's just an important distinction. We don't even know what the possible peak is.

包括我自己在内，很多人都惊讶于它的进展速度之快、幅度之大。这一点是一个重要区别——我们甚至不知道它的极限在哪里。

With the Internet, you could imagine everybody could communicate at high speed with everybody else. Every company would have a website, which they do now. But you could have realistically imagined that in 1990, let's say.

对于互联网，人们可以想象到人人高速互联，每家公司都会有自己的网站——事实如今确实如此。而在 1990 年左右，这种想象其实已经相当现实可行。

There were things before that, like Gopher, I'm probably dating myself, but there were things before the web that were kind of like that. But yeah, with AI, you don't know where the peak is or if there's a peak at all.

在那之前还有 Gopher 之类的东西——我说这个可能暴露年龄——它们在万维网之前就有些类似的功能。但对于 AI，你不知道它的高峰在哪里，甚至不知道是否存在峰值。

So that's one important difference. The other important difference, and for better or for worse, this is now ...gained, you know, profound international attention. I mean, the amount of resources and, you know,

这就是一个重要差别。另一个差别是——无论好坏——AI 现在已经获得了全球范围内深刻的关注。投入 AI 的资源数量，

money and compute and energy that are flowing towards AI is extraordinary. You know, the early days of web, we were a startup, you know, we whatever, we got a little seed loan, less than a million dollars, we're off to the races,

包括资金、算力、能源，都极为庞大。还记得互联网早期，我们只是一家初创公司，拿到不到一百万美元的种子贷款，就开始起跑了，

you know, I think we got like 10 million dollars in our venture round and that was that. You know, these days companies are spending billions and billions of dollars building, you know,

然后在风险投资轮中大概拿了 1000 万美元，就是这么回事。而现在的公司为了打造最好的 AI 模型，动辄投入数十亿美元，

the best AI models in the world and that's a good and bad thing, but it's definitely different. So I just think it's In as much as there are parallels, I just think we have no idea what this is going.

这既有好处也有坏处，但毫无疑问与当年完全不同。因而尽管存在若干相似点，我认为我们对它最终会走向何处根本没有概念。

Speaker 1:

So to that end, then, do you believe that AI is fundamentally more of a discovery than an invention? Do you feel like this is some emergent property of the universe and we're just stumbling upon it?

发言者 1：

既然如此，您认为 AI 从根本上说更像是一种 “发现” 而非 “发明” 吗？您觉得它是宇宙中的某种涌现属性，而我们只是在偶然间触及它吗？

Or is this like the ultimate test of human creation?

还是说，这更像是对人类创造力的终极考验？

Sergey Brin:

Well, I mean, I guess both of those things that you said. are distinct from what the web might have been. I mean, yeah, I think it's a discovery in the sense like we simply do not know what is the limit to intelligence.

谢尔盖·布林：

嗯，我想你提到的两点都与互联网不同。是的，我认为这更像一次 “发现”，因为我们根本不知道智能的极限在哪里。

Law that says, you know, can you be a hundred times smarter than Einstein? Can you be a billion times smarter? Can you be Google times smarter? You know, there's no I don't I think we have just no idea what the laws governing that are and So,

有没有什么定律规定，一个智能体能否比爱因斯坦聪明一百倍？一十亿倍？或者“谷歌倍”地聪明？我们对其背后的规律毫无头绪，

yeah, I guess you could call it a discovery of sorts maybe An analogy is sort of quantum computing where you don't really know How much computation are you really going to be able to get out of the universe?

因此可以说这是一种发现。可以拿量子计算来类比——你并不清楚最终能从宇宙中榨取多少计算力，

Kind of the basic laws of quantum mechanics suggests. It's extremely high but You don't know in practice if they're just other limitations. You don't know about right now.

量子力学的基本定律暗示极限非常之高，可实际中是否还会遇到别的限制，我们目前并不知道。

Speaker 1:

Yeah. Yeah Yeah, that's fascinating. But you ultimately think that AI is a more momentous and Discovery or invention than the Internet.

发言者 1：

嗯，确实很吸引人。但归根结底，您认为 AI 相比互联网更具划时代意义，无论是作为 “发现” 还是 “发明”？

Sergey Brin:

Yeah, I think the Internet was It was definitely very important, but it was sort of as much a kind of social development,

谢尔盖·布林：

是的，我认为互联网当然极其重要，但它更多是一种社会层面的发展，

like everybody agreeing to use these protocols and then kind of making their data and systems available for everybody else with, you know, TCP and IP and then, you know, HTML, HTTP,

比如大家共同使用 TCP/IP 协议，再到 HTML、HTTP，让彼此的数据和系统可以互通，

just agreeing on protocols and allowing it to grow and flourish. You know, maybe akin to how money was an invention thousands of years ago that let people, you know, really trade and stuff. But it's not like...

正是这种在协议上的共识让网络得以生长和繁荣。这或许类似于数千年前货币的发明，使得人们真正能够进行交易。但这并不是那种……

It's not neither money nor the internet are testing the limits of the universe.

无论是货币还是互联网，都没有在挑战宇宙的极限。

Speaker 1:

发言者 1：

But AI is.

但是人工智能就是如此。

Sergey Brin:

谢尔盖·布林：

But AI is, yeah, because we just... We don't know how intelligent things can be. We don't know, you know, we know some things about the brain. There's maybe, whatever, 100 billion neurons, 100 trillion synapses, and they run so fast.

但是人工智能确实如此，因为我们……我们并不知道智能可以达到什么程度。我们知道一些关于大脑的信息，譬如它可能拥有一千亿个神经元、一百万亿个突触，而且运转极其迅速。

But, you know, with our computers, can we, you know, simulate that? Can we go beyond that? And how far and what would that be like?

但是，你知道，用我们的计算机，我们能否模拟这种规模？能否超越它？如果可以，能超越到什么程度？那又会是什么样子？

Speaker 1:

发言者 1：

Yeah.

是的。

Sergey Brin:

谢尔盖·布林：

We just don't know.

我们真的不知道。

Speaker 1:

发言者 1：

I feel like in that way, the question of how do we approach this, what do we build, who should work on this are all questions that are philosophical just as much as they are technical or economic.

在我看来，关于我们该如何应对、应该构建什么、由谁来做，这些问题既是技术或经济层面的，也是哲学层面的。

Sergey Brin:

谢尔盖·布林：

100%. Consciousness is another kind of thing that gets brought into it. The Internet didn't raise questions of consciousness, for example, but if this AI is smart enough and self-aware enough, Does that matter? What does that mean?

确实如此。意识也是被牵涉进来的另一层问题。比如互联网从未引发对意识的质疑，但如果这套 AI 足够聪明并具有自我意识，那是否重要？这意味着什么？

I don't know.

我也不知道。

Speaker 1:

发言者 1：

Yeah. You started Google famously in a garage in Menlo Park with Larry Page in 1998. You were just two guys trying to build something because you saw an opportunity.

是的。1998 年，你和拉里·佩奇在门洛帕克的一间车库里创办了谷歌。那时，你们只是两个年轻人，因为看到了机会而想要打造些东西。

Now Google is a \$2 trillion company, which I hope I have this right, 180,000 employees. Plus or minus. Plus or minus.

如今谷歌是一家市值约 2 万亿美元、拥有大约 18 万名员工的公司，大致如此。

Sergey Brin:

谢尔盖·布林：

You track it as well as I do.

你对这些数据的追踪不亚于我。

Speaker 1:

发言者 1：

I'm sure riding the AI wave, so to speak, with all that infrastructure has many advantages. But I'm curious if there's any part of you that maybe just 1% wishes you were 20 years old again,

我相信凭借这些基础设施，乘着 AI 浪潮确实有许多优势。但我很好奇，你心里是否有哪怕 1% 的成分想要再回到 20 岁——

just graduated from Stanford, it was just two guys in a garage.

刚从斯坦福毕业，只是两个年轻人在车库里创业。

Sergey Brin:

谢尔盖·布林：

Oh, that's a good question. I mean, look, I'm just grateful As a computer scientist to be alive at any age during this time, I think if you were to walk across the street,

哦，这是个好问题。说实话，作为一名计算机科学家，能活在这个时代、无论处于什么年龄，我都感到非常感激。我想如果你走到街对面，

I don't know maybe you can talk some folks into letting you do that later, you know and you see how all the AI researchers are all gathered kind of around around the coffee machine and whatnot and you know everybody's excited.

也许你可以说服一些人晚点带你去看看——你会看到所有 AI 研究人员聚在咖啡机旁，人人都兴奋不已。

I mean it is a very startup-like feel. Technically, by the way, when we started, we had the garage and a couple bedrooms. We did have the garage, but it wasn't just a garage.

那种氛围非常像创业公司。顺便说一句，当年我们除了那间车库，还租了几间卧室。确实有车库，但不止于车库本身。

Speaker 1:

发言者 1：

My greatest fear was getting a historical detail wrong, so I'm glad that's done.

我最担心的是搞错历史细节，能澄清这一点我很高兴。

Sergey Brin:

谢尔盖·布林：

We tell the story kind of like a garage, but there were a couple rooms in addition, which helped a lot. I mean there is a very entrepreneurial sense, I think.

我们常把它说成是在车库创业，但其实还有几间房间，这帮了很大忙。我认为那种创业气息非常浓厚。

Given the kinds of compute requirements that are required to compete at the forefront right now and kind of the amount of science that goes into it,

考虑到如今要在前沿竞争所需的计算资源规模，以及背后投入的科研量，

it would be really hard to try to make a lot of headway, at least on the foundation model side, as a couple guys in the garage. Lots of folks in garages I can use these models to create new and amazing things.

至少在基础模型层面，靠几个在车库里的人要取得重大突破会非常困难。不过，很多身处车库的人完全可以利用这些模型创造出新的惊艳成果。

And I don't want to discount the possibility that somebody's just going to have some idea that's so brilliant that even in a sort of a couple people in the garage could pull it off.

当然，我并不排除有人会想到极其精彩的点子，以至于即便只有两三个人在车库里也能实现它的可能性。

But it seems like the frontier is being pushed by Pretty big companies like ours, I think, we're now at the frontier. I'm very proud of the progress we've made over the last year.

但如今看来，真正推动前沿的还是像我们这样的大公司。我认为我们位于最前沿，对过去一年取得的进展感到十分自豪。

Honestly, I'm really grateful to be able to be a part of that. I don't think I would take that teleportation to my younger self just yet.

说实话，能参与其中我非常感激。我想我暂时不会选择把自己传送回年轻时代。

Speaker 1:

What is the most sci-fi sounding thing that you actually believe has a decent chance of becoming real, let's say, in the next 10 years?

发言者 1：

在未来 10 年里，您认为听起来最具科幻色彩、但实际上有相当机会成真的事情是什么？

Sergey Brin:

I think the most exciting will be Gemini. Making some really substantial contribution to itself in terms of you know machine learning idea that it comes up with maybe implements and to develop the next version of itself.

谢尔盖·布林：

我认为最令人兴奋的将会是 Gemini——它可能为自身作出真正实质性的贡献，例如提出某些机器学习构想、亲自实现它们，并据此打造下一代的自己。

We already use Gemini a lot during like pieces like some AI researcher will be like oh I need to you know debug this code for me or You know, help me with this math or something like that. Those sort of one-offs.

我们现在已经常常用 Gemini 来处理零散任务，比如某位 AI 研究员说“帮我调试这段代码”或“帮我算一下这道数学题”，这种一次性的请求。

But as a sort of really substantive, some kind of new significant breakthrough. That the AI itself makes, I think that, to me, that's science fiction and I think it could well happen.

但如果 AI 能够凭自身实现某种真正意义上的重大突破——对我来说，这才是科幻，而我认为这完全可能发生。

这个定义比手下的CEO要好的多。

Speaker 1:

Yeah, if you had to ballpark guess, when do you think Gemini will create the next version of Gemini?

发言者 1：

那么，粗略估计一下，您觉得 Gemini 何时会创造出下一代的 Gemini？

Sergey Brin:

I mean, like I said, it's already assisting, that's already happening, but from like kind of some kind of from scratch sort of rewrite? I don't know, that's a tough question.

谢尔盖·布林：

就像我说的，它已经在辅助我们，这件事正在发生。但要让它从零开始彻底重写？这就很难说了。

I don't know if I don't know how high of a priority that is to some extent, because we sort of can guide it. You know, at what point is it possible? Maybe in three years, let's say, three, four years.

至于这件事在优先级上到底有多高，我也拿不准，因为我们可以对它加以引导。什么时候可能实现？或许三年？三四年左右吧。

I don't know if the vision, the virtue that it would make just by itself would be quite as good as itself, but You know, if you think about it with our new video model, the VO3, I mean...

我不确定它单凭自己造出的版本能否与当下一样好。不过，想想我们的新视频模型 VO3，我的意思是……

Speaker 1:

Which, by the way, brought me to tears in the demo area. Just a few.

发言者 1：

顺便说一句，我在演示区看 VO3 时都被感动得流泪了，真的。

Sergey Brin:

In a good way, I hope.

谢尔盖·布林：

希望是感动得好的那种泪水。

Speaker 1:

In a very good way.

发言者 1：

是非常好的那种。

Sergey Brin:

Okay, just checking.

谢尔盖·布林：

好的，我只是确认一下。

Speaker 1:

The sound, something about it just came to me.

发言者 1：

那个声音，某种细节一下子击中了我。

Sergey Brin:

Yeah, no, sound is such a huge deal.

谢尔盖·布林：

是的，声音确实非常关键。

Speaker 1:

I didn't realize how much that was missing until it was all there and it just, yeah, hit me like a ton of bricks.

发言者 1：

之前我没意识到它缺少了多少要素，直到演示全都呈现出来——那一刻真的像一块巨石砸在我心里。

Sergey Brin:

Oh, well, thank you. But you know in theory,

谢尔盖·布林：

哦，那太好了，谢谢。不过理论上说，

I guess I've never tried this but well first of all you don't I know if we support this in the user interface But you don't have to give it a prompt Like it will just generate a video our user interface might not actually support not having a prompt But then you would have no idea of what it's gonna make.

我倒没真这样试过——首先我不确定我们的界面是否允许——但理论上你可以不给它任何提示，它也能直接生成一段视频。只是如果没有提示，你完全不知道它会产出什么。

Yeah, you could just say make a Good video, I guess. That would be sort of the model just by itself going. But generally, I think when it's directed by a person, and presumably that's what you did,

是的，你大概也可以只对它说“做一段好视频”，让模型自行发挥。但总体来说，我认为有人类给出方向会更好——我猜你当时就是这样做的，

you gave it some prompt, some bargain, you know, you get really good results. So I guess I'm just saying if Gemini creates the next great version of Gemini, I think For the sort of foreseeable future,

你给它一些提示、一些约束，就能得到非常棒的结果。所以我的意思是，如果 Gemini 打造出下一个卓越版本的 Gemini，在可预见的未来，

it will probably do better if there's a human that sort of guides it, at least at some high level way. I mean, it could conceivably someday blank slate, just go to town, do everything without any guidance.

只要有人类在高层面做些引导，它的表现大概会更好。当然，也不排除有朝一日它能“白纸起家”，全程自洽、自主完成一切。

But yeah, that's level sci-fi I don't think we've gotten to yet.

不过，那种级别的科幻我们还没有真正触及。

Speaker 1:

Got it. So for the foreseeable future, you do foresee a world in which it's the Google employees helping the AI along to build the next versions of Gemini, the future.

发言者 1：

明白了。那么在可预见的未来，您确实预见到一个由谷歌员工协助 AI 打造 Gemini 后续版本的世界，对吗？

Sergey Brin:

Yeah, that's right.

谢尔盖·布林：

是的，没错。

Speaker 1:

I'm curious how much of your time and energy you spend on sort of these like The deeper, more meaty philosophical questions of AI versus, I'm sure, all of the technical questions, the practical questions, the business questions.

发言者 1：

我想知道，面对 AI 的那些更深层、更核心的哲学问题，与所有技术、实践和商业层面的事务相比，您会投入多少时间和精力？

How much of your energy goes towards all of that, given that it feels like we're discovering some fundamental underlying capability of the universe, at the same time as building cool tech?

考虑到我们在构建酷炫技术的同时，似乎也在揭示宇宙潜在的基本能力，您为此投入了多少精力？

Sergey Brin:

I mean probably not that much goes to the philosophical questions just as a practical matter. It's just like There are so many technical details you need to get right on the way there.

谢尔盖·布林：

实际上，投入到哲学问题上的精力可能并不多，因为从实践角度看，前进过程中需要处理的技术细节实在太多。

I'm fretting about our being able to sign up for whatever Ultra and VO3 and all the things that aren't quite working as I'd hoped and I'm right now hassling the engineers and the product managers about all the little snafus.

我担心诸如 Ultra 和 VO3 之类的服务上线进度，以及那些未如我所愿顺利运转的部分；此刻我正忙着催工程师和产品经理解决各种小问题。

I mean, it's definitely nice to take a step back. Some of the philosophical questions do kind of emerge out of the technical details. Like, you know, we have some new model. How are we going to evaluate it? Let's say.

当然，偶尔退后一步是很好的。有些哲学性的问题确实会从技术细节中浮现出来。例如，我们有了新的模型，那我们该如何评估它呢？

Like, what does it mean for the model to be good? We have sort of standard benchmarks and things. At some point, the models tend to get really good at those benchmarks. And, you know, every time you need to kind of redesign that,

比如说，模型“好”到底意味着什么？我们有一些标准基准，但模型终有一天会在这些基准上表现极佳，于是你就需要重新设计评估体系，

you do take a step back kind of philosophically and try to figure out, you know, what is important. When you have new kinds of AI models, the diffusion model, for example, that you can now play with,

这时你会在哲学层面后退一步，思考真正重要的是什么。比如现在有了新的 AI 类型——扩散模型——你可以操作它们，

text diffusion, that's not like an apples to apples thing. So now we're kind of asking, well, How do we compare a thing that doesn't, you know, go left to right, that's actually kind of spinning the whole thing out at once?

文本扩散并不是左右递归那一套，因而我们会问：如何比较一种不是从左到右、而是一次性生成整段内容的模型？

How do we measure that compared to our kind of normal autoregressive models? So a lot of these things bring up some philosophical questions. You are pretty grounded and nose to the grindstone, no pun intended, answering them.

我们怎样将其与传统的自回归模型进行度量？这些问题往往引出哲学层面的思考，而你又必须脚踏实地、埋头应对它们——双关语并非有意。

这种技术能力可能不是黄仁勋或者扎克伯格具备的，和Ilya Sutskever处于一个级别，甚至更好。

Speaker 1:

It's rooted in such practicality. It's not theory about what other people are working on when you're actually in the lab building.

发言者 1：

这一切都扎根于实践，而不仅仅是理论；当你真的在实验室里动手时，可不是探讨别人做什么。

Sergey Brin:

It would probably be great for me to be able to spend more time on some of the philosophical questions. But there's a lot going on.

谢尔盖·布林：

如果我能花更多时间探讨哲学问题当然很好，但现在的事务实在太多。

Speaker 1:

There's a lot there. I don't know how much time we have left, but one question I do want to make sure I get the chance to ask is what's a question or a topic that you wish people like me, interviewers, ask you more about?

发言者 1：

的确事情太多。我不知道我们还剩多少时间，但我想确保能问到：有没有什么问题或话题，是您希望像我这样的采访者能多问您的？

Sergey Brin:

A question or a topic that I wish people would ask me about? Oh my gosh.

谢尔盖·布林：

我希望人们多问我什么问题或话题？天哪。

Unknown Speaker:

Hmm.

未知发言者：

嗯。

Sergey Brin:

Okay, I have to formulate this as a question or I guess I could just sort of answer the question. Maybe that'd be easier.

谢尔盖·布林：

好吧，我得把这话组织成一个问题，或者我可以直接给出答案，可能那样更简单。

I guess it's this overall idea that people sort of react to let's say new AI announcements whatever we announced a bunch of things yesterday and now you know what can you do with those things and there's a bunch of cool stuff you can do those things and then there's a bunch of stuff you can't or it's not quite right But I think the interesting question is what are you going to be able to do with kind of those things next generations in one year,

我想说的是这样一个整体想法：人们对于新的 AI 公告会立刻做出反应。比如我们昨天发布了一堆东西，大家就开始思考这些东西能做什么——有很多很酷的用法，也有很多做不到或还不完善的地方。但真正有趣的问题在于，一年后、下一代技术出来时，你将能用它们做什么？

in two years. And that's that brings, you know, all kinds of exciting questions. I mean, language models two years ago made so many very embarrassing errors. It was kind of like Wow this thing actually did this correctly,

或者两年后能做什么。这会引出各种令人兴奋的问题。要知道，两年前的语言模型常常犯下令人尴尬的错误。当它偶尔做对时，人们会惊叹：“哇，它居然正确完成了！”

and that's super cool That's very different than oh my gosh I can actually use this as a tool for whatever it is because I You know if it's you know gonna be right 20% of the time I can you know whatever post a tweet with it Which is like wow that's cool,

这种“太酷了”的感觉，与“天哪我可以把它当工具用”截然不同。如果它只有 20% 的正确率，你最多用它发条推文，这固然酷，

but I can't actually use it day to day and Nevertheless, when you kind of look at the trend, you probably will be able to use it day to day with reasonable reliability.

但无法日常依赖。然而从趋势来看，你很可能在不久的将来就能以相当可靠的方式每天使用它。

I think as people kind of think about where they're going to plug these tools in to whatever it is they're trying to do. You know, you're trying to make a movie. VO is cool in that. It has sound now.

我想，当人们考虑将这些工具嵌入自己的工作时——比如你要拍电影——VO 现在很酷，因为它已经有声音了。

You know, maybe a year ago you've been like, well, it doesn't have the sound, it'll be a pain. We've done some work for continuity of characters and things like that, but it's, and we actually have some movies we are making with it,

一年前你可能会说：没声音会很麻烦。我们为角色连贯性等问题做了一些改进，实际上也在用它制作一些影片，

but it's probably still not ideal for making like some kind of two-hour film. But nevertheless, I think when you look at how far these tools have come in the last couple years,

但它可能仍不适合拍一部两小时的长片。尽管如此，当你回顾这些工具过去几年的进步，

you know, all the video models, not just ours, but you know, all of them, and you forecast in a couple more years, boy, you're going to be able to do a lot of really interesting things with them.

无论是我们的还是其他的视频模型，再往前展望几年，你会发现它们将能实现许多非常有趣的事情。

Speaker 1:

Can you give some examples of that? Because one of the arguments I've heard is like, By far, video is the most expensive. If you look at all the different modalities here, video is the most computationally expensive.

发言者 1：

能举些例子吗？我听到一种说法是，在所有模态中，视频无疑最昂贵——从计算消耗来看，视频开销最大。

And what's the practical application? Fun videos on YouTube, you know, AI brought on TikTok, like, there's a lot of controversy or conversation amongst,

那么它的实际应用是什么？YouTube 上的有趣短片、TikTok 的 AI 视频？围绕这点有许多争议和讨论，

you know, folks about why are we investing so much time and energy into doing it besides the fact that it's really, really cool. Can you share what are some of those other practical applications of these video models?

人们质疑：除了因为它“非常酷”，我们为什么要投入这么多时间和精力？您能分享这些视频模型的其他实际应用吗？

Sergey Brin:

Yeah, I mean, like I say, I mean, I think it's like, The difference between a cool toy and sort of a useful tool, you know, is sort of a matter of time and it happens gradually. You know, we're trying to aim for the useful tool.

谢尔盖·布林：

正如我所说，“酷玩意”和“实用工具”之间的差异是时间问题，会逐步转变。我们的目标就是让它成为实用工具。

We have some film producers that are here. I think Darren Aronofsky might have already spoken to, I don't know if he was on panel or something, but he's making a video. I have a close friend of mine, Dustin, is making a video.

这里有一些电影制片人。达伦·阿伦诺夫斯基可能已经在某个讨论环节谈到过，他正在拍一部影片。我的好友 Dustin 也在拍片。

But like, you know, some real artists are using these tools. But it's early days. I mean, they're obviously ...dealing with something that two years from now movie directors will think was kind of a joke,

一些真正的艺术家正在使用这些工具。现在还很早期——他们显然在处理一些两年后电影导演会觉得“简直是玩具”的东西，

but they're putting up with it to be, you know, on the frontier. And I think these models will be capable of producing really compelling videos.

但他们愿意忍受这些限制，只为了站在前沿。我相信这些模型终将能生成真正引人入胜的视频。

They'll be able to do it, you know, in concert with, you know, human directors, human actors and so forth. I think that Aronofsky film does have, has a combination of like, you know,

它们将能与人类导演、演员等协同创作。我记得阿伦诺夫斯基的那部影片就结合了

real life sort of acting combined with AI generation in a pretty cool way. But I you know look we today obviously we have the Industrial light magic all the you know Lucasfilm, you know, they do all the special effects, you know,

真人表演与 AI 生成内容，以非常酷的方式混合呈现。如今我们已经有工业光魔、卢卡斯影业等负责特效，

we already use technology to generate film this is just sort of a new dimension in that and obviously Early days and maybe it's not you know, the best resolution. It's not like the best.

电影制作本就依赖技术生成，而这只是其中的新维度。当然现在仍处早期，分辨率可能不够高，各方面还未最佳，

over a long period of time or whatnot, but I think you'll see all those things come along. We're trying to push the envelope so these things become real tools, not just toys.

但随着时间推移，你会见到这些问题都被解决。我们正努力突破极限，让它们成为真正的工具，而非仅仅是玩具。

Speaker 1:

Yeah, yeah. One of the things I always say on my platform is whatever you think about it today, this is the worst that it will ever look.

发言者 1：

是的，是的。我在我的平台上常说：无论你今天怎么看，它现在已经是最糟糕的状态了。

Sergey Brin:

Yeah, that's right. That's exactly correct.

谢尔盖·布林：

是的，没错。完全正确。

Speaker 1:

Okay, so I got the signal for one more question. So I'll ask this one on behalf of all of the, you know, the builders out there who are so excited about this moment, but are not a Google employee. They're not working at a Frontier Lab.

发言者 1：

好的，我收到提示还可以再问一个问题。我代表所有对这一时刻充满激情、却不是谷歌员工、也不在前沿实验室工作的建设者们来提问。

Is there any sort of direction that you would point these people into, whether they have a software engineering background or are just simply somebody who understands the depth and the importance of this moment and wants to be involved?

您会为这些人指明任何方向吗？无论他们是拥有软件工程背景，还是仅仅理解这一时刻的深度与重要性并希望参与其中？

Is there a direction that you would point them into to say, start here? Or maybe, is there a direction you would say, don't go in that way, don't waste your time?

有没有一个“从这里开始”的方向？或者，有什么路径您会劝他们不要走，以免浪费时间？

Sergey Brin:

There are so many great ideas, so I would be reluctant to stop anybody from doing anything because you never know. I think there's actually an increasing amount of really interesting academic work, even outside of the big labs.

谢尔盖·布林：

精彩的想法太多了，我并不愿意阻止任何人去尝试，因为谁也无法预料结果。我认为，即使在大型实验室之外，也有越来越多非常有趣的学术研究涌现。

And I think it has sort of happened with this advent of reasoning models that that happens more in the what we call post-training kind of reinforcement learning step,

随着推理模型的出现，这种情况更多地发生在我们称之为“后训练”的强化学习阶段，

which is more manageable with the amount of compute resources that lots of academic institutions or small companies have.

而这一阶段所需的计算资源，许多学术机构或小型公司都能够承担。

And there are lots of actually open weight models including our Gemma models that you can use to experiment with that. Increasingly, I think you'll see some kinds of reinforcement learning APIs from the top models so that,

现在有许多开源权重模型——包括我们的 Gemma——可供大家实验。我认为，你将越来越多地看到顶级模型提供某种强化学习 API，这样，

you know, you can send us your problems maybe with correct solutions or graders and, you know, we can contribute your problems to the mix of, like, if you want the model to be good at that. So I think that kind of thing is coming.

你可以把带有正确答案或评测器的问题发给我们；如果你希望模型擅长这方面，我们就能把你的问题纳入训练集合。我觉得这种方式即将到来。

Yeah, actually, I think it's pretty good time to be able to make an impact without having to train up a foundation model.

所以，现在正是无需训练基础模型也能产生影响的好时机。

Speaker 1:

Well, thank you so much for your time.

发言者 1：

非常感谢您抽出时间。

Sergey Brin:

Thank you.

谢尔盖·布林：

谢谢。

2025-05-24 Sergey Brin.CatGPT

2025-05-24 Sergey Brin.CatGPT

热门主题

Related Articles

2025-05-22 Sergey Brin.Google IO

2025-05-24 Sergey Brin.Google for Developers

2014 Sergey Brin’s Letters to Alphabet Shareholders

2025-05-20 Sergey Brin.All-In Podcast

2017 Sergey Brin’s Letters to Alphabet Shareholders