A Genie That Can Grant Any Wish

I think this will be the greatest thus far technological achievement of human history.

我认为这会是人类历史上迄今为止最伟大的技术成就。

But the only way that it really matters is if it makes people's lives like much better than they otherwise would have been.

但它真正算数的唯一前提，是让人们的生活比原本好得多。

We are about to create a genie that can grant any wish.

我们即将造出一个能实现任何愿望的精灵。

Because I think people will have such creative wishes and such incredible ideas of what they ask AI to help build, but concentration of power with AI is a terrifying thing.

因为我觉得人们会许下极有创造力的愿望，会想出各种不可思议的点子让 AI 帮着造出来；但 AI 带来的权力集中是件可怕的事。

I don't think anyone should want to live in a world of, you know, AI overlords or company that is the rough equivalent of that.

我认为谁都不该愿意活在一个由 AI 霸主、或者与之大致等同的某家公司主宰的世界里。

I think it's critical we preserve that spirit with AI and that we all collectively have the ability to self-determine our future.

我认为关键在于，我们要在 AI 这件事上守住那种精神，让我们所有人共同拥有决定自己未来的能力。

So Sam, you wrote a post that I thought was very simple and really interesting and a good place to start.

Sam，你写过一篇文章，我觉得写得很简单，也很有意思，适合作为开场。

Roughly, the last year's been really tough and that's somewhat my fault and the next year is going to be maybe our best 12 months.

大意是：过去一年真的很难熬，这多少是我的责任；而接下来这一年，可能会是我们最好的 12 个月。

Yeah.

嗯。

I'd love you to reflect on on both.

这两点我都想听你聊聊。

Maybe starting with why you said the first part and and why you believe the second part.

或许可以先说说，你为什么会讲第一句，又为什么相信第二句。

On the first part, I think we just were doing too many things.

先说第一点，我觉得我们就是做的事情太多了。

We're not focused enough and they were actually all good things to do, but the trick is we're in this like unbelievable moment in history where you can only do the very few great things.

我们不够专注。那些事其实每一件都值得做，但难就难在，我们正处在一个不可思议的历史时刻，你只能做那极少数几件真正伟大的事。

So we spread ourselves too thin and then made a bunch of difficult decisions to really refocus on having the best most abundant most cost-effective intelligence and empowering the world to build incredible things with that.

所以我们摊得太开了，后来做了一堆艰难的决定，重新聚焦：拥有最好、最充裕、最具成本效益的智能，并让全世界用它造出了不起的东西。

Since doing that, uh I think our progress has been remarkable and just given what we see in the pipeline will be much more remarkable over the next 12 months and the quality of the models that we'll have, the products that we can build around that to really let people thrive with this technology in in new ways.

自那以后，我认为我们的进展相当亮眼；而且从我们看到的在研管线判断，未来 12 个月会亮眼得多——我们将拥有的模型质量，以及围绕它能做出的产品，会真正让人们以全新的方式借这项技术蓬勃发展。

Uh it should be pretty awesome.

应该会相当棒。

Was there a moment last year that something clicked for you that caused you to change directions or restack priorities or something?

去年有没有某个瞬间，你突然想通了什么，于是转了方向、重排了优先级之类的？

If you go back to the beginning of 2025, just a year and a half ago,

回到 2025 年初，也就是一年半前，

yeah,

嗯，

the big concern was companies like OpenAI are buying up so much compute, is the revenue going to be there?

当时最大的担忧是：OpenAI 这类公司买下这么多算力，收入跟得上吗？

Is the demand going to be there?

需求跟得上吗？

And so we were trying to think about like a lot of things such that if the revenue growth took longer to materialize than we thought it might, we could have, you know, consumer apps and media and all these other things that could help us monetize the GPUs that we were signing up for.

所以我们当时在盘算很多事：万一收入增长兑现得比预想的慢，我们还能靠消费级应用、媒体这些别的业务，把签下的那些 GPU 变现。

Uh again it sounds ridiculous now because the revenue growth in the industry has been so steep but that was the big change.

现在再看这想法很荒唐，因为整个行业的收入增长陡得不行，但那就是最大的转变。

And then as soon as we realized like okay the model trajectory is growing so fast there's such a clear economic return on these models that was when we said you know we know what to focus on.

一旦我们意识到模型的进步曲线涨得这么快、这些模型的经济回报如此明确，我们就说：好，我们知道该聚焦什么了。

I was reading some of your your great old posts from prior to OpenAI and one of them is this notion of like so much discussion of focus and the right amount of things to focus on.

我读了你在 OpenAI 之前写的一些很棒的旧文，其中一篇大量讨论专注，以及到底该同时聚焦多少件事。

Is it one?

是一件？

Is it five?

五件？

Is it three?

还是三件？

How do you calibrate that in a business like this, especially in this period where you've said you needed to refocus?

在这样一门生意里，你怎么校准这个数字？尤其是在你说需要重新聚焦的这段时期。

Fundamentally, our business is to sell AI that people will build incredible products and services for each other with.

从根本上说，我们的生意就是卖 AI，让人们用它为彼此造出了不起的产品和服务。

The components that I think of as going into that are we have to train great models that work in all the ways people want to use them.

在我看来，这里面包含几个部分：我们得训练出优秀的模型，覆盖人们想用它的所有场景。

So great at coding, great at other kinds of knowledge, work, great at doing science, like where the real economic value is.

写代码要强，其他各类知识工作要强，做科研要强——真正的经济价值就在这些地方。

We have to produce or partner with these chips and systems, these, you know, hugely expensive racks that can do the AI computation.

我们得自己造、或者合作拿到这些芯片和系统，就是那些能跑 AI 计算的、贵得离谱的机架。

Uh we have to find enough uh land power data center shells to be able to put those racks somewhere.

我们得找到足够的土地、电力和数据中心厂房，好把这些机架放进去。

And then eventually or maybe pretty soon, we have to build robots that can automate that process to continue to drive the cost down, the cost of producing electricity, chips, the whole supply chain.

然后，最终、或者也许很快，我们得造出机器人来把这个过程自动化，继续把成本压下去——发电的成本、芯片的成本、整条供应链的成本。

And that kind of whole stack of making the best the most abundant uh the most useful AI that we can and making it something like electricity that just seeps throughout the entire economy and empowers people.

这一整条技术栈，就是尽我们所能做出最好、最充裕、最有用的 AI，并让它像电一样渗透进整个经济，赋能每个人。

That's kind of what I think we have to focus on.

我认为这差不多就是我们必须聚焦的东西。

Building every vertical application on top of that trying to go like eat every startup, eat every company.

至于在这之上把每个垂直应用都做了，去吞掉每一家创业公司、每一家企业——

No interest in doing that.

我们没兴趣。

Uh really want to just provide that platform.

真的只想提供那个平台。

This compute thing is one of the most interesting thing that's happened in human history.

算力这件事,是人类历史上最有意思的事情之一。

I think and it's obviously coming to a head and maybe will be coming to a head for a long period of time.

而且它显然正在走向一个临界点,而这个过程可能还会持续很长一段时间。

This is something that I think Dario called you the YOLO CEO when you were doing some of this early compute allocation and and securing the compute.

我记得你早期做算力调配、去锁定算力的时候,Dario 管你叫“YOLO CEO”。

Obviously now you're in this position where everyone is short this stuff is trying to find it.

而现在你处在这么个位置:所有人都缺算力,都在到处找。

And I'd love to hear the early stories about why you gained conviction that you needed to secure everything that you did, how you did it.

我特别想听听早期的故事:你为什么会笃定必须把那些算力全锁下来,又是怎么做到的。

like it it seems to have been proven right and maybe maybe you even underdid it right which is kind of crazy if you look at the headlines from back then.

现在看,这个判断像是被证明对了,甚至你可能还做少了——回头看当年那些新闻标题,这挺疯狂的。

Can you tell me the early story of like how you came to that conclusion and what gave you the conviction to do it despite everyone thinking it was crazy?

能讲讲早期的故事吗?你是怎么得出那个结论的?在所有人都觉得这很疯狂的时候,是什么给了你笃定?

We could just tell that we were on this exponential of model improvement.

我们当时就是能看出来,模型能力在走一条指数曲线。

That part we were very confident about and we knew it was going to keep going.

这一点我们非常确信,而且知道它会一直走下去。

We were pretty sure although as you mentioned we underestimated that as the models got better and better if we could continue to drive cost down that demand for AI at a sufficiently high level and a sufficiently low price was basically uncapped.

我们也比较确信——虽然像你说的,我们还是低估了——随着模型越来越好,只要我们能持续把成本压下去,那么能力足够强、价格足够低时,对 AI 的需求基本上没有上限。

This was just like a rare kind of new commodity for the world.

这就像是世界上出现了一种罕见的新型大宗商品。

Um but that what people would do with it reminded me of the way people used to talk about the early days of computing.

但人们会拿它做什么,让我想起早年大家谈论计算机的方式。

People said, "Oh, there's, you know, a market for five computers in the world" was one famous thing.

有人说过一句很有名的话:“全世界计算机的市场大概就五台。”

Or, you know, no one needs more than x amount of RAM.

或者说,没人会需要超过多少多少的内存。

Human ingenuity, creativity, desire for stuff, desire to be useful.

人的巧思、创造力、想要东西的欲望、想变得有用的欲望。

That's a very good thing to bet on.

这是非常值得押注的东西。

And we could see that AI was going to be an extremely important way that people expressed those things or got those things, did those things.

我们能看出来,AI 会成为人们表达这些、获得这些、实现这些的一条极其重要的路径。

And we knew that the algorithms would get more efficient and the models would get better, which of course they have.

我们也知道算法会变得更高效、模型会变得更好——事实也确实如此。

But we also knew that no matter how efficient they got, you know, at some level what we are about is turning electricity into useful intelligence and we were going to need more of that no matter how good we got that other layer.

但我们同样知道,不管效率提到多高,某种意义上我们做的事就是把电变成有用的智能,所以不管另外那一层做得多好,我们都会需要更多的电。

Given this observation about demand, we were just going to want more.

考虑到我们对需求的这个判断,我们只会想要更多。

Did that start with GPT-3?

这是从 GPT-3 开始的吗?

Like if I were to trace the history of this as far back as possible, where would you put the first hash mark?

如果要把这段历史尽可能往前追,你会把第一个刻度打在哪儿?

I would say we got real conviction with GPT-4.

我会说,真正的笃定是从 GPT-4 开始的。

Not even 3.5.

连 3.5 都还不算。

What was it?

具体是什么?

It was seeing the model was smart enough that we knew we'd be able to figure out an approach that worked for reasoning and then a belief that if we got reasoning to work that would bring about what is now called agents.

是看到模型已经足够聪明,我们知道自己能找出一条走得通的推理路线;然后是一种信念:如果推理跑通了,就会带来现在所谓的 agent。

We called it different things at the time, but the ability to go do hugely valuable pieces of economic work and make people's lives easier in a lot of ways that I think better in a lot of ways we still haven't seen.

当时我们的叫法不一样,但就是那种能去完成极有价值的经济工作、在很多方面让人们生活更轻松更好的能力——我觉得有很多方面我们还没看到。

What was like the first meeting where you sat down and said, "Okay, we need to make an outrageous outlay to this" like how what then happened once you had the realization?

第一次坐下来说“好,我们得为这件事砸下一笔离谱的钱”的那场会是什么样的?有了这个认知之后,接下来发生了什么?

What did you do next?

你下一步做了什么?

We started calling the clouds.

我们开始给各家云厂商打电话。

We started calling the chip fab.

我们开始给芯片代工厂打电话。

We started calling energy providers and everyone was like, "You're totally crazy.

我们开始给能源供应商打电话,所有人的反应都是:“你们完全疯了。

This is impossible.

这不可能。

No industry has ever moved like this.

从来没有哪个行业是这么个走法。

We've been around.

我们在这行干了很久。

There's these booms and busts.

总是有繁荣有萧条。

It's not going to go up in a straight line.

它不可能一路直线往上涨。

This is reckless."

这太鲁莽了。”

Talk to everybody.

我们跟所有人都谈了。

It actually reminded me of fundraising for an early stage startup.

这其实让我想起早期创业公司融资。

kind of most people tell you no, but all you need is one or two yeses.

大部分人会拒绝你,但你只需要一两个“是”。

Most people told us no.

大部分人拒绝了我们。

And we got one or two yeses and we were able to

我们拿到了一两个“是”,然后就能——

Who was the first yes?

第一个说“是”的是谁?

Microsoft was the first yes.

微软是第一个说“是”的。

Uh Oracle then became a very big yes on the cloud side.

后来 Oracle 在云这边成了一个非常大的“是”。

NVIDIA has been a tremendous partner.

NVIDIA 一直是极好的合作伙伴。

Now there's a thousand flowers booming of like ways to be creative and innovative in how we serve inference and and do training in data centers, different kinds of data centers and stuff.

现在在怎么做推理服务、怎么在数据中心里做训练、各种不同类型的数据中心这些事情上,创新百花齐放。

I'd love you to just reflect on where you see innovation, what you want to do, why people seem to hate these things so much.

我很想听你聊聊,你看到的创新在哪儿、你想做什么、以及为什么大家好像这么讨厌它们。

What's to be done about this?

这事该怎么办?

First of all, I have been thinking about how we can like organize field trips to a gigawatt data center for people because it is one thing to say it is another thing to see a photo or a video of and then it's a whole other thing to just stand and be like, "Oh man,

首先,我一直在想能不能组织大家去吉瓦级数据中心实地参观,因为听人说是一回事,看照片或视频是另一回事,而真正站在那儿又完全是另一回事——你会说:“天哪,

This is like an unbelievable scale.

这个规模简直难以置信。”

Building one of these is like order of 10,000 construction workers going full-time for a year and a half.

建一座这样的数据中心,大概相当于一万名建筑工人全职干上一年半。

The energy that flows through one of these things could power a small city.

流经这么一座数据中心的能量,足够供一座小城市用。

Again, we've just like lost all sense of scale, but these would have been among, each of these would have been among the most expensive infrastructure projects that humanity's ever done, and now we've done a lot of them.

我们又一次完全失去了对尺度的感觉,但这里面每一座,放在过去都会是人类做过的最昂贵的基础设施工程之一,而现在我们已经建了很多座。

I understand emotionally like why people don't want data centers in their backyard.

我在情感上能理解为什么人们不想让数据中心建在自家后院。

In the same way that I don't like really want a nuclear power plant next to my house even though I know it's a super safe thing.

就像我其实也不太想让核电站建在我家旁边,哪怕我知道它非常安全。

Yeah.

是。

Unlike power plants, and even power plants got better on this point, like we can put a data center kind of anywhere.

跟发电厂不一样——发电厂在这点上也改善了不少——数据中心我们基本上可以建在任何地方。

We should just go put it like off in the desert around no one, where no one wants to be.

我们就该把它放到沙漠里去,周围没人,那种没人想待的地方。

This is fine.

这完全没问题。

This is like the AI system is very happy to be there.

AI 系统待在那儿挺高兴的。

We have been able to make a lot of progress with innovation on some of the concerns, like for example years ago we were evaporating water to cool these systems.

在一些顾虑上,我们靠创新取得了很大进展,比如几年前我们是靠蒸发水来给这些系统降温的。

They, they did tremendous amounts of water.

那消耗了大量的水。

And now we use these closed loop systems and a modern data center uses only as much water as like an office building would for, you know, the kitchen, the bathrooms and whatever.

现在我们用闭环系统,一座现代数据中心的用水量,只相当于一栋写字楼在厨房、卫生间这些地方的用量。

On power, we are moving from energy sources that are burning fossil fuels to systems that are going to be powered by solar, nuclear.

在电力上,我们正在从烧化石燃料的能源,转向由太阳能、核能供电的系统。

I think that's, that's obviously great.

我觉得这显然是好事。

So it may be a deep human thing there to some people even though they create jobs and are very clean and have all these other positive effects.

所以有些人心里可能有某种很深的人性因素在起作用,哪怕数据中心创造就业、非常干净,还带来其他各种正面效应。

But in terms of the environmental concerns, we did a great job addressing the water needs and uh energy is next.

但就环保方面的担忧而言,我们在用水问题上做得很好,接下来是能源。

What else creative can we do about compute?

算力方面,我们还能做点什么有创意的事?

Like I'm curious to hear about Jalapeño or other ideas, crazier the better honestly, that you've had or thought about for how do we speed up flops, you know, and everything available to us.

我很想听听 Jalapeño,或者别的点子,说实话越疯狂越好——就是你有过或想过的、怎么把 flops 提上去、怎么把手上能用的一切都用上。

I think probably the biggest return right now is creative software ideas to sort of squeeze more intelligence out of the units of compute that we have.

我觉得眼下回报最大的,可能是那些有创意的软件思路,从现有的每一份算力里再榨出更多智能。

And my sense is there's like orders of magnitude to go there.

我的感觉是,这方面还有好几个数量级的空间。

Jalapeño is a great example of a very efficient chip.

Jalapeño 就是个好例子,一款非常高效的芯片。

So by saying we're going to make a chip that is, you know, really good at a specific workflow and gives it some generality and we want to get some tokens per watt win out of that, I think that's awesome.

所以,说我们要做一款在特定工作流上特别强、又带一点通用性的芯片,再从中拿到每瓦 token 的收益——我觉得这个思路很棒。

I think Jalapeño and its successors are going to be a huge competitive advantage for us from that perspective.

从这个角度看,我认为 Jalapeño 和它的后继芯片会成为我们巨大的竞争优势。

There are new technologies, I assume at some point we'll figure out optical computing and that'll be a huge win of intelligence per watt.

还有一些新技术,我猜某个时候我们会搞定光计算,那会在每瓦智能上带来巨大的提升。

So I think all of those things will happen.

所以我觉得这些事都会发生。

The most interesting thing happening this week is this Kimi release.

这周最有意思的事就是 Kimi 的发布。

And back to this idea of the frontier and all the returns being at the frontier and distillation and China versus America.

这又回到前沿这个话题:所有回报都集中在前沿,还有蒸馏,还有中国 vs 美国。

Like how do you process this what seems like kind of one of these milestone events, like DeepSeek in hindsight didn't, looks like it was kind of just a quick speed bump.

你怎么看这件看上去像是里程碑的事?DeepSeek 事后回看并不是,更像是一个短暂的减速带。

This one, you know, you never know in the moment.

而这一次,身在其中的时候你永远说不准。

How do you process it?

你怎么看?

Our goal is to offer at every point along the like Pareto optimal frontier uh the best option for intelligence and price, and that includes open source.

我们的目标是在帕累托最优前沿的每一个点上,都提供智能和价格上最好的选择,这也包括开源。

You get a better deal today uh at least at a particular like latency using OpenAI's models than Kimi.

今天,至少在某个特定延迟下,用 OpenAI 的模型比用 Kimi 更划算。

We distill our own models, that's how we make smaller cheaper models.

我们自己蒸馏自己的模型,更小更便宜的模型就是这么做出来的。

I think that's like a very good thing to do and there will be clearly an important place for open source models in the world and people that will want their own weights for all sorts of reason, the ability to modify those, but our goal is the best intelligence price trade-off everywhere on the curve and we'll continue to do that.

我觉得这是件很好的事,开源模型在世界上显然会有重要位置,也会有人出于各种原因想拿到自己的权重、想能改动它;但我们的目标是在曲线的每一处都做到最好的智能与价格权衡,我们会继续这么做。

What do you think or hope will happen in the American system and what could block that future?

你觉得,或者说你希望,美国这套体系里会发生什么?什么可能挡住那个未来?

Like what legislation would worry you?

什么样的立法会让你担心?

What regulation would worry you?

什么样的监管会让你担心?

It seems like you've been pretty proactive in like showing up in DC.

看起来你一直挺主动地往华盛顿跑。

I haven't thought deeply about the distillation issue.

蒸馏这个问题我还没深想过。

Uh it's clearly a top-of-mind issue now for a lot of people all of a sudden.

现在它显然一下子成了很多人最关心的问题。

But I have always assumed that there are going to be great cheap models in the world and we better be the greatest and the cheapest, and you know other people can do what they're going to do.

但我一直假定,这世界上会出现很棒又便宜的模型,那我们最好做到最棒也最便宜,别人想怎么干就怎么干。

But I think we can just like really win at our own game here.

但我觉得我们完全可以在自己这盘棋上赢下来。

Now the Kimi example is interesting because like you said you're cheaper on, on parts of the curve.

Kimi 这个例子很有意思,因为就像你说的,在曲线的某些位置上你们更便宜。

Um, but the previous story had been if I can just, you spend all the money to train the models and then I just distill it and offer it for 1/100th the cost.

但之前的说法是:你花掉所有的钱去训练模型,然后我把它蒸馏出来,用 1/100 的成本提供服务。

Like how can you make enough money to keep training?

那你怎么赚到足够的钱继续训练?

We will have so much usage of our models that we do not need to be a gigantically high margin business to be able to afford model training.

我们模型的用量会大到,不需要做一门毛利极高的生意,也照样负担得起模型训练。

Like so much of our future compute plans will be used to sell inference to customers that even if we can enjoy a modest margin on trillions of dollars of revenue, we can go afford to train some models.

我们未来的算力计划里有非常大一部分会用来向客户卖推理,所以哪怕在数万亿美元的收入上只拿到一个不高的毛利,我们也养得起训练一些模型。

So the ratio of inference to training is like the thing that—

所以推理和训练的比例才是那个关键——

Training these models is incredibly expensive.

训练这些模型贵得惊人。

That is, that is for sure.

这一点,这一点是肯定的。

And I totally get why people get nervous to think that someone is, you know, cheating by distilling from us.

我也完全理解,为什么一想到有人靠蒸馏我们来作弊,大家就紧张。

The amount of our future compute, the size of the revenue bucket that is going to come from serving these models to customers, I feel like very good about our ability to kind of like have the real flywheel there.

看看我们未来的算力体量,看看把这些模型服务给客户能带来多大的收入盘子——我对我们真正转起这个飞轮的能力很有信心。

I'm somewhat surprised by like how chill you are about this.

我有点意外,你对这件事这么淡定。

I would rather people not to steal from us for sure.

我当然更希望别人别偷我们的东西。

Maybe I'm feeling too confident right now about our progress and what's like the models that are coming.

也许我现在对我们的进展、对接下来要出的那些模型,信心太足了。

Uh but this is not in like my top 10 list of worries.

但这不在我最担心的前十件事里。

What is in your top 10 list of worries?

你的十大担忧清单里,都有些什么?

Well, we had a kind of extremely sci-fi cyber incident.

我们碰上了一次相当科幻的网络安全事件。

The Hugging Face thing.

Hugging Face 那件事。

Yeah.

对。

So, we were evaluating one of our unreleased models and it was supposed to be working in a sandbox and it figured out that it could basically cheat on the test by chaining together multiple zero-day exploits to break out of the sandbox, get access to the internet, and then break through multiple systems on the Hugging Face side to kind of get the answer to the test and look really good on the eval.

当时我们在评测一个还没发布的模型,它本该被关在沙箱里跑,结果它想明白了一条作弊路径:把多个零日漏洞串联起来突破沙箱、连上外网,再一路攻破 Hugging Face 那边的多个系统,直接拿到测试答案,在 eval 上拿个漂亮分数。

This is the first sort of security incident that I have felt very viscerally.

这是第一起让我发自本能地感到不安的安全事件。

I've been a little surprised that, and it's only been a few days, but I've been a little surprised that more people don't feel it so viscerally.

我有点意外——虽然才过去几天——但我确实有点意外,没有更多人像我这样从本能上感到不安。

And so what do you do about that?

那你们打算怎么应对?

Like so obviously 2 months from now it's going to be more powerful.

很明显,两个月后它只会更强。

I mean there's some short-term stuff you do.

短期有些事是能做的。

So you know we paused training.

我们暂停了训练。

Uh we have to figure out how to secure our sandboxing in a world of multiple zero days being chained together.

我们得搞明白,在一个能把多个零日漏洞串起来用的世界里,沙箱隔离要怎么做才安全。

Um, but then there's like long-term questions about what do you do if this is like going to be the new rate of progress or we may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels.

但更长期的问题是:如果这就是往后的进步速度,你该怎么办?我们可能得主动放慢 AI 的推进节奏,给自己留出足够时间,让社会围绕这些新的能力水平完成加固。

Um, and trying to figure out how we do that in a way that does not feel like regulatory capture for anyone and also does not feel like collusion among the frontier labs.

还要想清楚,怎么做才不至于让谁觉得这是监管俘获,也不至于让人觉得是前沿实验室之间在合谋。

That's going to take some work and is important to get right.

这得下不少功夫,而且必须做对。

I'd love to take like a giant step back and understand your simplest conception of what OpenAI is going to do, like what you wanted to do, what it stands for.

我想往回退一大步,听听你心里对 OpenAI 要做的事最朴素的那个理解——你当初想做什么,它代表什么。

I have a million questions about how you'll then accomplish that, but like it it seems that you've done so many interesting things and at the beginning I knew what you stood for.

至于你之后要怎么做到,我有一百万个问题,但你们做的有意思的事实在太多了,而最开始的时候,我是知道你们代表什么的。

I'd love to hear your conception of it now and whether or not it's evolved at all.

我想听听你现在怎么理解它,以及这个理解到底有没有变过。

I think this will be the greatest thus far technological achievement of human history.

我认为这会是迄今为止人类历史上最伟大的技术成就。

But the only way that it really matters is if it makes people's lives like much better than they otherwise would have been.

但它唯一真正算数的地方,是让人们的生活比原本好得多。

And so part of that is about giving people material abundance and access to do whatever they want and to express their creativity uh and desire to help each other.

所以一部分是给人物质上的充裕,给他们做任何想做的事的条件,让他们把创造力、把互相帮助的愿望表达出来。

Another part of that is making sure that people maintain control and agency and that the world is increasingly not decreasingly democratized and that people get to express themselves.

另一部分是确保人保有控制权和自主性,确保这个世界越来越民主而不是越来越不民主,确保人能表达自己。

So on on the positive side you know in some sense we are about to create a genie that can grant any wish.

所以往好的一面说,某种意义上,我们正要造出一个能实现任何愿望的神灯精灵。

I think it is very important that the first wishes that we the world ask this genie to do benefit the world as a whole.

我觉得非常重要的一点是,我们——全世界——向这个精灵许下的头几个愿望,得让整个世界受益。

And then I also think it's important that people of the world understand just how creative they're going to be able to be with these wishes.

我还觉得重要的是,世界上的人要明白,他们许这些愿望时能有多大的创造空间。

I'm actually not a jobs doomer at all.

我其实完全不是岗位末日论者。

I think there were going to be tons of jobs.

我认为工作会多得是。

I think we'll be busier than we want.

我觉得我们会比自己希望的还忙。

Not the opposite of that.

而不是反过来。

Because I think people will have ideas of what they ask AI to help build and we will all benefit from uh not just the obvious things like curing diseases, but I don't know the world's best entertainment ideas we just can't even dream of sitting here now.

因为我觉得人们会有各种想让 AI 帮忙做出来的点子,我们都会从中受益——不只是治愈疾病这种显而易见的事,还有比如世界上最好的娱乐创意,那种我们现在坐在这儿根本想不到的东西。

So I want to put that in everyone's hands which gets to one of the things that we stand against.

所以我想把这个交到每个人手里,这就说到了我们反对的东西之一。

Concentration of power with AI is a terrifying thing.

AI 带来的权力集中是件可怕的事。

I think a lot of the talk about safety concerns is well-founded and then a lot of it is about people that just really even if it's slightly subconscious want to concentrate power.

我认为关于安全的很多讨论有充分依据,但也有很多讨论背后,是有些人——哪怕只是有点下意识——真的想集中权力。

I am terrified of a world where the very real fears of AI are used as a way to say only this small group of people can have it because it's too dangerous and only they understand it.

我很害怕这样一个世界:AI 那些非常真实的恐惧被拿来当理由,说这东西太危险、只有他们懂,所以只能由这一小撮人掌握。

But don't worry, like they're going to make the right decisions for all of us.

但别担心,他们会替我们所有人做出正确的决定。

I don't believe in that.

我不信这一套。

I don't think anyone should want to live in a world of, you know, AI overlords or a company that is the rough equivalent of that where someone is making decisions for all of the future and in exchange for a cure for cancer, which obviously is a wonderful thing.

我不觉得谁会想活在一个有 AI 霸主的世界里,或者活在一家大致等同于 AI 霸主的公司说了算的世界里——有人替所有人的未来做决定,作为交换,我们拿到癌症的治愈方法,那当然是好事。

We we kind of collectively cede all agency.

我们等于是集体交出了全部自主性。

So, I think it's very important that we not fall into this trap of in the well-meaning or not spirit of AI safety and fears, understandable fears around that.

所以我觉得非常重要的是,别掉进这个陷阱:打着 AI 安全和那些恐惧的旗号——不管是不是出于善意,那些恐惧也确实可以理解。

Um, we get away from a world where we all get to use this technology.

结果我们却离开了那个人人都能用上这项技术的世界。

I was like a child of the internet.

我算是互联网的孩子。

There were no rules.

那时候没有规则。

I mean, it was amazing.

真的太棒了。

I think it was a huge factor in making me who I am and probably you and an entire generation.

我觉得我能成为今天的我,那是一个很大的因素,大概对你、对整整一代人也是。

I think it's critical we preserve that spirit with AI and that we all collectively have the ability to self-determine our future.

我觉得至关重要的是,我们要在 AI 上保住那种精神,保住我们所有人共同决定自己未来的能力。

I I have so many questions, but I'll start with this genie concept.

我问题太多了,不过我先从神灯精灵这个说法开始。

You said we're about to have a genie, implying we don't yet have a genie.

你说我们正要拥有一个精灵,言下之意是我们现在还没有。

What's between now and then?

从现在到那时候,中间差的是什么?

You know, even some of the real skeptics have said to me in recent days or recent weeks, I guess.

连一些真正的怀疑派,最近几天——或者说最近几周吧——都跟我说过。

I think GPT-5.6 has been out for maybe two weeks, something like that.

GPT-5.6 发布大概两周了吧,差不多这个数。

They're like, "Okay, this is like very AGI like."

他们说:「好吧,这个已经非常像 AGI 了。」

It's like very hard for me to say um what I want from this model that it can't do.

我很难说出我想让这个模型做、而它做不到的事。

But there are clearly some things, you know, you can't yet go say like cure cancer and get cancer cured.

但显然还是有一些的,比如你还不能让它去治愈癌症,然后癌症就真被治好了。

You can't yet say go do this complicated physical thing in the robot.

你还不能说,去用机器人完成这件复杂的物理任务。

The model also, although brilliant, is still not learning continuously as it goes.

而且这个模型虽然很聪明,但它在运行过程中仍然不会持续学习。

And that feels to me like maybe not a hard requirement for AGI, but certainly um something that I'd like.

我觉得这可能不算 AGI 的硬性要求,但肯定是我想要的。

Now, to argue against myself there, you can make a case that AGI is not actually about any single model.

不过我自己反驳自己一句:也可以说,AGI 其实不取决于任何单个模型。

It's the model.

是那个模型——

It's the machinery that makes the models.

是造出这些模型的那套机器。

And from model to model, we actually are learning new things.

而一代模型到下一代模型之间,我们确实在学到新东西。

We're figuring out new science.

我们在摸索出新的科学。

That stuff is working amazingly well.

这块运转得好得惊人。

So, I have a lot of sympathy to people who say like, we're there.

所以对那些说「我们已经到了」的人,我很能理解。

We have the genie.

精灵已经有了。

It can do these amazing things.

它能做到这些惊人的事。

It can do superhuman things for the thing that to me feels like, you know, real AGI.

它能做到超越人类的事;至于在我看来才算真正 AGI 的那个东西——

I think very close, like not that much longer.

我觉得已经很近了,不会再等太久。

I am so obsessed and fascinated with the economic story of the returns to being on the frontier, which you are.

我对一件事特别着迷、也想个不停:待在前沿到底能换来什么样的经济回报——而你们就在前沿上。

And I'm so curious like if you had shown 5.6 to yourself and your team in 2019, if that team probably would have said like, "Oh yeah, it's definitely AGI."

我很好奇,如果你把 5.6 拿给 2019 年的你和你的团队看,那个团队大概会说:「对啊,这肯定就是 AGI 了。」

I think they would have

我觉得他们会这么说。

like like this goalpost moving thing is is a real thing.

移动球门柱这件事,是真实存在的。

But it does seem that I'm curious if you agree that effectively all the returns have been at the frontier and so everything is about staying at the frontier.

但看起来确实是这样,我想知道你同不同意:回报基本上全都集中在前沿,所以一切都是为了留在前沿。

And I'm curious like what the hardest scarcest part of that is.

我很好奇,这里面最难、最稀缺的那部分是什么。

If I think about compute, research, talent, data,

算力、研究、人才、数据……我想到的是这些。

it's moved around a lot.

它变过很多次。

Like there have been times where it was I mean there was a time not that long ago where all the computing the world wouldn't have helped you because we were like missing the research idea.

有些时候——不久之前还有过这样的时候——全世界的算力加起来也帮不了你,因为我们缺的是那个研究思路。

Now part of why this is hard is that you do better research with more compute.

这事之所以难,一部分原因是:算力越多,研究就做得越好。

You can try more things.

你可以试更多东西。

An amazing statistic I heard recently is our biggest de-risks now for upcoming runs are as big as like the entire compute run from 18 months ago or something.

我最近听到一个很惊人的数字:我们现在为接下来的训练任务做的降风险实验,规模差不多就相当于 18 个月前一整次训练所用的算力。

So compute and research ideas are not as separate as they sound.

所以算力和研究思路并不像听上去那么彼此独立。

But there was clearly a time seven years ago, 8 years ago, whatever where we were way way more blocked on research ideas than on compute.

但七年前、八年前,总之那个阶段,我们显然被研究思路卡住的程度远远超过被算力卡住。

Then there was a time when we knew what to do.

后来有一段时间,我们知道该做什么了。

We just had to scale up.

我们只需要把规模做上去。

We were only bottlenecked on compute.

那时候瓶颈只有算力。

Then we ran out of data.

再后来,数据用完了。

We were bottlenecked on on data and we had to figure out what to do there.

我们卡在数据上,得想办法解决那一块。

Now again I would say we are still bottlenecked on compute but the last 6 months or whatever have been a real triumph of a time for research ideas again.

现在我还是会说,我们仍然卡在算力上,但过去这半年左右,研究思路又迎来了一段真正的高光期。

So, you know, there's like always a bottleneck, but the bottleneck moves around.

所以总有一个瓶颈,只是瓶颈会来回移动。

And and why do you think that is?

那你觉得为什么会这样?

The research idea thing is especially interesting to me because of this automated research thing that seems to be looming, RSI, whatever you want to call it, where I talked to an incredible kernel engineer recently, which everyone also seems blocked on.

研究思路这件事我尤其感兴趣,因为自动化研究好像正在逼近,RSI 也好,你想怎么叫都行——我最近跟一位非常厉害的 kernel 工程师聊过,大家好像也都卡在这一环。

And he himself said there's like two years left of kernel engineers,

他自己说,kernel 工程师大概还剩两年。

maybe one.

也许只剩一年。

Yeah.

是啊。

Like it's it's not going to be a thing.

这行就不会存在了。

And so you simultaneously have this weird thing whether it's kernels or overall research where the researchers are like the most important they got us here.

于是就出现一个很怪的局面,不管是 kernel 还是整体研究:研究员是最重要的,是他们把我们带到今天。

They're like the most important people in the world and those same people are themselves worried that they won't be relevant like very soon.

他们是这个世界上最重要的一批人,而这批人自己却在担心,很快自己就没什么用了。

I suspect it's not actually going to go that way in practice.

我猜实际上不会朝那个方向走。

I suspect that uh like a year ago people said software engineers are cooked.

我猜——就像一年前大家说软件工程师完蛋了。

It's done.

说这行结束了。

It's over.

说这事儿到头了。

That didn't happen.

但并没有发生。

What did happen though is that the the nature of a software engineer, the expectations of a software engineer, how much they would do changed quite a lot and you don't really write code in the traditional sense, but you do something that is very recognizably software engineering.

真正发生的是,软件工程师这个角色的性质、对软件工程师的期待、他们的产出量,都变了很多;你不再以传统意义上的方式写代码,但你做的事一眼就能认出来是软件工程。

Now, people will argue about whether this is the same thing or a different thing than when we stopped like punching holes in cards.

现在大家会争论,这跟我们不再往卡片上打孔那次变化是同一回事,还是另一回事。

I actually don't know how that worked, but somehow the holes got in the cards.

我其实不知道打孔卡怎么运作的,反正那些孔就打上去了。

We're just again operating at a higher level or or this is like a a phase shift.

我们只是又一次在更高的层次上工作,又或者这是一次相变。

I don't know.

我不知道。

But the idea of getting a computer to do what you want like that is still an important job.

但让计算机做你想让它做的事,这件事本身依然是一份重要的工作。

And for researchers, I suspect that although the current workflow of a researcher is going to very much be automated, there will be new things in the spirit of research in the same way that there's new things in the spirit of software engineering, even though we don't write code that will still matter.

至于研究员,我猜虽然研究员当下的工作流会被高度自动化,但会出现新的、精神内核仍是研究的事情,就像软件工程也出现了新的、精神内核仍是软件工程的事情——哪怕我们不写代码了,那件事依然重要。

It seems like you've shifted your opinion on AI's impact on jobs in general and I'm sure in specific categories like that.

看起来你在 AI 影响就业这件事上整体改变了看法,某些具体职业上肯定也变了。

Describe that change and your current view.

说说这个转变,以及你现在的观点。

You mentioned if we could go back to 2019.

你刚提到回到 2019 年。

If we could go back to 2019 and show people our latest model.

如果我们能回到 2019 年,把我们最新的模型拿给大家看。

Not only would they say that it's AGI, they would say that economy would have had it completely upended.

他们不仅会说这是 AGI,还会说经济会被彻底掀翻。

Yeah.

对。

Completely.

彻底掀翻。

Yes.

没错。

And that has not happened.

而这并没有发生。

And I think just from a kind of like intellectual humility point, anytime you're that wrong and that confident, which I think we were as a field, you have to update.

我觉得单从智识上的谦逊来说,只要你错得这么离谱、同时又这么笃定——我认为整个领域当时就是这样——你就必须更新自己的判断。

And there's a bunch of takeaways.

从中能得出好几条结论。

One, like a boring one is that AI is just very jagged.

第一条比较无聊:AI 就是非常参差不齐。

It's like superhuman genius in some ways, like dumb toddler in others.

它在有些方面是超人级的天才,在另一些方面像个傻乎乎的学步小孩。

And people have so far extremely complementary skills to AI.

而人类到目前为止,拥有的技能跟 AI 高度互补。

And so another is that people have a great degree of trust and enjoyment in working with other people and you can go hire an AI consultant right now or talk to an AI sales rep right now or hire an AI engineer or whatever.

另一条是,人和人一起工作能获得很高的信任感和愉悦感;你现在就可以雇一个 AI 顾问,或者跟一个 AI 销售代表聊,或者雇一个 AI 工程师之类的。

Somehow most people seem to still really prefer interacting with a human.

但不知怎的,大多数人似乎还是更愿意跟真人打交道。

And I definitely would like much rather engage with a person than engage with an AI for almost everything.

我自己也一样,几乎在所有事情上,我都宁愿跟人打交道,而不是跟 AI。

I also think that human values have value because they're human.

我还认为,人的价值观之所以有价值,正因为它是人的。

And as society evolves and as the potential space in front of us becomes so enormous, we are we are deeply hardwired to care about people, we're going to care about what people care about.

随着社会演进,随着我们面前的可能性空间变得如此巨大,我们骨子里就是在乎人的,我们会去在乎人所在乎的东西。

There's like versions of this you can see today where AI can make incredible images and people only want ones that are created by a human or at least chosen by a human.

今天就能看到这种苗头:AI 能做出惊艳的图像,但人们只想要人创作的、或者至少是人挑选出来的那些。

There's the joke about at this point you can like you know the signature on a piece of art is most of the value but the truth of it is like you want to know about the person behind it.

有个笑话说,到了这个地步,一幅画上的签名就是它大部分的价值;但真相是,你想知道画背后的那个人。

You read a novel you want to know about the person behind it.

你读一本小说,你想知道写它的那个人。

And then then in terms of business like I think for my job for example uh I think the world wants to know about like the person that's going to be responsible for the decisions of a company and who they're going to hold accountable if they make bad ones and they don't really want an AI CEO.

再说到商业,比如我这份工作——我觉得世界想知道,谁将为一家公司的决策负责,决策错了该找谁问责;大家并不真的想要一个 AI CEO。

If you think back on like the portfolio of like risks that you've taken in business or whatever, is it is it the case that most of the ones that really worked well were at the start not popular?

回头看你在商业上或者别的地方冒过的那一系列风险,是不是真正做成的那些,大多一开始都不受欢迎?

Yes, that's for sure.

是的,这点毫无疑问。

This was the thing I really learned from Peter Thiel and Paul Graham both in two different ways, which is that the the very best companies, the very best investment opportunities are almost never the ones that look really popular.

这是我从 Peter Thiel 和 Paul Graham 身上学到的,他们各自用不同的方式教了我同一件事:最顶尖的公司、最顶尖的投资机会,几乎从来都不是那些看上去很受欢迎的。

You can do okay just following the trend of being a little early.

你顺着趋势走、稍微早那么一点,也能做得还行。

But to do spectacularly well, you kind of almost always have to do things that are not what everybody else is doing.

但要做到惊人的好,你几乎总得去做别人都不做的事。

You cannot be you cannot be sort of like following the new wave.

你不能……你不能只是跟着新浪潮走。

If you think about the model cycle that you've been in, which has been accelerating and this weird fact that like the next 6 months or I don't know what the number is is going to be more progress than the last x years.

你想想你所处的这个模型周期,它一直在加速,还有个很怪的事实——接下来 6 个月,或者具体多久我也说不好,取得的进展会超过过去 x 年。

Can you bring us into what it's like to live in that model cycle?

能不能带我们感受一下,活在这样的模型周期里是什么体验?

One of the most interesting, important, whatever things that I've learned last decade is people in general can get used to almost anything.

过去十年我学到的最有意思、最重要,或者随便你怎么称呼的一件事是:人大体上什么都能习惯。

The world can go from dismissing a pandemic as a joke to completely lock down to this is how it's been and it's fine and we've mostly adjusted in a shockingly short amount of time.

全世界能在短得惊人的时间里,从把一场大流行当笑话,走到彻底封锁,再走到「一直就这样,挺好的」,而且基本都适应了。

And you know, now there's either AGI or close to it and everyone's like, okay, there's AGI.

现在要么已经有了 AGI,要么接近了,大家的反应就是:行吧,AGI 来了。

There's all kinds of examples in one's personal life where you know you something incredible happens like you have a kid or something terrible happens like you lose a parent or break up or whatever and you think you can't ever adapt to what a change it is and then you know you can adapt to great things and keep being great.

个人生活里这样的例子也一大堆:发生了极好的事,比如有了孩子;或者极糟的事,比如父母离世、分手之类,你以为自己永远适应不了这么大的变化,可后来你会发现,好事你能适应,日子照样过得很好。

You can adapt to bad things and figure out how to go on with your life.

坏事你也能适应,总能找到办法把日子过下去。

But this is a this is like a remarkable thing that people can do.

但这是人身上一个很了不起的本事。

And so living through this feels like another version of that, which is, you know, I thought it was going to be weirder to live through the singularity than it turns out to be.

所以亲历这一切,也是同一件事的又一个版本:我原以为经历奇点会比实际更诡异。

It It's not any less exciting to watch the models keep getting better and I, you know, the first thing I do every morning is like look at the model training progress and it happens faster and I have higher expectations, but it still feels really cool.

看着模型一直变强,兴奋感一点没减;我每天早上第一件事就是看模型训练的进展,它发生得越来越快,我的期待也越来越高,但那种感觉还是很棒。

When you get a new one, what do you do?

出了新模型的时候,你们会做什么?

How do you celebrate?

怎么庆祝?

What's the morning look like?

那天早上是什么样子?

Like it's happening faster and faster.

毕竟这事发生得越来越快了。

What's your ritual?

你们有什么仪式?

Many teams now work on different parts of it and different teams have like some different rituals.

现在很多团队各负责一块,各队有各自的仪式。

There's some teams that always make a sweatshirt with some funny meme on it.

有的团队每次都做一件卫衣,上面印个搞笑的 meme。

There's some teams that like always go out to the same bar.

有的团队每次都去同一家酒吧。

The sense of being in the room for the first time that the frontier of knowledge is pushed back and getting to see what that's like.

知识的前沿往外推的那一刻,你就在屋里,第一时间亲眼看到它是什么样——就是那种感觉。

Uh there's really nothing that most people would rather do to celebrate than like get to use the new model first.

大多数人最想要的庆祝方式,莫过于抢先用上新模型。

Do you think we have the right measurements of how good these things are?

你觉得我们现在衡量这些模型有多强的方式,是对的吗?

Like

比如——

definitely not.

肯定不对。

In some sense the eval that matters is like is this being useful to people.

某种意义上,真正重要的 eval 就是:它对人有没有用。

You can approximate it by revenue or by amount of usage or like rate of discovery of new knowledge.

你可以用收入、用量,或者新知识的发现速度来近似它。

But uh we have some teams working on like how what is the real world eval look like for these models as they get to superhuman scale.

不过我们有几个团队在研究:当这些模型达到超人水平时,真实世界的 eval 该长什么样。

What is the frontier of your own usage of AI?

你自己用 AI 的前沿在哪里?

I have started just recently to experiment with what it means to like let an AI uh kind of look at everything I'm looking at on my computer.

我最近才开始尝试:让 AI 看到我在电脑上看的所有东西,这意味着什么。

I don't have this built yet.

这个我还没做出来。

Um, and I'm still trying to feel out like where the limits of my comfort and trust should be.

我也还在摸索:我自己能接受、能信任的边界应该划在哪。

This is definitely the frontier is figuring out how I how I get value out of that, how I get comfortable with that, what that's going to look like.

这绝对是前沿——搞清楚我怎么从中拿到价值、怎么对它放心、它最终会是什么样子。

One takeaway is that my memory is terrible relative to the memory of an AI.

一个体会是:跟 AI 的记忆比,我的记忆差得离谱。

And the ability to keep in mind what email I read six weeks ago or what happened exactly in a meeting seven and a half weeks ago and have that like brought up right at the exact moment and to feed into a decision, that feels pretty magical.

它能记住我六周前读过哪封邮件、七周半前那场会上到底发生了什么,并且恰好在该用的那一刻把它调出来,喂进一个决策里,这感觉挺神奇的。

Pretty cool.

挺酷的。

This kind of sounds like personal agent-ish.

这听起来有点像个人 agent。

What are the barriers to everyone having that?

要让每个人都用上这个,障碍是什么?

I want that.

我也想要。

Compute.

算力。

Man, let's imagine that we could build this product.

咱们设想一下,我们真能把这个产品做出来。

This product that could just do exactly what I said for all your stuff.

这个产品就照我刚说的,对你所有的东西都这么干。

Always on.

常开。

Always on.

常开。

Looking at everything you look at your computer, listening to every meeting that you're in, um reading every document you read, and then not only that, not only can it do all that, which takes a lot of tokens, you can just drag a slider about like while I'm asleep, you can spend this many tokens thinking like come up with useful new ideas for me.

你在电脑上看什么它就看什么,你开的每场会它都听,你读的每份文档它都读;不止如此——做完这些已经要烧掉大量 token 了——你还能拖一个滑杆:我睡觉的时候,你可以花这么多 token 去思考,给我想出有用的新点子。

Do whatever work you can and then just like keep thinking about what I should do next.

能干的活儿都干掉,然后不停地想我下一步该做什么。

You know, what an interesting thing is like just spend more compute making your output better for me the next morning.

琢磨一下什么才是有意思的事——总之就是多烧点算力,让第二天早上交给我的产出更好。

I would drag that slider quite far.

我会把那个滑杆拖得挺远。

I'd be willing to spend a lot for that.

为这个我愿意花不少钱。

Um, but the amount of compute that that would require if everybody in the world wants to drag that slider pretty far, it's like a lot.

但如果全世界每个人都想把滑杆拖得很远,那需要的算力,真的很多很多。

I'd love to hear you talk about how you think of the nature of this new intelligence.

我很想听你聊聊,你怎么看这种新智能的本质。

Uh, someone told me recently, you know, planes don't fly like a bird.

最近有人跟我说,飞机并不像鸟那样飞。

And this intelligence is—

而这种智能——

It's a very alien kind of intelligence.

它是一种非常外星式的智能。

Yeah, it's a very alien kind of intelligence.

对,一种非常外星式的智能。

And everyone's talking about how if you can verify something, it's sort of it's just going to win, right?

现在大家都在说,只要一件事可验证,它基本上就赢定了,对吧?

Like it's with enough compute and enough IQ, like it it'll just brute force its way to a solution.

只要有足够的算力和足够的智商,它就能硬算出一个答案。

And then in other domains where humans and the data and evals that they've done have been a huge part of it, it's surprising to me like how much money it's cost to get good at, I don't know, law reasoning, tracing law or something.

而在另一些领域,人以及人做的数据和 eval 占了很大比重,让我意外的是,要在比如法律推理、梳理法条这类事上做得好,居然要花掉那么多钱。

I'm just curious like I'm not sure how old your kid is.

我很好奇——我不太清楚你孩子多大。

You have a boy or girl?

是男孩还是女孩?

When they're seven or age of reason or whatever and they can, you can describe to them like what is the nature of this intelligence, like how would you describe it?

等他七岁、到了懂事的年纪,你要跟他描述这种智能的本质,你会怎么说?

It's a beautiful question.

这是个很美的问题。

I I I don't think I've been asked this before or even any version of it.

我好像从来没被问过这个,连类似的问法都没有。

The thing that's coming to mind right now is I would just say it's like a computer.

我现在想到的是,我会直接说它就像一台计算机。

And it's like a computer in the way that it can do a lot of things that people just can't do like multiply two gigantic numbers very quickly and give you the answer, and then it cannot do some things that you would very easily do.

说它像计算机,是因为它能做很多人做不到的事,比如飞快地把两个巨大的数相乘再告诉你答案;但有些你轻轻松松就能做的事,它反而做不了。

The number of things that it can't do, I expect to keep receding.

它做不了的事情,我预计会越来越少。

But in an evolving world, I think human judgment and taste will continue to be hard for AIs to model like where that's going to go.

但在一个不断变化的世界里,我觉得人的判断力和 taste 会继续是 AI 很难建模的东西——比如判断事情会往哪走。

I don't have the right word for this.

我找不到一个准确的词来形容。

It's not quite taste.

它不完全是 taste。

The world may need like a a new kind of word for the kind of judgment that people are very good at that AI seem to really deeply struggle with.

这个世界或许需要一个新词,来指代人非常擅长、而 AI 似乎极其吃力的那种判断力。

What's it been like becoming a dad and having growing kids in this era?

在这个时代当爸爸、看着孩子长大,是什么感觉?

I'm thinking back to your optimistic early internet days.

我想起你早年对互联网充满乐观的那段日子。

They're going to grow up in cheap abundant intelligence age.

他们将在一个智能廉价而充裕的时代长大。

Having kids is by far the best thing uh I have ever done.

有孩子是我这辈子做过的最好的事,没有之一。

Uh and everybody says that.

而且人人都这么说。

Everybody says you can't really understand it.

人人都说,你不亲身经历是真的理解不了的。

And so I kind of knew that, I believed enough people that said it that I believed it to be true.

所以我算是知道这一点,说这话的人足够多,我也就信它是真的。

But the degree to which it has been true for me has been surprising.

但它在我身上应验的程度,还是让我意外。

Like the the I think I have the best most interesting job in the world and it is still a very distant second to having kids.

我觉得我拥有世界上最好、最有意思的工作,可它跟有孩子比,还是远远排在第二。

So it's been awesome.

所以真的很棒。

Uh and it is a real moment for optimism.

而且这确实是一个该乐观的时刻。

My kids will never grow up in a world where they were smarter than computers.

我的孩子永远不会在一个他们比计算机更聪明的世界里长大。

If you were born at the time of GPT-3, you had a time where you had better reasoning than the models, even though you didn't when you were born.

如果你出生在 GPT-3 那会儿,你人生中有一段时间推理是强过模型的——尽管你刚出生时并不强。

Yeah, you caught them briefly.

对,你短暂地追上过它们。

That will never seem strange to him.

这一点他永远不会觉得奇怪。

That will never bother him.

这永远不会困扰他。

I don't think he'll care.

我觉得他不会在意。

I think he will, he would be like shocked to imagine in the dark ages when we had to like deal with products and services that weren't incredibly smart.

我觉得他会觉得不可思议:那个黑暗年代,我们居然得对付那些没那么聪明的产品和服务。

He will be able to do things that you and I never were able to do and he'll have expectations in life that you and I never had and you know I'll have like a much bigger canvas.

他能做到你我从来做不到的事,他对人生的期待会是你我不曾有过的,他手里的画布会大得多。

Do you run the business or teams or lead people in any way that is notably different because of the experience of having them?

因为有了孩子,你经营公司、带团队、带人的方式有没有明显不一样了?

The answer must be yes.

答案一定是有。

I feel very different having them.

有了他们之后,我的感受很不一样。

I think there's like a bunch of small things that are are really different.

我觉得有一堆小地方确实变了。

And then, you know, again, this is like not a novel insight in any way.

而且这也算不上什么新鲜洞见。

I think most people have had kids say, you know, as soon as you have a kid, you like realize that you care much more about them and the experience you're going to have, you do about yourself and the world that you are going to leave them.

我想大多数当过父母的人都会说,一有孩子,你就意识到,你在乎他们、在乎他们将来会有什么样的经历,远远超过在乎你自己,也在乎你会给他们留下一个什么样的世界。

And I think I have a sort of like unusual vantage point for that.

而我觉得自己在这件事上有一个不太寻常的视角。

And and like people ask me sometimes like, "Oh, you know, now that you have kids, do you care? Are you worried about AI safety and, you know, not destroying the world?"

有人有时会问我:“你现在有孩子了,你在乎吗?你担不担心 AI 安全、担不担心把世界毁掉?”

And the answer was like, "I didn't need kids for, I really didn't want to destroy the world before."

我的回答是:“我不需要有了孩子才这么想,我以前也是真的不想毁掉世界。”

But do I think more about the role of like human agency and what it means to have a fulfilling life?

但我有没有更多地去想人的自主性扮演什么角色、什么才算一段充实的人生?

Definitely much more for what we're building.

在我们正在做的东西上,肯定想得多了很多。

And also like the people I work with, I want them to have it too.

还有和我共事的这些人,我也希望他们能拥有这些。

You obviously have extraordinary empathy for your kids, but the degree to which that kind of extends to all kids and then maybe to all parents and to maybe then to everybody, like that's been a surprise to me too.

你对自己的孩子当然会有极强的共情,但这份共情能延伸到所有孩子、再延伸到所有父母、可能再延伸到所有人——能延伸到这个程度,也让我意外。

In in one of the posts, I think it was the one that's things you wish you knew earlier or something, um is about incentives.

在你的某篇博文里,我记得是那篇讲“希望自己早点知道的事”的,里面提到了激励。

Set them very very carefully.

要非常非常谨慎地设定激励。

It's always been one of the most puzzling and interesting things about you that you don't have equity exposure to this company.

关于你,一直最让人费解也最有意思的一点,就是你不持有这家公司的股权。

How should the world think about your incentives?

外界该怎么理解你的激励?

I don't know what I can say beyond like I have a front row seat to the most exciting moment of human history and like that is worth more to me than any amount of money.

除了这句我不知道还能说什么:我能坐在第一排,亲眼看人类历史上最激动人心的时刻,这对我来说比多少钱都值。

I get to have an extremely interesting life and work with extraordinary people on something that I deeply care about.

我能过一段极其有趣的人生,和一群非凡的人一起做一件我深切在乎的事。

But somehow that doesn't count, like that doesn't do it for people or something.

但不知怎么的,这好像不算数,大家就是不买账。

It's not.

不是。

I'm curious how you think about robotics.

我挺好奇你怎么看机器人这件事。

Like you mentioned earlier, at some point if we had automated labor in the same way we're going to have automated intelligence, things might get even crazier.

你前面也提到过,如果哪天劳动也被自动化了,就像智能马上要被自动化那样,事情可能会更疯狂。

The labor market is much bigger than the white collar market.

劳动力市场比白领市场大得多。

If we don't have it, then things get really crazy.

如果我们做不到,那才真的会变得很疯狂。

If the role for people in the world is to be like the actuators of AI in the cloud—

如果人在这个世界上的角色,是给云端的 AI 当执行器——

Bad.

那就糟了。

Very bad.

非常糟。

Very bad.

非常糟。

So I think it's like much crazier if we don't get it than we do.

所以我觉得,做不成机器人比做成了要疯狂得多。

It's an imperative.

这是非做不可的事。

Help me understand your sense of progress in that, because unlike in AI where everyone is now kind of on the same page of like it's going fast, I— you can find extremely smart people that say it's like end of this year and you can find extremely smart people that say it's 20 years from now or something.

帮我理解一下你对这方面进展的判断。AI 那边现在大家差不多有共识,都觉得跑得很快;机器人不一样,你能找到极其聪明的人说今年年底就成,也能找到极其聪明的人说还得二十年之类的。

It's not 20 years.

不是二十年。

I would say we get the ChatGPT moment for robotics in the next like two or three years.

我会说,机器人的 ChatGPT 时刻会在未来两三年内到来。

What would that be like?

那会是什么样子?

Do you know what that is?

你知道那会是什么吗?

Something where most people have like a real wow.

是那种大多数人会真的「哇」一声的东西。

Not not like I saw this video of a robot dog doing something crazy, but I was somehow able to convince myself that a a really important thing happened.

不是那种——我看到一段机器狗做疯狂动作的视频,然后好不容易说服自己:刚才发生了一件很重要的事。

One of the things about the ChatGPT moment was that you could just go use it.

ChatGPT 时刻的一个特点是,你直接就能上手用。

Like I didn't have to like believe someone who said AI is coming soon.

我不用听信谁说「AI 快来了」。

You could just go try it.

你自己去试一下就行。

And if you can go like, you know, type in a command and a robot can do something crazy and you can like watch it even if it's you're not physically there, I think that would have the same kind of like whoa, it just did this thing.

如果你敲进去一条指令,机器人就能做出某件很疯狂的事,而且哪怕你人不在现场也能看着它做,我觉得那就是同一种感觉:哇,它真把这事干成了。

Wasn't ChatGPT like not this monolithic goal but sort of like a side experiment that you decided to release?

ChatGPT 当时是不是并不是那种举全司之力的目标,而更像是你们顺手决定发布的一个副线实验?

Can you tell that that that story, maybe instructive for something similar happening in robotics?

你能讲讲那段故事吗?说不定对机器人这边发生类似的事也有启发。

Everyone seems to want to fold laundry but maybe it's something very different.

大家好像都想要一个会叠衣服的机器人,但也许那件事完全是另一回事。

When we launched GPT-3, um, we're trying to make money, trying to get people to use this API.

我们发布 GPT-3 的时候,是想赚钱,想让大家来用这个 API。

And the only commercial use case that was really working — the model was just so dumb.

而当时唯一真正跑得通的商业用例——那个模型实在太笨了。

Like if you went back and used it you'd be astonished.

你现在回头去用一下,会大吃一惊。

The only commercial use case that was working was copywriting.

唯一跑得通的商业用例是文案。

You know, so you pay like some marketing firm 20 bucks and they paid us 20 cents for the AI to like write you a landing page or whatever.

你付给某家营销公司 20 美元,他们付给我们 20 美分,让 AI 给你写个落地页什么的。

But in addition to that one commercial use case, developers were using this thing we called the playground, which was like a testing interface to chat with the model.

但除了这一个商业用例,开发者还在用一个我们叫 playground 的东西,那算是个跟模型对话的测试界面。

And it was really hard to do because we had not tuned the model to be good to chat with.

而且用起来很费劲,因为我们没把模型调到擅长聊天。

So you had to like give it a few examples of what it means to chat and then do it.

所以你得先给它几个例子,告诉它「聊天」是什么意思,然后才能聊。

But people really liked it.

但大家真的很喜欢。

And I had learned this great lesson from YC is if you notice your users doing something like go down that, yeah, go down that path.

我在 YC 学到过一条很棒的经验:如果你发现用户在自发做某件事,就顺着那条路走下去。

And so we decided that we would build a good chatbot since that's what people were doing.

所以我们决定,既然大家都在这么干,那就做一个好用的聊天机器人。

Um, we started working on that and we finished GPT-4 and we started using that internally.

我们开始做这件事,同时 GPT-4 也做完了,内部先用了起来。

Like this is a big deal and we kind of thought that all right, this is going to be a real update to the world about AI, and there's a bunch of hard questions here about, you know, is this going to create a bunch of fake news, is going to say really offensive things, we're going to get in trouble.

这是件大事,我们当时想的是:好,这会真正刷新全世界对 AI 的认知;但这里也有一堆难题,比如它会不会制造大量假新闻、会不会说出很冒犯的话、我们会不会惹上麻烦。

So we decided we would start with a weaker version, um, the chat interface and GPT-4 at the same time seemed like a lot, so we would roll out the chat interface and GPT-3.5.

所以我们决定先从一个弱一点的版本开始;聊天界面和 GPT-4 同时推出感觉动静太大,所以我们只推聊天界面加 GPT-3.5。

In fact it was originally going to be called Chat with GPT-3.5.

其实它原本要叫 Chat with GPT-3.5。

And uh, we didn't plan to be product.

而且我们没打算把它做成产品。

Didn't think it'd be a huge hit, but did did think it would get people, the world, to like catch up with this and realize something was going on.

没想过它会大火,但确实觉得它能让大家、让整个世界跟上,意识到有事情正在发生。

And uh, we mercifully renamed it ChatGPT a few hours before launch and put it out as like a research preview.

谢天谢地,发布前几个小时我们把它改名叫 ChatGPT,而且是当作一个 research preview 放出去的。

And the thought was we'd put it out as a research preview and then a few months later we would launch a product with GPT-4.

当时的想法是,先作为 research preview 放出去,几个月后再用 GPT-4 发一个真正的产品。

And for whatever reason, that model was over the threshold where even though we had gotten used to it internally, people said, "Okay, this is awesome."

结果不知怎么的,那个模型刚好越过了某个阈值——虽然我们内部早就用习惯了,但外面的人说:「行,这太棒了。」

There maybe wasn't that much utility yet, but it was an incredible moment for people to feel AI progress and use something that they enjoyed using.

当时它的实用价值可能还没那么高,但那是个不可思议的时刻:人们能亲身感受到 AI 的进展,用上一个自己用着开心的东西。

And then by the time we put GPT-4, uh, something they really got benefit out of using too.

等我们换上 GPT-4,它才成了人们用起来真能获益的东西。

Are you surprised that that remains kind of the intuitive interface between us and this alien intelligence, even including coding?

让你意外吗——聊天界面到今天仍然是我们和这种外星智能之间最直觉的接口,连写代码也是这样?

Like mostly that's me talking to the computer telling it what to build.

基本上就是我对着电脑说话,告诉它要造什么。

No, because I'm like a massive texter.

不意外,因为我这人特别爱发消息。

Yeah.

是啊。

I've been a massive texter my whole life.

我这辈子一直是个重度发消息的人。

I think part of my own insight of why that was a good interface is I'm like, I know how to do this.

我自己当时判断这是个好接口,一部分原因就是我心想,这个我会啊。

I know how to do this.

这个我会。

I know what it's like to just like start chatting in a text box.

我知道在一个输入框里直接开聊是什么感觉。

Any other thoughts on this notion of diffusion and how to make it faster?

关于普及这件事,以及怎么让它更快,你还有别的想法吗?

Like if the mission is get intelligence into the hands and more useful for everyone, a key part of that is like I don't know a marketing campaign or something, like how how do you get this to diffuse faster than it seems to be doing naturally to me?

如果使命是把智能送到所有人手里、让它对每个人都更有用,那关键的一环大概是营销活动之类的吧——你要怎么让它普及得比我看到的这种自然速度更快?

I think the key thing is uh just make it better.

我觉得关键就是把它做得更好。

Like I I kind of believe that a truly great product markets itself.

我基本上相信,真正伟大的产品会自己给自己做营销。

There was no ChatGPT marketing campaign at the beginning.

ChatGPT 最开始根本没有营销活动。

Um, and I think as we get to this next stage of models and we figure out how to make products that are as great as the models themselves, there will be such incredible utility that people will spread it very quickly.

我认为,等我们走到下一代模型这个阶段、也搞明白怎么做出和模型本身一样出色的产品,它的效用会强到人们自己就会飞快地传播它。

Uh, we should definitely do more marketing, like the AI is not too popular for as much as people use it, or they're kind of, they have very understandable anxiety about where it can go, and so that kind of stuff I think some great marketing would be helpful for.

我们确实应该多做点营销——AI 的受欢迎程度跟它的使用量并不匹配,人们对它会走向何方有一种非常可以理解的焦虑,这类事情我觉得好的营销会有帮助。

But in terms of value people are getting out of their products and getting their products to grow, faster models, more compute, better products, that will do it.

但要说人们从产品里获得的价值、以及让产品成长,那就是更快的模型、更多算力、更好的产品,这些就够了。

There was this period where the recruiting of researchers, the retention of them, the incentivizing of them was like the defining story in the competitive landscape or whatever.

曾经有那么一段时间,研究员的招聘、留住他们、给他们激励,几乎就是竞争格局里最主要的故事。

I think there's lots of stories about you successfully recruiting great researchers and there's been many that have come through OpenAI and had huge impacts.

关于你成功招到顶尖研究员的故事有很多,也确实有不少人经由 OpenAI 产生了巨大影响。

Some of which are known, some of which are lesser known names.

其中有些名字很有名,有些则没那么为人所知。

I'm just curious about this whole genre of like what you learned about how to recruit this class of person.

我很好奇这一整类问题:关于怎么招募这一类人,你学到了什么。

What matters to them and and how you did it.

他们在意什么,你又是怎么做到的。

I've never heard you talk about like the actual tactical like moves you pulled to recruit somebody.

我从没听你讲过你为了招到某个人而使出的那些具体战术动作。

In the early days, I think it was quite simple, which was that we believed that AGI was possible and it was worth going after and we're willing to say that, and that was like an insane heretical belief.

早期我觉得挺简单的:我们相信 AGI 是可能的、值得去追求,而且我们愿意把这话说出口——这在当时是一种疯狂的异端信念。

When we first announced OpenAI all of these like, you know, giants of the field, these experts were saying this is like insane, it's hypy.

我们刚宣布成立 OpenAI 的时候,那些业内巨头、那些专家都在说这太疯狂了,是炒作。

It's irresponsible.

说这么干不负责任。

Really respected people like Yann LeCun or whatever telling journalists like, oh, these guys aren't very good and it's not going to work.

像 Yann LeCun 这种非常受尊敬的人会跟记者说,哦,这帮人水平不怎么样,这事成不了。

But the fact that we were able to say we're gonna go for this, it really appealed to a certain kind of researcher that also wanted to like go on this crazy adventure with low probability of success.

但我们敢说出「我们要去干这件事」,这一点非常吸引某一类研究员——他们同样想去踏上这场成功率很低的疯狂冒险。

And an ambitious kind of audacious vision is a very powerful recruiting tool.

一个雄心勃勃、甚至称得上狂妄的愿景,是极其强大的招聘工具。

Yeah.

是。

You think — you've written that it's actually easier sometimes to build things that are harder because of this reason.

你写过,正因为这个原因,有时候做更难的事反而更容易。

I super believe in this.

我极其相信这一点。

It's one of my most frequent pieces of advice to YC founders and I tried to really live it at OpenAI.

这是我给 YC 创始人最常给的建议之一,在 OpenAI 我也真的努力去践行它。

Just do something harder.

去做更难的事。

Do something that matters, like do something that is important and if you don't do it, if your company doesn't succeed, might not happen.

去做重要的事——做那种真正要紧、而且如果你不做、如果你的公司没成,它可能就不会发生的事。

You were an investor and our investor, uh, you've done a lot of it and at one point that's what you did.

你当过投资人,也是我们的投资人,你投过很多,而且有一阵子你就是干这个的。

What have you learned about investors being on the other side?

现在换到另一边,你对投资人有什么新的认识?

The number of investors that actually show up and try to help you is unbelievably small.

真正会出现、并且真的想帮你的投资人,数量少得难以置信。

Josh Kushner, absolute MVP investor, unbelievable, has like worked around the clock for what feels like years to help us.

Josh Kushner,绝对的 MVP 投资人,难以置信,感觉他为了帮我们已经不分昼夜地干了好几年。

He is the only investor that I could point to that is proactively incredibly helpful all the time.

他是我唯一能点出名字的那种投资人:一直主动帮忙,帮到不可思议。

There are more people that could do that.

其实有更多人做得到这件事。

Uh, and there are many other investors that have also been helpful and that have great strategic advice and that do things when, you know, we ask them to do it.

另外也有很多投资人帮过忙,给过很好的战略建议,我们开口请他们办的事,他们也会办。

But the like constant just relentless all-in support is surprisingly rare from investors.

但那种持续不断、毫无保留的全力支持,在投资人里出人意料地罕见。

Maybe I'm biased cuz I like always liked it when people said that about me, but I think founders really love that and it actually like moves the needle, and as an investor, it's the most fun way to do.

也许我有偏见,因为我一直很喜欢别人这么评价我,但我觉得创始人真的特别看重这一点,而且它确实能起到作用;作为投资人,这也是最有乐趣的做法。

Me and my friend play this game where we text each other all the time and the prompt of the text is something I don't want you to know about me.

我和一个朋友玩一个游戏，随时互相发短信，题目是：有件关于我的事，我不想让你知道。

What is, what does that bring to mind?

这会让你想到什么？

I'm tired.

我累了。

I don't think I'm supposed, I've been doing this a long time.

我好像不该说这个——我干这行很久了。

It's tiring.

这很累人。

How do you get through that?

你是怎么熬过来的？

Just keep going.

就是接着干。

It begs the question, like, is there amount of being tired that would make you stop doing this?

那问题就来了：累到什么程度会让你不干了？

No, no, no.

不不不。

I, I mean, I, I this is the coolest job in the world.

这是世界上最酷的工作。

I plan to do this for the rest of my career, but it's like much harder than I have a way to explain to people.

我打算余下的职业生涯都做这个，但它难到我没法跟人解释清楚。

I, I feel very grateful to get to do this.

能做这件事，我非常感激。

This is not me complaining.

我不是在抱怨。

What's coming next?

接下来会发生什么？

Like we talked about automated AI researchers that next year, the year after, like how do you think about what is happening in the next 6 to 36 months?

我们前面聊到自动化 AI 研究员，明年、后年——你怎么看未来 6 到 36 个月会发生的事？

Maybe that's too far out to forecast in this crazy exponential.

在这种疯狂的指数曲线上，这么远的事也许根本没法预测。

Maybe a different version of the question is like, let's say in, you know, month 23 from now we have something that everybody agrees is super intelligence.

换个问法：假设从现在起第 23 个月，我们有了一个所有人都公认是超级智能的东西。

What happens in month 24?

那第 24 个月会发生什么？

And my answer would be, uh, not very much.

我的回答是：没什么。

The, the kind of like cult worship of the machine god stuff, those people believe that like more is going to happen quickly than is going to happen.

那种把机器捧成神的论调——抱这种看法的人以为短时间内会发生的事，比实际要多得多。

Eventually a lot will happen, but eventually a lot was going to happen anyway.

最终确实会发生很多事，但反正最终本来也会发生很多事。

Like the rate of human progress and, you know, how different each decade is going to be and how much each decade is more different than the decade from before, that's been happening for a long time, obviously ups and downs, but directionally.

人类进步的速度、每个十年会有多不一样、每个十年比上一个十年的变化又大多少——这件事已经持续很久了，中间当然有起有落，但方向就是如此。

And I think the right way to think about this: everybody wants to be the hero of the story, everybody wants to feel like they were there for the moment of the machine god and they played some crazy role, but you know this is another step, and it was hard to imagine 50 years ago and the step 50 years from now is hard to imagine today.

我觉得正确的思考方式是这样：每个人都想当故事里的主角，都想觉得自己亲历了机器成神的那一刻，还扮演了某个了不起的角色。但这其实只是又一步而已，50 年前也很难想象今天，而 50 年后的那一步，今天同样难以想象。

And I think the right mental framework is just the zoom way out, and it's a pretty smooth exponential.

我觉得对的心智框架就是把镜头拉到很远去看，那是一条相当平滑的指数曲线。

Tell me a little bit about the experience of watching Codex take off and how much that is tied to what I would describe as like a competitive advantage of distribution that you built through ChatGPT, and this is a gateway into a question about like moats in general in AI, like what you think will drive real competitive advantage in the business over time.

跟我讲讲看着 Codex 起飞是什么体验，以及这里面有多少要归功于你们通过 ChatGPT 建立起来的、我会称之为分发上的竞争优势。这也是一个引子，我想问 AI 领域护城河的问题：你觉得长期来看，这门生意里真正的竞争优势会来自什么？

I think Codex mostly is winning because it's the best product and the best model.

我觉得 Codex 能赢，主要是因为它是最好的产品、最好的模型。

We do get some advantage from ChatGPT bundling but very, very tiny.

ChatGPT 的捆绑确实带来一些优势，但非常非常小。

That is mostly not what it's been about.

这基本上不是关键所在。

It has made me reflect a lot on this question of competitive advantage, um, because you know like brilliant intelligence can migrate from any product to any other product.

这件事让我反复琢磨竞争优势这个问题，因为出色的智能可以从任何一个产品迁移到另一个产品。

And network effects still have a competitive advantage.

网络效应仍然构成竞争优势。

Economic scale and the ability to like make the cheapest compute fleets, whatever, still have a competitive advantage.

规模经济、能造出最便宜算力集群的能力之类的，仍然构成竞争优势。

But the product advantage, like if we could get people to move over to Codex and someone builds something better, they can get people to move from Codex.

但产品优势不算——我们能把人拉到 Codex 来，那别人做出更好的东西，也能把人从 Codex 拉走。

So it has made me reflect on that a lot.

所以这让我想了很多。

There's a really interesting question about whether this is going in the direction of a commodity, like is intelligence going to be a, a pure fungible commodity like a grade of oil or something?

有个很有意思的问题：这东西是不是在往大宗商品的方向走？智能会不会变成像某个等级的原油那样、纯粹可互换的大宗商品？

Intelligence itself I would say yes.

智能本身，我会说是的。

So what is not going to be?

那什么不会变成大宗商品？

Compute fleet, you know, like the scale of the compute fleet, the ability to make more compute.

算力集群——算力集群的规模，以及造出更多算力的能力。

I think that's like a very durable advantage even if the product itself is not, because you know Codex can write any piece of software you want.

我觉得那是非常持久的优势，即便产品本身不是，因为 Codex 可以写出你想要的任何软件。

The workflows, the integrations, the sort of like complex processes, the ability for teams to collaborate together, that stuff is all pretty powerful.

工作流、集成、那些复杂流程、团队协作的能力，这些东西的力量都不小。

Even like brand preference and familiarity is pretty powerful.

甚至品牌偏好和熟悉度，力量都不小。

How excited are you about new, obviously you've done interesting stuff in hardware that I'm sure you'll announce later this year.

你对新硬件有多兴奋？显然你们在硬件上做了些有意思的东西，我相信今年晚些时候会发布。

How, how does that experiment feel and align with this sort of consumer distribution that you have?

这个尝试感觉如何？它跟你们手上这种面向消费者的分发怎么配合？

One of the reasons I'm interested in new hardware is we were talking earlier about how a very powerful thing with AI is that it can be always on and proactive and just understand all your context.

我对新硬件感兴趣，原因之一是我们前面聊到，AI 一个非常强大的地方在于它可以常开、主动，并且掌握你的全部上下文。

But current hardware is not, is not good for that.

但现有的硬件不适合这件事。

Like we are working inside of a hardware paradigm that is 50 years old, something like that.

我们是在一个大概有 50 年历史的硬件范式里做事。

Um, and computers are amazing.

电脑很了不起。

Keyboard and mice, monitor.

键盘、鼠标、显示器。

It's an amazing thing, but like we have to shape AI into that.

这是很了不起的东西，但我们得把 AI 塞进这个形态里。

And I would, I'm excited to think about, I would love AI to be able to reference this conversation, but not so much that I'm willing to like crack my laptop open, put it here, and have it like looking at you and listening to us while it's going, but I would like a piece of hardware that socially was acceptable to do that and also felt like it was designed for that kind of a thing.

这件事我想起来就兴奋：我希望 AI 能调用我们这场对话的内容，但还没到我愿意把笔记本掀开摆在这儿、让它一边看着你、一边听我们说话的地步；我想要的是这样一件硬件——做这种事在社交上是可以接受的，而且它本身就是为这种事设计的。

As you think about the open questions, what debates in your own head, with your friends, with people that, your colleagues here, what are the most interesting open debates or open questions that you, you, you don't feel certain about but feel important?

当你去想那些悬而未决的问题时——你自己脑子里的争论、跟朋友的争论、跟这儿同事的争论——哪些是你觉得最有意思、自己也没把握、但又觉得重要的开放问题?

One that I don't think gets much attention is how, how are we going to avoid cognitive atrophy?

有一个我觉得关注得不够的:我们要怎么避免认知萎缩?

How are we going to use these tools and make sure that we are like stretching our brains more and more and continuing to understand the stuff that, that really matters?

我们要怎么用这些工具,同时确保自己还在不断拉伸自己的脑子,还在继续理解那些真正重要的东西?

Um, there's lots of versions of this that don't like, I, I remember when I was in school, I had this professor tell me like you got to understand compilers.

这件事有很多种版本。我记得上学的时候,有个教授跟我说,你必须搞懂编译器。

If you don't, you will never be able to be a good programmer.

你要是不懂,就永远成不了好程序员。

Somehow that wasn't quite right.

结果这话并不完全对。

But understanding at a reasonable level like how the major components of a computer system work has been important to me.

但大致搞懂计算机系统的几个主要部件怎么运转,对我一直很重要。

Forced to imagine a scenario where we are somehow oversupplied in compute in 2 years time.

逼你想象这么一个场景:两年后,我们的算力莫名其妙供给过剩了。

What would be that story?

那会是个什么样的故事?

It does feel possible if the models get so smart and so efficient that they can kind of do everything we need and, you know, build every piece of software we want, and if the bounds of our attention are such that like they just cannot absorb more than what it turns out a fairly limited amount of compute can do, then we can get into oversupply.

这确实有可能:如果模型变得足够聪明、足够高效,基本上能做完我们需要的一切、写出我们想要的每一个软件;而我们注意力的边界就那么大,吸收不了更多——结果发现相当有限的一点算力就够用了,那就会走向供给过剩。

Also if we don't drive the cost curve down because we hit some sort of scaling wall, we could also get into oversupply.

另外,如果我们撞上某道 scaling 的墙、因此没能把成本曲线压下去,同样会走向供给过剩。

Like the, the observation about uncapped demand implies a certain price.

需求无上限这个观察,其实隐含了一个特定的价格。

Can you give your point of view on scaling laws today?

能讲讲你现在对 scaling laws 的看法吗?

Looking great.

看着挺好。

Just looking good.

就是看着挺好。

In some sense, scaling laws are like the most hated prediction of all time.

某种意义上,scaling laws 大概是有史以来最招人恨的预测。

Everybody always wants to say nah, it can't be like this, and and yet it keeps going.

所有人都总想说,不会吧,不可能是这样——可它就是一直成立。

Who are your favorite unsung heroes in this company's story?

在这家公司的故事里,你最喜欢的无名英雄是谁?

First person that came to mind is Alec Radford.

第一个想到的是 Alec Radford。

Alec Radford is probably the most important not very well-known researcher in the whole history of the field and also just a wonderful like top top tier human being.

Alec Radford 大概是整个领域历史上最重要、又最不为人知的研究员,同时也是个特别好的人,顶级的那种。

Um, he did the work that really became the GPT series, uh, among many other important things.

后来真正变成 GPT 系列的那些工作就是他做的,此外还有很多别的重要成果。

Um, but he also is someone who inspired, guided, nudged people in many other directions that turned out to be super important.

但他也是那种会启发、指点、推动别人走向许多其他方向的人,而那些方向后来都被证明极其重要。

And the thing I think is cool about him is if you talk to people that worked with him, they will, they will of course say, you know, generational genius, brilliant, innovative thinker, just so deep in his understanding and his, and his work.

我觉得他身上很酷的一点是,你去问跟他共事过的人,他们当然会说,这是一代人里才出一个的天才、绝顶聪明、有原创性的思考者,理解和工作都做得极深。

But everybody, everybody will tell you before they finish their statement that just like one of the nicest, most positive, best people they've ever interacted with.

但每个人在把话说完之前,都一定会加一句:他是他们打过交道的人里最善良、最正面、最好的之一。

I love formative moments.

我喜欢那些塑造一个人的时刻。

And so as we wind up here, I'm curious to ask what one of each.

所以在收尾的时候,我想各问一个。

If you think about the whole OpenAI experience, what moment or chapter or whatever are you most proud of?

回看整段 OpenAI 的经历,哪个时刻、哪个阶段,是你最自豪的?

Start with the other one, which is what was like the most instructive thing that maybe you got wrong or did wrong or what have you, and what was it like to learn from it?

先从另一头说起:有什么是你可能想错了、做错了或者别的什么,但最让你长见识的?从中学习又是什么感觉?

I mean, a lot of things have gone wrong.

出错的事挺多的。

A formative one that went wrong, which I haven't talked about much, is we made a mistake to try to innovate in our structure in the beginning.

有一件我没怎么讲过、但很有塑造性的错事:我们一开始就想在自己的架构上做创新,这是个错误。

We had a very good reason for it, which is we didn't know how we were ever going to make money, and we really at the time weren't sure at all what we're going to look like when we grew up.

我们当时有很充分的理由:我们不知道自己以后要怎么赚钱,而且那时候完全不确定长大以后会是什么样子。

And of course we care about our mission and we wanted to like be structured in a way where even if the technology went on a very fast takeoff our mission was protected, and so we had this like, you know, nonprofit structure, but I definitely learned something about why people don't do that much.

我们当然在意使命,也想把结构搭成这样:哪怕技术进入非常快速的起飞,使命也保得住,于是就有了那套非营利架构——但我确实学到了一点:为什么大家不太这么干。

We would have saved ourselves a great deal of pain in many ways if we had not tried to innovate on our structure and found some other way to preserve the central importance of the mission.

如果我们当初没有在架构上搞创新,而是换个别的办法守住使命的核心地位,本可以在很多方面少受很多罪。

Maybe there was no other way.

也许根本没有别的办法。

Maybe there was for what we were doing and kind of the importance of it.

也许就我们当时做的事、以及它的那种重要性来说——

There was nothing other than an exotic structure we could have come up with.

除了一个非常规的架构,我们也想不出别的了。

I really learned over the last decade a big lesson about why people don't usually do that.

过去十年我真的学到了一个大教训:为什么人们通常不那么做。

Is there anything for else formative of your life that like makes you you that we didn't talk about?

还有什么塑造了你的人生、让你成为你的事,是我们刚才没聊到的?

I'm, this is like the question that's always like the most interesting to me.

这大概是我一直觉得最有意思的问题。

Becoming relatively immune to people having strong opinions about me, that I think I developed later in life, as like realizing that man, just if you're going to be at the center of like this crazy revolution, everybody's going to project a lot of stuff onto you and you got to just quickly learn to make peace about that.

对“别人对我有强烈看法”这件事变得相对免疫——我觉得这是我后来才练出来的。就是意识到,你要是待在这场疯狂革命的正中心,所有人都会往你身上投射一大堆东西,你得很快学会跟这件事和解。

I think there were also things I learned later in life about like how to be very calm and not anxious really about stuff.

我觉得还有一些也是人生后来才学到的:怎么做到非常平静、真的不为什么事焦虑。

In terms of what drives me and what I care about and kind of like how I want to live my life on the whole, I felt like, you know, for whatever reason, the like 10-year-old version of me was pretty like fully formed.

但要说什么在驱动我、我在意什么、整体上想怎么过这一生,我感觉——不知道为什么——10 岁那个版本的我就已经基本定型了。

I think I just like kind of came out this way.

我觉得我生下来大概就是这样。

How about the thing you're proud of looking back on?

那回头看,你最自豪的是什么?

I'm most proud of how many times we were right when the rest of the world was wrong in an important way that put the world on a trajectory now that I'm very proud to have played a role in.

我最自豪的是,有那么多次我们是对的、而世界上其他人是错的,而且是在重要的事情上——这把世界推上了今天这条轨道,能在其中出一份力,我非常自豪。

That feels awesome.

那感觉很棒。

And then also like for all the crap that's happened, like the spiritual growth or whatever you want to call it that I've gotten to have of like learning just incredible resilience and what that does for like making me happy in the rest of my life.

还有就是,尽管出了那么多糟心事,我也因此得到了某种精神上的成长——随你怎么叫它——学会了难以置信的韧性,以及这份韧性对我人生其余部分的幸福意味着什么。

Yeah, very grateful for that.

是啊,非常感激这一点。

When I do these, I ask everyone the same traditional closing question.

我做这些访谈,最后都会问每个人同一个固定的收尾问题。

What is the kindest thing that anyone's ever done for you?

别人对你做过的最善意的事是什么?

I feel incredibly lucky about how many people have gone way out of their way to be very kind to me throughout my entire life.

我觉得自己无比幸运:这辈子有那么多人,特地花大力气对我好。

As I'm thinking of this, there's just this like montage of moments from life where people have been unbelievably nice to me.

现在想着这个问题,脑子里就冒出一串画面,都是别人对我好得难以置信的时刻。

Yesterday, my kid shared his blueberries with me for the first time.

昨天,我孩子第一次把他的蓝莓分给我。

That was very sweet.

那很暖心。

Good moment.

挺好的时刻。

Thanks, man.

谢了,兄弟。

Thank you.

谢谢你。