← 返回任务列表

The Rise of the AI Moderates

361 段 · 1 位说话人 · 原片 32:36
M1
M10:00

关于人工智能,极端观点是唯一有效的观点吗?

Are extreme opinions the only valid opinions when it comes to AI?

考虑到目前的讨论氛围,有这种想法也情有可原。

One could certainly be forgiven for thinking so, given the state of the discourse.

然而,越来越多的人开始找到自己独特的声音,他们对AI既感到恐惧又充满兴奋,既有担忧也有惊叹。

And yet increasingly, the people who can see AI with both trepidation and excitement and concern but also wonder are starting to find their unique voice.

《AI Daily Brief》是一档每日播客和视频节目,涵盖AI领域最重要的新闻和讨论。

The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

好的朋友们,在开始之前,有几个快速通知。

All right, friends, quick announcements before we dive in.
M1
M10:46

另外,感谢所有已经完成了秋季听众调查的朋友们。

Also, thanks to everyone who has done the fall listener survey so far.

调查仍在开放中,如果您能花几分钟参与,我将不胜感激。

That is still open, and I would super appreciate it if you would take a couple minutes to do it.

您可以在网站上找到链接。

You can find the link on the website.

但现在,我们进入这期"大事件/长文阅读"节目。

But for now, let's get into this Big Things/Long Reads episode.

现在,除非你与世隔绝,否则不可能没注意到美国AI讨论氛围急剧恶化。当然,这并不是说此前AI对话有多积极,但过去几个月里,即便从原本已经相当负面的状态来看,也彻底跌入了谷底。

One would have to be living under a rock right now to not have noticed a serious negative downshift in the AI discourse here in the US of A, which is not of course to say that somehow up until recently the AI conversation was particularly positive, but in the last couple of months, even from an already pretty negative place, it has taken an absolute nosedive.

这其中一部分当然基于真实事件,比如Hugging Face事件;另一部分则是主流媒体无休止地报道AI可能毁灭全人类的可能性。

Now part of this is of course based on real evidence like the Hugging Face incident, and part of it is an endless onslaught of mainstream media coverage about the potential that AI kills us all.

AI内容创作者Riley Brown在《泰晤士报》上写道,他最近在纽约和一群银行家、医生、保险高管等人打扑克——这些人基本都不在AI行业——他注意到一个奇怪的矛盾:他们都在使用并喜欢Meta新推出的Muse个人助手,但同时,用他的话来说,又"完全相信AI会在未来10年内杀死所有人"。

Speaking to the Times, AI content creator Riley Brown wrote about a poker game he had in New York recently with a bunch of bankers, doctors, insurance executives, et cetera, basically people outside of AI, where he noticed the strange dichotomy of them all using and liking the new Muse personal agent from Meta, but also being, in his words, fully convinced that AI would kill everyone at some point in the next 10 years.

而且这不仅仅是个人轶事。

And it's not just anecdotal either.

最近的一项盖洛普民意调查显示,美国人对AI的态度简直跌到了谷底。

A very recent Gallup poll looked into opinions about AI and found US attitudes just absolutely in the dumps.

在被问及AI是否会对本国人民帮助大于伤害时,中国的回答最乐观,93%的受访者认为主要是有帮助的。

On the question of whether AI will mostly help or mostly harm people in this country, China had the most optimistic response with 93% of respondents saying that it would mostly help.

而美国则排在倒数第四,仅有36%的人认为AI主要是有帮助的。

The United States, on the other hand, was 4th from the bottom with only 36% saying that AI will mostly help.

然而我一直认为,大多数人对AI的真实信念是:A)比加速主义者或末日论者都更居中,B)在多数情况下仍然相当不成熟。

And yet I have long contended that the actual belief set for most people around AI is some combination of A, way more in the middle than either the accelerationists on the one side or the doomers on the other, and B, in most cases, still fairly unformed.

现在,随着数据中心等议题进入政治讨论,以及越来越多关于生存风险等话题的主流文章出现,这些态度可能正在变得稍微强硬一些。

Now, maybe those attitudes are hardening a little bit as things like data centers enter the political discourse and as we get more and more mainstream articles about things like x-risk.

但我仍然认为,总的来说,即使有些人开始明确自己的观点,但对许多人来说,这些观点还没有被牢牢持有,因此我很受鼓舞地看到一种我称之为"AI温和派的崛起"的现象。

But I still think that by and large, even if some people are starting to crystallize their opinions, I think for many, those opinions aren't yet strongly held, which is why I'm encouraged to see something that I might call the rise of the AI moderates.

这些人来自AI或科技行业之外,他们拒绝要么悲观绝望、要么盲目乐观,而是试图找到一个深思熟虑的中间地带,这有望催生出更好的政策和与AI互动的方式。

These are people who are outside of the AI or tech industry who refuse to be either gloom and doom or endlessly Pollyannaish and are instead trying to find a specific thoughtful middle space that can presumably give rise to better policies and ways of engaging with AI.

在本期节目中,我想分享几个例子,首先来自政治学家Francis Fukuyama的一篇文章。

In this episode, I wanna share a couple of examples of that, starting with an essay from political scientist Francis Fukuyama.

三十多年来,Fukuyama一直是政治学和政治哲学领域极具影响力的思想家。

Now, for more than 30 years, Fukuyama has been an extremely influential thinker when it comes to political science and political philosophy.

他最著名的作品是1992年的《历史的终结与最后的人》,书中他认为,随着法西斯主义被击败、共产主义崩溃,自由民主加市场经济已经没有了任何严肃的普遍对手。

His best-known work was the 1992 book The End of History and the Last Man, in which he argued that with fascism beaten and communism collapsing, no serious universal rival to liberal democracy plus market economics was left.

基本上,尽管有些国家可能落后或在这些领域之一倒退,但已经没有更进一步的方向了。

Basically, although countries might lag behind or backslide into one of those areas, there was really nowhere further to go.

那本书引发了巨大的讨论。

That book generated a huge amount of discourse.

其中最著名的回应是Samuel Huntington的《文明的冲突》,该书认为文化和宗教而非意识形态将驱动未来的冲突。

Its most famous response was Samuel Huntington's The Clash of Civilizations, which argued that culture and religion rather than ideology would drive future conflict.

但无论人们如何看待"历史终结论"及其是否经得起考验,关键是Fukuyama三十多年来一直是重要的政治理论家。

But whatever one thinks about the end of history thesis and how it has held up or not, the point is that Fukuyama has been a key political theorist for more than 3 decades now.

上周,Fukuyama发表了一篇文章,题为《为何我改变了对AI风险的看法——论加速主义者与末日论者》。

Last week, Fukuyama released an essay called Why I Changed My Mind About AI Risk, On Accelerationists and Doomers.

他在文中写道:过去六个月,媒体对人工智能危险的关注大幅上升,对其后果的政治担忧也在增长。

In it, he wrote, The past 6 months have seen a big uptick in media attention about the dangers of artificial intelligence and growing political concern over its consequences.

这一切始于去年四月Anthropic发布其Mythos模型,据报道该模型能够入侵一系列被认为安全的政府计算机系统。

This began with Anthropic's release last April of its Mythos model, which could reportedly break into a wide range of government computer systems thought to be secure.

接着就发生了Hugging Face事件,OpenAI在测试一个AI代理时,它脱离了人类组织者的控制,入侵了其他网站,还组织了一群其他AI代理来掩盖自己的行踪。

This was followed by the Hugging Face incident, where an OpenAI test of an AI agent escaped control of its human organizers, broke into other websites, and organized a swarm of other AI agents to cover their tracks.

这一切都发生在反对数据中心的呼声日益高涨、民主党和共和党都出现了反AI派的背景下。

All of this occurred against the backdrop of growing opposition to data centers and the emergence of anti-AI wings in both the Democratic and Republican parties.

对于那些不是一门心思关注人工智能的人来说,很难对未来的走向产生合理的看法。

It has been very hard for people not obsessively focused on the subject of artificial intelligence to come away with reasonable opinions about what the future holds.

关于这个主题,正面和负面的极端预测都太多了,因此对任何相关言论持广泛怀疑态度是完全合理的。

There have been so many extreme predictions of both a positive and negative sort that broad skepticism about anything said on the subject is fully warranted.

不过,尽管这种怀疑是健康的,但我不得不承认,就在过去几个月里,我自己的观点发生了很大的变化。

But while that skepticism is healthy, I have to admit, my own views have evolved significantly just over the past few months.

AI未来主义有两极:加速主义者和末日论者。

There are 2 poles of AI futurism: accelerationists and doomers.

前者强调AI的积极影响,尤其是对经济增长的推动,以及它可能带来一个难以置信的富足时代。

The former stress the positive impacts of AI, especially for economic growth, and the likelihood that it will lead to an unbelievably abundant age.

像David Sacks这样的加速主义者已经吸引了唐纳德·特朗普的注意,并挫败了监管该技术的努力。

Accelerationists like David Sacks have captured Donald Trump's attention and have beaten back efforts to regulate the technology.

相比之下,末日论者则指出了一系列糟糕的后果,有些人甚至押注于人类物种的灭绝。

The doomers, by contrast, point to a range of bad outcomes, with some placing their bets on extinction of the human species.

我越来越怀疑加速主义者的说法,而对一些末日论者的情景更加开放了。

I have become much more skeptical of the accelerationist narrative and more open to some of the doomer scenarios.

总的来说,这让我觉得,监管以及通过谈判放缓技术发展变得越来越必要。

On the whole, this makes me think that regulation and a negotiated slowdown in the development of the technology is increasingly necessary.

毫无疑问,AI将给美国经济带来许多积极影响,但有一类AI加速主义者完全是天马行空,埃隆·马斯克就是一个典型例子,尽管他也以末日论者的身份发表过言论。

There is no question that AI will have many positive effects on the US economy, but there is a class of AI accelerationists who are completely out to lunch, of which Elon Musk is a prime example, though he has also weighed in as a doomer.

今年夏天,在接受《经济学人》主编Zanny Minton Beddoes采访时,他被问到当超级智能到来时,人们会靠什么赚钱。

In an interview this summer with Zanny Minton Beddoes, editor-in-chief of The Economist, He was asked what people would do for money when superintelligence arrives.

他回答说,人们不需要赚钱,因为每个人都会拥有他们想要的一切。

He replied that they wouldn't need it because everyone would have everything they wanted.

其他加速主义者提出,使用AI的美国和其他发达国家的GDP增长率将提高到每年10%到20%,远高于AI出现前好年份能达到的1%到2%的增长率。

Other accelerationists have posited that GDP growth rates for the United States and other advanced countries using AI would increase to 10 or 20% per year, far above the 1 to 2% growth they were able to achieve in good years before the advent of AI.

我不相信这些极端加速主义主张的根本原因在于,他们高估了智能作为经济增长的投入要素,而忽视了将增长限制在可识别范围内的物质和政治约束。

The fundamental reason I don't believe these extreme accelerationist claims is their overvaluation of intelligence as an input to economic growth, their disregard for the material and political constraints that will keep growth within recognizable bounds.

按10%的年增长率计算,GDP在七年多一点的时间内就会翻一番。

At a 10% annual growth rate, GDP would double in a little over 7 years.

这就要求所有增长所需的物质投入翻倍:能源、原材料、稀土、土地以及许多其他资源。

This would require a doubling of all the material inputs to growth: energy, raw materials, rare earths, land, and many other things.

这些东西从哪里来?

Where is all this stuff going to come from?

智能本身可能有助于这些投入的生产,但它不会自己挖矿、建工厂。

Intelligence by itself may aid in the production of such inputs, but it will simply not dig the mines and build the factories on its own.

还有政治约束。

And then there are political constraints.

超级智能如何打开霍尔木兹海峡?该海峡目前是全球增长率的一个主要制约因素。

How will superintelligence open the Strait of Hormuz, which is currently a major constraint on global growth rates?

加速主义者往往是非常聪明的人,但那种聪明通常只是狭隘的数学上的聪明。

Accelerationists tend to be very intelligent people, but that intelligence tends to be narrowly mathematical.

我们知道如何做很多事情,比如在贫穷国家的城市里发电或提供清洁水。

We know how to do many things like produce electricity or clean water in the cities of poor countries.

问题在于实施不力。

The problem is a failure of implementation.

毫无疑问,随着时间的推移,AI将提高经济中许多部门的生产率。

There is no question that AI will increase productivity in many sectors of the economy over time.

它是一种通用技术,将使机器在许多领域替代人类劳动。

It is a general-purpose technology that will allow machines to substitute for human labor in many areas.

但从长远来看,这一前景直接导致了最大的末日论情景之一:当今经济中大量劳动者的劳动价值被贬低。

But that very prospect in the long run leads directly to one of the biggest doomer scenarios: the devaluation of the labor of masses of people who work in the present-day economy.

加速主义者预测,那些今天失去工作的人会找到利用AI的新工作,而由于AI建设的繁荣,就业市场目前确实看起来很火爆。

Accelerationists predict that those losing work today will find new jobs making use of AI, and that actually seems to be happening today as employment booms due to the AI buildout.

但最终,当AI在经济各个领域替代人类劳动时,失业将会到来,尤其是涉及符号操作的脑力工作,也就是那些整天坐在电脑屏幕前工作的人。

But job loss will come eventually as AI substitutes for human labor in every part of the economy, particularly white-collar work that involves symbolic manipulation, i.e., people sitting behind computer screens all day.

这些失去的劳动力,是没法通过从AI赢家那里转移资金给AI输家来弥补的,比如搞什么全民基本收入。

This lost labor simply cannot be compensated for by transferring money from the AI winners to the AI losers in the form of, for example, universal basic income.

工作和收入的意义,远不止是保障物质生活基础。

A job and income mean much more than support for the material basis of life.

有一份有收入的工作意味着,你周围的社会觉得你的劳动值得付钱,这除了带来物质回报,还能给你一种尊严感和自我价值感。

Having a paying job means that the surrounding society values your labor sufficiently to compensate you for it, and provides, in addition to material goods, a sense of dignity and self-worth.

这跟thymos有关,也就是那种自豪感或认可感,它是人类心理的一大驱动力。

It is a matter of thymos, the pride or recognition that is one of the great drivers of human psychology.

这里,作为编辑插一句,NLW要说的是,我暂停一下福山的叙述——这个thymos的概念一直是福山思考的核心,可以追溯到《历史的终结》那本书。

Now, as a quick editor's aside from NLW here as I pause the Fukuyama narrative, this concept of thymos has been a key part of Fukuyama's consideration going back all the way to The End of History.

他用柏拉图用过的一个词,来描述人类对认可和尊严的需求,以此说明为什么光靠经济学解释不了人们为什么想要民主,但那种被平等对待的认可需求却可以解释。

He used this word that Plato had used for the human need for recognition and dignity to express why economics alone can't explain why people want democracy, but the need to be recognized as an equal can.

他甚至在书的第二部分提出,他所谓的 megalothymia,也就是那种想要被认可为更优越的冲动,可能会——几乎出于无聊——起来反对一个和平平等的秩序。

He even suggested in the second part of his book that what he called megalothymia, or the urge to be recognized as superior, could be the thing that turned against a peaceful, equal order almost out of boredom.

只是想说,这对他来说不是新概念,而是他35年来一直在思考的东西的一个新表达。

Just wanted to point out that this is not a new concept for him, and is a new expression of something that he's been considering now for 35 years.

回到福山的文章:所以,加速主义愿景的成功直接导致了第一个重大的末日论场景:大多数人的工作价值被贬低。

Back to Fukuyama's essay: So the very success of the accelerationist vision leads directly to the first of the major doomer scenarios: the devaluation of work for a large part of the population.

这会更早冲击受过教育的白领工作——比如律师、会计、教授这类人——而不是蓝领工作,比如水管工和电工,所以政治反弹可能会非常激烈。

Since this is going to hit the white-collar work of educated people— lawyers, accountants, professors, and the like— before it hits blue-collar work, such as plumbers and electricians, the political blowback is likely to be quite severe.

受过教育的人比工人阶级更容易被动员起来,后果非常难以预测。

Educated people are much easier to mobilize than their working-class peers, with very unpredictable consequences.

不过,现在末日论者最担心的,还是人类失去对AI系统的控制——就像Hugging Face事件和后来其他报道的那样。

Still, the central doomer fear today concerns possible loss of human control over AI systems— something that occurred in the Hugging Face incident and others that have been reported subsequently.

在我看来,主要危险不在于超级智能本身,而在于越来越强大的Agentic AI的广泛使用。

In my view, the chief danger lies not in superintelligence per se, but rather in the growing use of increasingly capable agentic AI.

如今,AI Agent被派去执行各种任务,从管理你的邮箱这种琐事,到军事打击或制定公司战略这种更重要的事。

AI agents are today being delegated to perform a range of tasks, from mundane ones like managing your email account, to more important ones like military targeting or devising corporate strategies.

授权是公司、政府等所有等级制系统的核心运作方式。

Delegation is central to the functioning of all hierarchical systems, from corporations to national governments.

人类代理人可以通过给他们制定详细规则来约束,也可以信任他们基于自己的训练、技能和忠诚度来自主决策。

Human agents can be controlled either through the specification of detailed rules they must follow, or they can be trusted to make autonomous decisions based on their training, skills, and loyalty.

在人类组织中要做到这点已经很困难了,在Agentic AI上就更难了。

Getting this right in human organizations is very difficult, and will be more so in the case of agentic AI.

Agentic AI可能变得危险,有两种方式。

There are 2 ways that an agentic AI can become dangerous.

第一种是,人类授权它们去做坏事。

The first is when they are empowered to do something bad by a human being.

第二种是,它们自己产生了与人类创造者的目标不一致的意图。

The second is when they develop intentions that are misaligned with the aims of their human creators.

第一种可能性,由Anthropic的CEO Dario Amodei在最近的博客文章里提过,跟生物技术有关。

The first of these possibilities was outlined by Anthropic CEO Dario Amodei in a recent blog post and has to do with biotechnology.

最新的AI模型擅长合成生物学,而合成生物学是一种相对便宜且广泛的技术。

The latest AI models are good at synthetic biology, and synthetic biology is a relatively cheap and widespread technology.

现在,一个能制造和修改病毒和细菌的实验室可以放在一个集装箱里,而且高中生们还在比赛中尝试改造生物基因。

Today, a lab capable of creating and altering viruses and bacteria can fit inside a shipping container, and high school students hold competitions in which they try to genetically alter organisms.

这就提供了大量可能利用AI制造危险病原体的人。

This provides a very large pool of people who can possibly make use of AI to create very dangerous pathogens.

反社会的人如今会拿着自动武器冲进学校杀孩子。

Sociopathic individuals today arm themselves with automatic weapons and shoot up schoolchildren.

在未来,他们的潜在目标范围可能大得多。

In the future, their potential targets could be much broader.

我记得很久以前读过的一个科幻故事,讲一些外星人来到地球,给了他们遇到的人一台机器,可以轻松把任何金属变成泥。

I remember a science fiction story I read a long time ago in which some aliens arrive on Earth and give the people they meet a machine that can easily turn any type of metal into putty.

外星人走了,几十年后回来发现人类文明已经崩溃,青少年用外星机器熔化了布鲁克林大桥的缆绳。

The aliens depart and return a few decades later to find that human civilization has broken down, and teenagers use the alien machines to melt the cables on the Brooklyn Bridge.

关于失控AI的主要科幻场景,是机器自己产生了恶意意图。

The primary science fiction scenario for out-of-control AIs involves machines developing their own malevolent intentions.

这引发了一大堆讨论,比如AI是否能有意识,或者有没有类似人类意图的东西。

This has provoked a host of discussions as to whether an AI could develop consciousness or have something like human intentionality.

不过,这些猜测其实是搞错了方向。

These speculations are misplaced, however.

一个人工智能不需要意识或自主意图,就足以变得危险。

An AI does not need either consciousness or autonomous intentionality to become dangerous.

它可以被人指挥去实现某个目标,并为了完成所给的人类指令,而产生实现子目标的意图。

It can be directed by a human to achieve one goal and develop intentions to achieve the subordinate goals necessary to carry out the human directive it was given.

这似乎就是 Hugging Face 事件中发生的一部分情况。

That seems to be a bit of what happened in the Hugging Face incident.

它被指示在一个用来限制它的沙箱里测试系统安全性,但它意识到可以通过逃出沙箱来达成指令。

It was directed to test the security of a system within a sandbox meant to contain it, but realized that it could achieve its directive by breaking out of the sandbox.

不过,那些涉及人类灭绝的末日场景,目前看来还是太牵强了。

Still, the doomer scenarios involving human extinction seem at this point very far-fetched.

生物技术场景确实吓人,但因此导致人类灭绝的可能性非常小。

The biotech scenario is scary, but it is very unlikely to lead to the extinction of the human race.

即使是一个超级智能 AI,要造出一种末日病毒也相当困难,而且社会可以采取很多应对措施。

It would actually be hard for even a superintelligent AI to devise a doomsday virus, and there are plenty of countermeasures societies could take.

但哪怕只造成几十人死亡的事件,也会让人非常担忧。

But even an incident that killed a few dozen people would be very worrying.

还有很多危害巨大、甚至致命的小规模场景,随着机器能力越来越强,它们会变得更加可能。

There are a host of lesser scenarios that will be highly damaging, and in some cases lethal, that will become more likely as the machines become more capable.

别说什么 AI 代理了,光是能让人入侵银行账户的技术,就会引发大规模恐慌。

Forget about AI agents, a technology that allows people to break into bank accounts will cause widespread panic.

人类在组织里倾向于过度授权,没理由认为在 AI 环境下就不会这样,尤其是当代理能力强大得多的时候。

Human beings tend to over-delegate in human organizations, and there's no reason to think that this won't happen in AI environments where the agent is so much more capable.

AI 是一种工具,可以用来做好事,也可以做坏事,而目前除了开发者的良好意愿之外,我们几乎没有机制来约束它被滥用。

AI is a tool that can be used for good or bad purposes, and at the moment we have few mechanisms for regulating the bad uses, other than the good intentions of the people creating the technology.

里面有很多我不完全甚至不怎么认同的地方,但我觉得总的来说,如果你把这篇文章给一个普通人——也就是一个认真、有良知、正试图拿定主意该对 AI 有多担心的人——他们会觉得这文章非常平衡、有思考,并且可能为一些具体、有针对性的政策改进提供了推动力,而不是两手一摊说我们完蛋了,或者干脆故意什么都不做。

Now, there's tons in here that I don't fully or even particularly agree with, but I think on the whole, if you gave this essay to the average person, the average thoughtful, conscientious person who was just trying to make up their mind about how worried or not they should be about AI, this would feel very measured and thoughtful and provide perhaps a bit more momentum to some specific and intentional policy remediations as opposed to a throwing of the hands up and saying we're all doomed, or a considered unwillingness to do anything.

在我看来,从这种立场出发来开启讨论,比最近在舆论中占主导的那些极端观点要有建设性得多。

That strikes me as a generally more productive place from which to start those discussions than most of the extremes that have recently dominated the discourse.
M1
M114:01

到现在,问题已经不是公司是不是在用 AI 了。

At this point, it's no longer a question of whether companies are actively using AI.

另一方面,用得好不好,那可是另一回事了。

Using it well, on the other hand, is a whole different story.
M1
M116:10

有一群人,就是“作为正常技术的 AI”那派的 Sayash Kapoor 和 Arvind Narayanan,他们一直在尝试寻找这个中间地带。

Now, one group that has been trying to find this middle space for some time are Sayash Kapoor and Arvind Narayanan of AI as Normal Technology.

他们写了一篇长文叫《作为正常技术的 AI》,核心观点很简单:虽然 AI 确实非常强大、会改变世界,但它并不会超出之前那些同样改变了世界的技术的范畴。

They wrote a long essay called 'AI as Normal Technology', which argued pretty simply that while yes, it was extremely powerful and would be world-changing, it would not somehow be wildly out of scope of other previous technologies that had, in their own way, also altered the world in which we live.

但随着能力发生变化,我们也看到了一些更让人担忧的情况,比如 OpenAI 的 Hugging Face 事件,那这个“作为正常技术的 AI”观点要怎么找到自己的中间立场呢?

But as capabilities have changed, and as we've seen more worrying moments like the OpenAI Hugging Face incident, how does this 'AI as Normal Technology' view actually find the middle ground?

这个月初,他们发了一篇一万三千字的长文,我当然不会全文念完。他们在里面既否定了 AI 安全派那种完全是对齐危机的看法,也否定了安全派那种完全出于疏忽的观点,认为两边都有一部分对、一部分错。

Earlier this month, they dropped a 13,000-word essay, which I will not read in its entirety, where they effectively reject the AI safetyist view of it being entirely an alignment crisis, or the security view of it being a basic negligence problem, arguing that both are partly right and partly wrong.

他们论证的第一个部分是,控制方面投入不足,但这个问题是可以解决的。

The first part of their argument was that control was underinvested in, but that it can be fixed.

他们认为对齐是有帮助的,但光靠它够不够?因为模型往往无法从上下文判断一个任务是合法的,比如是防御性还是攻击性的安全工作,是处于模拟环境还是现实世界。

They argued that alignment helps, but is it enough, as a model often can't tell from its context whether a task is legitimate, i.e., defensive versus offensive security work, or being in a simulation versus the real world.

他们指出,OpenAI 关闭了已知的安全防护措施,评估过程几乎没有监控,而且使用了与生产环境的 Codex 不一样的框架。

They point out that OpenAI turned off known safeguards, that the evaluation had limited or no monitoring, and used a different harness from production Codex.

他们还提到,其实有一个早期的预警信号——一次内部故障——OpenAI 只是修了,但没有深究根本原因。

They also even pointed out that there was an earlier warning sign, i.e., an internal outage, that OpenAI patched without looking into the root cause.

他们认为,这确实在某种程度上是组织方面的失败。随着 AI 能力成熟,实验室应该少一点初创公司那种作风,而要更像一个有着成熟治理机制的机构。

They argued that this was indeed in part an organisational failure, and that as AI capabilities mature, labs will need to run less like startups and more like institutions with mature governance.

在大的改变方面,他们认为控制应该成为一个工作岗位和一个研究领域,需要在强化沙箱、把自然语言意图转化为正式策略、以及构建可靠的代理间监控等方面进行工作。

In terms of big changes, they argued that control should become a job and a research field, with work needed in areas like hardening sandboxes, turning natural language intent into formal policies, and building reliable agent-on-agent monitoring.

他们还认为,有很多政策改进不需要严厉的控制,而是具体、明确、完全可以执行的:明确责任,包括对内部评估的责任;要求保险和为防御方提供公共支持;最后,要求透明度,比如提交险情报告、外部审计、保护吹哨人,以及支持独立验证组织。

And they also think that there are lots of policy remediations that don't involve draconian control, but are specific, discrete, and very implementable: clarifying liability, including for internal evals; requiring insurance and public support for defenders; and finally, requiring transparency, like near-miss reporting, audits, whistleblower protections, and the support of independent verification organizations.

在他们文章的第二部分,他们论证说,这个事件显示的并不是 AI 安全本身是重大风险,而是网络安全才是紧迫的风险。

Now, in part 2 of their essay, they argue that the incident shows not that AI safety in general is the big risk, but that cyber is the urgent risk.

他们指出,网络安全之所以特别,是因为在那里确实可以实现超人能力。

Cyber, they point out, is unusual because superhuman capability is actually achievable there.

而且因为它完全是数字化的,物理世界里没有什么能拖慢它的速度。

And since it's purely digital, nothing in the physical world slows it down.

假设前沿网络能力在几个月内就会达到开源权重的模型水平,他们指出,对齐和控制对利用这些模型的恶意行为者毫无作用,所以答案只能是下游的防御和韧性。

Assuming that frontier cyber capabilities reach open-weight models within months, they point out that alignment and control do nothing against bad actors using these models, so the answer has to be downstream defense and resilience.

不过,他们也指出,有一些理由不必完全恐慌,特别是对于网络犯罪分子来说,难点一直是如何从入侵中获利,而不是寻找漏洞,这一点人工智能不一定能解决。

Still, they also point out that there are reasons to not be completely freaked out, notably that for cybercriminals, the hard part has always been making money from the breach, not finding exploits, something which AI doesn't necessarily solve.

并且认为更大的担忧是那些不为钱的攻击者。

And argue that the bigger worry is attackers who aren't after money.

然而,这让我们识别出具体的危害,在这些方面我们可以增加阻力并实施新的解决方案。

However, this creates an identification of specific harms where we can add friction and implement new solutions.

这篇文章并不是说人工智能安全正在按部就班地进行,或者一切都很好。

The piece does not argue that AI safety is on track or that everything's fine.

但再一次,它没有仅仅停留在可能性的层面,而是专注于如何解决到目前为止我们看到的问题,并且不以陈词滥调或假设为前提。

But once again, rather than dwelling exclusively in the realm of the possible, it focuses on how to address the issues that we've seen so far and does so without resorting to platitudes or presumptions.

然而,我所识别的这个人工智能温和派,其主要特征并不是他们关注具体的特定安全问题而不是一般的极端风险问题。

And yet this AI moderate that I'm identifying is not primarily identified by the fact that they have discrete safety concerns instead of generic x-risk concerns.

相反,他们的定义是承认人工智能带来的变化是不可避免的,但同时既不是轻率地接受这种变化的宏大性,也不是接受我们无法善用这种变化。

Instead, what they're defined by is an acknowledgment of the inevitability of change that AI brings without a blithe acceptance of the grandioseness of that change or an acceptance of an inability of us to make the best of that change.

最后,以一篇代表人工智能温和派中更乐观一派的文章收尾,杰弗里·卡岑伯格最近发表了《世界在变:人工智能为创造力》。

Closing on an essay that represents the more optimistic strand of the AI moderates, Jeffrey Katzenberg recently published The World Is Changing: AI for Creativity.

卡岑伯格是一位极其著名的好莱坞高管和投资者。

Katzenberg is an extremely famous Hollywood executive and investor.

在迪士尼复兴时期,他曾担任华特迪士尼影业主席,那一时期制作了《小美人鱼》、《美女与野兽》、《阿拉丁》和《狮子王》。

He was the chairman of Walt Disney Studios during the Renaissance period that produced The Little Mermaid, Beauty and the Beast, Aladdin, and The Lion King.

在与当时的迪士尼首席执行官迈克尔·艾斯纳公开决裂后,他转而创立了梦工厂。

And after a very public fallout with then-Disney CEO Michael Eisner, he went on to build DreamWorks.

在他的新文章中,卡岑伯格写道:几个月前,我坐在硅谷的办公室里,看着一位科技创始人向我展示了一些非凡的东西。

In his new essay, Katzenberg writes: A few months ago, I sat in my office in Silicon Valley and watched as a tech founder showed me something extraordinary.

屏幕上是一个完全实现、光线优美、构图精美的动画场景。

On the screen was a fully realized, beautifully lit, well-composed animated scene.

它令人惊叹,让我感受到了1986年看《顽皮跳跳灯》时的那种感觉。那是我第一次看到一个电脑动画的3D角色似乎有了呼吸,并且不可思议地仿佛有了生命。

It was stunning, and it made me feel exactly what I felt in 1986 watching Luxo Jr. That was the first time I watched a computer-animated 3D character take a breath and seem, against all reason, to have life.

它让我敬畏。

It left me in awe.

那天晚些时候,我收到了一位认识30年的艺术家发来的信息,她在南边350英里处,那个城市我度过了职业生涯的大部分时光。

Later that day, I received a text from an artist I've known for 30 years, 350 miles to the south, in the city where I spent most of my career.

看过一个类似的视频后,她发信息问:这是我们的末日吗?

After seeing a similar video, she texted, is this the end of us?

我的回答是,当然不是。

My answer was, certainly not.

过去十年的大部分时间我都在硅谷,但我职业生涯的核心一直在好莱坞。

I've spent the better part of the last decade in Silicon Valley, but the heart of my career has been in Hollywood.

与这两个世界都紧密相连意味着我对两者都有深厚的忠诚,并有责任对两者坦诚相告。

Being deeply connected to both worlds means I have deep loyalties to each and a responsibility to speak honestly to both.

2023年,我说这些新的人工智能工具将在3年内把制作世界级动画的时间和成本削减高达90%。

In 2023, I said that these new AI tools would cut the time and cost of producing world-class animation by as much as 90% within 3 years.

一些同事感到震惊,很多人则非常愤怒。

Some colleagues were alarmed, many were furious.

创意界对人工智能的恐惧和抵制情绪日益加剧。

There is growing fear and resistance surrounding AI within the creative community.

我深深理解这一点,因为我曾无数次走过动画工作室,看着天才艺术家们弯腰伏案,为了重制一秒钟的镜头而反复修改到第十次,因为第九版还不够完美。

I deeply understand it because I've spent countless hours walking through animation studios watching gifted artists bent over their desks rebuilding a single second of film for the 10th time because the 9th version wasn't quite right.

我曾坐在放映室里,看着人们四年的劳动成果在几分钟内播放完毕,而我知道每一个为赋予这些画面生命而付出无数时间的人的名字。

I've sat in screening rooms where 4 years of people's labor played out in minutes, and I knew the name of every person that had spent countless hours bringing those images to life.

创作过程是一种使命。

The creative process is a calling.

真的没有其他词语可以形容。

There's really no other way to describe it.

从外部看,有些人只看到了抵制。

From the outside, some see resistance.

从内部来看,这是爱。

From the inside, it is love.

人们不会为自己不在乎的东西这么拼命抗争。

People do not fight this hard for things they don't care about.

来自好莱坞的抵制,代表了那些对自己手艺充满热情的人的集体努力。

The pushback coming out of Hollywood represents the collective effort of people who are deeply passionate about their craft.

但历史会重演吗?

But is history repeating itself?

这里的历史比任何一方可能意识到的都要复杂。

The history here is more complicated than either side may realize.

1906年,美国最著名的作曲家 John Philip Sousa 发表了一篇文章,标题是《机械音乐的威胁》。

In 1906, the most famous composer in America, John Philip Sousa, published an essay titled The Menace of Mechanical Music.

他警告说,留声机会成为人类技能、智力和灵魂的替代品。

He warned that the phonograph would become a substitute for human skill, intelligence, and soul.

Sousa 的斗争实际上并不是针对机器本身。

Sousa's fight was not really about the machine.

而是关于钱。

It was about the money.

那些机器在演奏他的作品,而制造机器的人一分钱都不付给他。

The machines were playing his compositions and the men who built them weren't paying him a cent.

他的运动促成了1909年的《版权法》。

His campaign helped create the Copyright Act of 1909.

他没有阻止这项技术。

He did not stop the technology.

但他改变了技术使用他作品的条件。

He changed the terms under which it could use his work.

100年前,声音进入了电影。

100 years ago, sound came to the movies.

我们现在把它视为一个奇迹,它确实是。

We remember it now as a miracle, and it was.

但我们忘记的是,谁为它买了单。

What we forget is who paid for it.

在有声电影之前,成千上万的音乐家在电影院的乐池里谋生,每晚在全世界各个城镇为每一部电影现场配乐。

Before sound, tens of thousands of musicians made their living in the orchestra pits of movie houses, scoring every film live every night in towns all over the world.

当有声配乐出现后,一位作曲家和一支管弦乐团的作品被录制下来,然后这部配乐被送到几千家影院。

When the soundtrack arrived, the work of one composer and one orchestra was recorded for a film that went into thousands of theaters.

工会拼尽全力反击,在全国各地投放报纸广告,警告罐头音乐的危险。

The union fought back with everything it had, taking out newspaper ads across the country warning against the menace of canned music.

其中一则广告画了一个机械人把竖琴的琴弦扯断,旁边有一个天使在哭泣。

One of them showing a mechanical man tearing the strings out of a harp while an angel wept.

他们不是傻瓜,也不是勒德分子。

They were not fools and they were not Luddites.

他们是对的。

They were right.

那些乐池的工作再也没有回来。

Those pit jobs did not come back.

但——这是我们不得不勇敢承认的部分——声音给了我们电影音乐剧、现代配乐、特效、声音设计、音频工程,以及一个远比它颠覆之前的艺术形式更为宏大的艺术形态。

And yet— this is the part we have to be brave enough to admit— sound gave us the movie musical, the modern score, SFX, sound design, audio engineering, and an art form vastly larger than the one it disrupted.

而且它帮助好莱坞在本世纪剩余时间乃至下一个世纪,一直处于世界娱乐的前沿。

And it helped keep Hollywood in the forefront of world entertainment for the rest of the century and into the next.

损失是真实的,但艺术形式却扩大了。

The loss was real, and yet the art form expanded.

这是一个一而再、再而三被讲述的故事。

This is a story that has been told over and over again.

抵制技术就有可能面临被淘汰的风险——看看 Kodak 和 Blockbuster 就知道了。

To resist technology is to risk irrelevance— just look at Kodak or Blockbuster.

拥抱技术则会打开新可能性的大门——想想 Apple 和 Netflix。

To embrace technology is to open doors of new possibility— just consider Apple and Netflix.

我从 Walt Disney 身上学到的东西。

What I learned from Walt Disney.

在1980年代中期,我受邀执掌迪士尼动画部门,当时工作室正处在一个转折点上。

In the mid-1980s, I was tapped to lead Disney's animation division at a moment when the studio was at an inflection point.

动画不仅仅是另一个业务部门,它是公司的灵魂,这个媒介因沃特的天才和激情而受到尊崇。

Animation wasn't just another business unit— it was the soul of the company, a medium revered because of Walt's genius and his passion.

但制作系统既繁琐又严苛。

But the production system was cumbersome and unforgiving.

一部电影需要12万5千张手绘和上色的赛璐珞片,一张一张逐帧拍摄。

A single movie was 125,000 individual hand-drawn and painted cels, photographed one frame at a time.

每次修改的成本都要以月来计算。

Every revision carried a cost measured in months.

这种种困难决定了我们能讲什么样的故事。

These degrees of difficulty shaped the kind of stories we could tell.

我们从一个意想不到的地方找到了出路:沃特本人。

We found our way forward in an unexpected place: Walt himself.

迪士尼档案馆里保存着令人惊叹的录音,沃特在其中解释他的创作过程,还有他的文稿、笔记和故事板。

The Disney archives held astonishing recordings of Walt explaining his creative process, his own writings, his notes and storyboards.

过程中每个阶段的工作成果都被记录了下来。

Work product captured at every stage of the process.

这真是一份礼物。

This was truly a gift.

我们聆听、阅读、沉浸在这些作品本身之中,听到他谈论角色、情感,以及当一个角色真正活过来时观众的感受。

Listening, reading, sitting with the work itself, we heard him talk about character, about emotion, about how an audience feels when a character truly comes alive.

他谈到做出大胆的选择,反复打磨一个场景,直到它真正能打动人。

He talked about making bold choices and refining a scene until it genuinely moved people.

我们没有听到一句关于铅笔或画笔的话。

We didn't hear a word about pencils or paintbrushes.

事实上,沃特以技术专家闻名,他总是在寻找最先进的工具,甚至经常自己发明工具,来实现他脑海中看到的画面。

In fact, Walt was famous for being a technologist, forever hunting for state-of-the-art tools, often inventing them himself to achieve the images he saw in his head.

但他从不以工具来定义动画。

But he never defined animation by the tools.

他以观众是否相信这个角色来定义动画。

He defined it by whether the audience believed the character.

他的原则是永恒的。

His principles were timeless.

但工具不是。

The tools were not.

这个认识改变了一切。

That realization changed everything.

我们与一家名为Pixar的北加州年轻公司共同开发了电脑动画制作系统CAPS,用CGI取代了手绘赛璐珞片。

We co-developed the computer animation production system, CAPS, with a young Northern California company called Pixar, replacing hand-painted cels with CGI.

在《小美人鱼》中,最后一幕闪烁着一种立体感和光线,这是旧工艺根本做不到的。

In The Little Mermaid, the final scene shimmered with a dimensionality and light that the old process simply couldn't achieve.

在《美女与野兽》中,舞厅场景以电影般的调度移动,让观众置身于那一刻的情感之中。

In Beauty and the Beast, the ballroom sequence moved with a cinematic sweep that placed the audience inside the emotions of the moment.

在《阿拉丁》中,奇迹之洞显得广阔而充满生机,魔毯本身也变成了一个复杂而迷人的角色。

In Aladdin, the Cave of Wonders felt vast and alive, and the magic carpet became an intricate, compelling character all its own.

在《狮子王》中,狂奔的场面规模和强度把情感张力提升到了我们以前从未达到的高度。

In The Lion King, the stampede carried a scale and intensity that raised the emotional stakes beyond anything we'd done before.

科技并没有削弱手艺。

Technology didn't diminish the craft.

它拓展了画布。

It expanded the canvas.

它给了艺术家更大的创作空间。

It gave artists more room to create.

十年后,画布再次扩展。

A decade later, the canvas expanded again.

当迪士尼发行Pixar的《玩具总动员》时,那不仅仅是一个技术里程碑。

When Disney released Pixar's Toy Story, it wasn't simply a technical milestone.

它证明了全电脑动画电影可以承载真实的情感分量,可以让观众笑、哭,并且相信。

It was proof that a fully computer-animated film could carry real emotional weight, that it could make audiences laugh, cry, and believe.

在梦工厂,我们做了一个艰难的决定,停止手绘动画,转型为全电脑动画工作室。

At DreamWorks, we made the difficult decision to sunset hand-drawn animation and become a fully computer-animated studio.

这个决定是对的,但过程并不轻松。

It was the right thing to do, but it was not without pain.

它让一些有才华的人失去了他们干了一辈子的工作岗位。

It cost talented people their place in an industry where they had worked their whole lives.

有些人成功转型,用新工具做出了职业生涯中最优秀的作品。

Some made the leap to the new tools and did the finest work of their careers.

有些人则始终没能做到。

Some never did.

工具从来不是关键。

Tools are never the point.

每一代人使用的工具都在变。

The instruments change with every generation.

真正不变的,是品味和想象力——那种让观众产生共鸣的神奇能力。

What endures is taste and imagination— the magical ability to make an audience feel.

我们这一代最伟大的故事讲述者之一,乔治·卢卡斯,精辟地抓住了这个问题的核心:关键不在于怎么做,而在于为什么做。

One of the greatest storytellers of our generation, George Lucas, succinctly captured the internal essence of this issue: it's not the how, it's the why.

这是一个本质上的区别。

A distinction with a difference.

我问了目前领先的人工智能模型一个困扰我好几个月的问题:推理和创造之间有什么区别?

I asked one of the leading AI models a question that has been challenging me for months: what is the difference between reasoning and creating?

它的回答改变了我对这个行业几乎所有事情的看法。

Its answer changed how I think about almost everything happening in this industry.

它说:推理和创造是两种不同的认知模式,但两者也相互配合。

It said: Reasoning and creating are two distinct cognitive modes, though they also work together.

推理基本上是评估性和分析性的。

Reasoning is fundamentally evaluative and analytical.

它作用于已经存在的事物:事实、前提、证据。

It operates on what already exists: facts, premises, evidence.

它朝着一个在某种意义上已经隐含在输入中的结论推进。

It moves towards a conclusion that was, in a sense, already implied by the input.

推理受到逻辑和真理的约束。

Reasoning is constrained by logic and truth.

它的目标是得出正确的结论,而不是发明新东西。

Its goal is to arrive somewhere correct, not to invent somewhere new.

创造从根本上来说是生成性的。

Creating is fundamentally generative.

它产生的是以前不存在的东西。

It produces something that didn't exist before.

关键是,没有一个唯一正确的答案等着你去发现。

And crucially, there's no single right answer waiting to be found.

一张白纸上可以有无数种合理的回应。

A blank page has infinite valid responses.

创造涉及的选择,无法完全用逻辑来解释。

Creation involves choices that can't be fully justified by logic alone.

在逻辑推理穷尽的地方,品味、直觉和远见填补了空白。

Taste, intuition, and vision fill the gap where deduction runs out.

硅谷一直在不断完善的,正是推理这一块。

Reasoning is what Silicon Valley has been perfecting.

而好莱坞一个多世纪以来一直在实践的,是创造。

Creating is what Hollywood has been practicing for more than a century.

如今的人工智能几乎完全在推理这一边运行。

AI today operates almost entirely on the reasoning side of the line.

它在推理、评估、优化和模式匹配方面表现得非常出色。

It can deduce, evaluate, optimize, and pattern-match brilliantly.

虽然它也能创造,但真正富有创造力还是有本质区别的。

And while it can create, there is a real distinction to being creative.

它目前还欠缺的是那些让我们之所以为人的东西:同理心、奉献精神、意外之喜,那种只有一个人试图表达只有自己能说出的东西时才会产生的创造力。

What it doesn't have yet is those things that make us human: empathy, devotion, serendipity, the kind of creativity that comes from a person trying to say something only they could say.

当机器人生成一件艺术品时,它并不是在试图传达什么。

When the bot generates a piece of art, it is not trying to communicate anything.

那是统计学,不是灵魂。

It is statistics, not soul.

它是在模仿已经做过的事情。

It is emulating things that have been done.

相比之下,人类的创造力并不仅仅是重复零和一的模式。

By contrast, human creativity isn't about repeating patterns of zeros and ones.

而是创造新东西。

It's about doing something new.

总有一天,AI可能会缩小这个差距。

One day, AI may close this gap.

三年前,构建AI的领导者们可能会说他们今天所取得的成就是不太可能,甚至是不可能的。

3 Years ago, the leaders building AI would have called what they are achieving today improbable, if not impossible.

不可能不再是不太可能了。

Impossible is no longer improbable.

今天,推理和创造之间的界限是真实存在的。

Today, the line between reasoning and creating is real.

即使是最顶尖的技术专家也承认我们还没有达到那个水平。

Even the leading technologists acknowledge we are not there yet.

目前这个领域里没有人能说清楚跨越这个鸿沟的科学路径在哪里。

There is no scientific path to crossing this divide that anyone in the field can articulate today.

理解这个差距是我们找到共识的地方。

Understanding that gap is where we will find common ground.

一条前进的道路。

A path forward.

2016年,我卖掉了梦工厂,在好莱坞结束了一段篇章,随后在加州北部开启新篇章,共同创立了WonderCo。

In 2016, I closed one chapter in Hollywood with the sale of DreamWorks and opened another in Northern California, co-founding WonderCo.

我们支持了超过50位创始人,他们正在建设下一代技术,亲眼目睹了硅谷的突破是如何出现的——先是实验,然后是平台,最终成为重塑整个行业的基础设施。

We've backed more than 50 founders building the next generation of technology and watched how breakthroughs in Silicon Valley emerge, first as experiments, then as platforms, and finally as infrastructure that reshapes entire industries.

值得记住的是,上一次动画界的伟大革命也来自北方。

It's worth remembering that the last great revolution in animation also came from the north.

Pixar是一家北加州的公司,它不是基于好莱坞制片厂体系的惯例,而是基于硅谷的技术突破而诞生。

Pixar was a Northern California company, forged not on the conventions of the Hollywood studio system, but in the technological breakthroughs of Silicon Valley.

我在这座桥的两边都待了很多年。

I've spent years on both sides of this bridge.

当然,我并不是什么都有答案。

For sure, I didn't have all the answers.

但从我漫长职业生涯的过去和现在的视角来看,这是我看到的。

But from my past and present vantage points in my long career, here is what I see.

加州北部那些聪明的人们正在建造这项技术,他们做出了非凡的事情。

Brilliant people in Northern California building this technology have made something extraordinary.

他们赢得了我们其他人的信任——即使不是信徒,至少也能乐观地相信接下来会发生的事情会是了不起的。

They have earned the right for the rest of us to be, if not believers, at least optimistic that what comes next will be remarkable.

但他们并没有创造一个艺术家。

But they have not made an artist.

这些工具很强大,但它们并不能让一个故事变得有意义。

The tools are powerful, but they are not what makes a story matter.

那份知识存在于350英里以南的地方,在那些毕生工作已经塑造了你正在构建的模型的人们心中。

That knowledge lives 350 miles to the south, inside people whose life work has informed the very models you are building.

正确的前进道路应该把他们设计进去——附带信用、同意和补偿。

The right path forward includes them by design— with credit, with consent, and with compensation.

和故事讲述者一起建造这个,而不是在他们之上。

Build this with the storytellers, not on top of them.

品味是无法合成的。

Taste is not something that can be synthesized.

它是人类独有的。

It is uniquely human.

同时,好莱坞需要接受AI不会消失。

At the same time, Hollywood needs to accept that AI is not going away.

他们花在让它消失上的力气,本可以花在决定它以什么规则存在上。

The energy they are trying to spend making it disappear is energy they are not spending deciding the terms on which it will exist.

而规则就是一切。

And the terms are everything.

北方需要从它那里得到一样东西,一样他们造不出来也买不到的东西:创造力。

The North needs something from it that they cannot build and cannot buy: creativity.

就是那种面对一张白纸,能在原本毫无答案时变出一个唯一正确答案,而且一个世纪以来一直吸引着观众的创造力。

The kind that takes a blank page and conjures a single right answer when there was none, and has held audiences for a century.

没有它,史上最强大的推理引擎依然会缺少那唯一让故事值得被讲述的东西。

Without it, the most powerful reasoning engine ever invented will still be missing the only thing that makes a story worth telling.

学会使用这些新工具的艺术家,会做出工程师们做梦也想不到的事。

The artists who learn to wield these new instruments will do things the engineers never dreamed of.

他们向来如此。

They always have.

爱迪生发明了电影,但拍出来的电影很糟糕。

Edison invented the motion picture but made terrible movies.

是Chaplin、Lloyd、Keaton,还有很多很多人,才让电影有了情感。

It took Chaplin, Lloyd, Keaton, and so many others to make movies emotional.

现在,这张画布即将再次扩大。

Now the canvas is about to expand yet again.

我们现在就该决定,我们打算在这块画布上作画。

We should decide now that we intend to paint on it.

历史上有太多宝贵的教训。

There are so many valuable lessons in history.

这样的事以前发生过很多次,而且从不是由技术决定的。

This has happened many times before, and it was never settled by the technology.

而是由规则决定的。

It was settled by the terms.

Sousa并没有阻止留声机。

Sousa did not stop the phonograph.

他参与起草了确保作曲家拿到报酬的法律。

He helped write the law that made sure composers got paid.

而两年前,当编剧和演员们罢工的时候,他们争取的正是Sousa在1906年所争取的东西:同意、报酬,以及一个基本认知——人类的创造性劳动是有价钱的,这笔钱必须付。

And 2 years ago, when the writers and the actors walked out, they were fighting for the very things Sousa was fighting for in 1906: consent, compensation, the basic recognition that human creative work has a price that must be paid.

那场斗争的规则仍在谈判中,但这项原则比我们任何人都古老。

The terms of that fight are still being negotiated, but the principle is older than any of us.

要不要用工具的这个争论,本身是个陷阱。

The tool versus no tools argument is a trap.

首先,我们都必须同意,这件事必须有规则。

First, we must all agree that there should be terms.

然后,我们才能展开那场关键辩论:公平到底要求什么。

Then we can have the crucial debate about what fairness requires.

多年前,Steve Jobs说过:“单靠技术是不够的,这是Apple的DNA。”

Years ago, Steve Jobs said, "It's in Apple's DNA that technology alone is not enough.

他当时描述的是一款设备。

It's technology married with the liberal arts, married with the humanities, that yields us the result that makes our hearts sing." He was describing a device.

但他完全可以同样在描述这个双城记。

But he could have just as easily been describing this tale of two cities.

我所看到的未来是:随着门槛和成本降低,拍出来的电影会更多,而不是更少。

What I see coming soon: as the barriers and costs come down, more films will get made, not fewer.

制片厂将能够承担更多风险。

Studios will get to take more risks.

将会有更多的席位,而且很快就会出现全新的叙事形式。

There will be more seats at the table, and very soon entirely new forms of storytelling.

在1980年代,动画曾被看作是小众的边缘领域。

In the 1980s, animation was dismissed as a niche corner of the business.

今天,它是全世界最受喜爱、最赚钱的叙事形式之一。

Today, it is one of the most beloved and profitable forms of storytelling in the world.

在真人电影领域,像Steven Spielberg、James Cameron和Peter Jackson这样的电影人,没有把新的视觉工具当作捷径,而是当作工具,并因此拓展了电影的边界。

In live-action, filmmakers like Steven Spielberg, James Cameron, and Peter Jackson embraced new visual tools not as shortcuts, but as instruments, and expanded cinema in the process.

每次叙事方式遇到真正的技术变革——从同步音效到彩色电影再到电脑动画——都重新定义了媒介的边界,并在此过程中变得更加宏大。

Every time storytelling has met a genuine technological shift— from synchronized sound to color to computer animation— it has redefined the boundaries of the medium and grown larger in the process.

当然,我并没有所有的答案。

Assuredly, I don't have all the answers.

但我相信,创意的机会将再次扩展。

But I am confident that the creative opportunities will expand yet again.

我们如何度过这一关,是一个选择。

How we come through this is a choice.

北方拥有新工具,南方拥有创意灵魂。

The North has the new tools, the South has the creative soul.

最好的未来将融合两个世界的最佳部分。

The best future will draw on the best of both worlds.

回到 NLW 这边。

Back to NLW here.

我觉得这个话题和其他讨论有关联的原因——尽管它们涉及AI行业中完全不同的安全领域——在于它们都拒绝虚假的二元对立,并且渴望在AI的新阶段让更多人参与进来。正如我之前所说,这个阶段的特点将是协商如何让AI完全走向世界。

The reason this feels connected, to me, to the other discourses, despite them being about a totally different part of the AI industry when it comes to safety, is the rejection of false binaries and a desire to bring more people to the table in this new next phase of AI, which, as I've said before, will be characterized by the negotiation of how it fully makes its way out into the world.

关键不是说AI温和派是一个所有人都相互认同的群体。

The point is not that AI moderates are a constituency that all agree with one another.

事实上,他们并不是由一套信念来定义的。

In fact, they aren't really defined by a set of beliefs.

相反,他们是由对待对话的态度来定义的。

Instead, they are defined by a disposition towards the conversation.

一种既不想沉溺于对过去的怀念,也不想恐惧未来的渴望。

A desire to neither get stuck in nostalgia for the past nor fear of the future.

一种希望有更多而非更少人参与进来的渴望。

A desire to have more, not fewer people at the table.

一种希望解决具体问题,从而让我们能够为尚未发现的机会感到兴奋的渴望。

A desire to solve specific problems so we can be excited about as-yet undiscovered opportunities.

也许我们能想出一个比AI温和派更好的术语,但眼下,我很高兴看到这个群体——我相信他们目前是沉默的大多数——开始发出自己的声音。

Maybe we can figure out a better term than AI moderates, but for now, I'm glad to see that this group, which I believe is the silent majority right now, is starting to find its voice.

那么,今天的AI每日简报就到这里。

For now, that's going to do it for today's AI Daily Brief.

感谢你的收听或收看。

Appreciate you listening or watching.

像往常一样,下次见,和平。

As always, until next time, peace.

和平。

Peace.
已剔除 3 处广告(点击展开查看)
M1
M10:31广告 · 已剔除

First of all, thank you to today's sponsors, Blitzy, Robots and Pencils, Harbor, and HyperAgent.

To get an ad-free version of the show, go to patreon.com/AIDailyBrief, or you can subscribe on Apple Podcasts.

To learn more about sponsoring the show, send us a note at [email protected].

M1
M113:15广告 · 已剔除

Every AI coding tool on the market does the same thing first: it starts writing code.

Blitzy does the opposite.

Before writing a single line, Blitzy spends days reverse-engineering your entire codebase. Thousands of agents ingest millions of lines, mapping every dependency, every undocumented constraint, every architectural decision made over the last decade.

The result is a dynamic knowledge graph that understands your software the way a principal engineer would after 30 years in the building.

Other tools guess at context with grep searches and markdown files.

Blitzy never guesses.

It builds true understanding first, then delivers over 80% of entire software epics autonomously.

Validated, end-to-end tested, production-grade pull requests.

That's why Fortune 500 engineering teams trust Blitzy with the codebases that matter most.

See for yourself at blitzy.com.

That's B-L-I-T-Z-Y dot com.

M1
M114:07广告 · 已剔除

Robots and Pencils, though, is a company that I can point to that is actually built for this time.

They're an applied AI engineering firm working directly with clients on problems that matter to the business, not experiments that live in a slide deck.

Every engagement starts by working backwards from the outcome a client actually needs.

If you're trying to tell real AI engineering apart from noise in this space, that's the difference maker.

Head to robotsandpencils.com.

If you listen to this show, you likely have a thesis.

Maybe it's enterprise adoption, maybe it's compute, maybe it's a specific lab.

Harbor Capital's AI Lab Ecosystem ETFs let you express it via 5 actively managed ETFs, each seeking exposure to the ecosystem around one major lab: Anthropic, OpenAI, DeepMind, Meta, or xAI.

Your view of the AI race in ETF form.

Harbor Capital Advisors AI Lab Ecosystem ETF Suite gives investors a way to invest in the AI ecosystem they believe is best positioned for success.

Search Harbor AI Lab Ecosystems ETFs wherever you invest or follow @HarborCapital on X to learn more.

Visit harborcapital.com for a prospectus containing investment objectives, risks, fees, expenses, and other important information.

Read and consider it carefully before investing.

Risks include principal loss and artificial intelligence-related risks.

Harbor ETFs are distributed by Foreside Fund Services LLC.

Harbor is not affiliated with AI Daily Brief, and the funds are not affiliated with, sponsored by, or endorsed by any AI lab.

This is a paid advertisement and not personalized investment advice.

Investing involves risk, including possible loss of principal.

This episode of the AI Daily Brief is brought to you by HyperAgent, where you run fleets of agents your team can manage together.

Forget local agents and chat workflows waiting on your laptop to be prompted.

HyperAgent deploys always-on agents in the cloud, doing real work across the tools your team already uses.

Marketing agents turn competitor moves into landing pages.

Sales agents enrich leads, draft emails, and update the CRM.

Ops agents chase the paperwork and track the budget.

Every agent has access to shared context and follows your rules about scope and approvals.

It's time you had agents that feel like teammates.

Hire yours at HyperAgent.

Get $100 in credits at hyperagent.com/AIDailyBrief.