人工智能正在吞噬劳动力市场吗？+ 五角大楼、OpenClaw与Alpha School的最新动态

凯西，最近怎么样？

Casey, how's it going?

你上周末有个大新闻啊，朋友。

You had some big news over the weekend, my friend.

是的。

I did.

是的。

I did.

我们得更新披露信息了。

We're gonna have to update the disclosure.

对。

Yes.

为什么？

Why is that?

嗯，过去一年左右在节目中，我一直披露我男朋友在Anthropic工作，但以后不会再提了，因为我没有男朋友。

Well, for the past year or so on the show, I've been disclosing that my boyfriend works at Anthropic, but we're not gonna say that anymore because I don't have a boyfriend.

我订婚了。

I have a fiance.

嘿。

Hey.

所以是这样的。

So yeah.

是的。

Yeah.

这太令人兴奋了。

That's so exciting.

他们说，结婚是和一个男人建立的最严肃的关系之一，仅次于和他一起做播客。

They say that getting married is the second most serious kind of relationship you can get into with a man besides starting a podcast with him.

所以咱们走着瞧吧，但我非常乐观。

So we'll see how it goes, but I'm very optimistic.

你们已经决定婚礼的主题了吗？

Have you decided on a theme for your wedding yet?

你知道吗，我得承认，我们还处于规划的最初阶段。

You know, I I have to admit, we're we're at the very earliest stages of the planning.

所以如果你有什么想法，我很乐意听取。

And so if you have any, you know, ideas, I'm very open to that.

嗯，我确实开始 brainstorm 一些可能的婚礼标签了，你知道的，每对情侣都需要一个这样的标签。

Well, I did start brainstorming possible wedding hashtags, you know, because you every every couple needs one of those.

当然。

Absolutely.

那这些怎么样？

So how about these?

好的。

Okay.

A g I do。

A g I do.

还有其他的吗？

Any others?

答应接受媒体采访。

Say yes to the press.

我喜欢这个。

That that that one I like.

我喜欢这个。

That one I like.

不错。

That's good.

或者，当然，经典的版本：我丈夫在Anthropics工作。

Or, of course, the classic, my husband works at Anthropics.

我是凯文·罗斯，《纽约时报》的科技专栏作家。

I'm Kevin Roose, a tech columnist at the New York Times.

我是来自Platformer的凯西·牛顿。

I'm Casey Newton from platformer.

这有点难。

And this is hard for.

本周，另一篇走红的AI文章引发了股市震荡。

This week, another viral AI essay shakes up the stock market.

到底发生了什么？

What's really going on?

经济学家安东·科里内克将为我们解释一切。

Economist Anton Korinek is here to explain it all.

此外，还有Anthropic与五角大楼的较量，以及我们的系统更新内容。

Plus, Anthropic versus the Pentagon and more in our system update.

我们之后必须重启电脑吗？

Do we have to restart our computer after that?

是的。

Yes.

好的。

Okay.

好吧，凯文，又一周过去了，又一篇预测AI将带来灾难的爆款文章搅动了股市。

Well, Kevin, another week, another viral essay predicting AI caused doom, roiling the stock markets.

到底发生了什么？

What is going on?

是的。

Yes.

这周的大新闻是，一家名为Citrini Research的公司撰写的一篇论文走红了。

So the big news from this week was that an essay written by a company called Citrini Research went viral this week.

这篇论文名为《2028年全球智能危机》，大致描绘了一个近未来图景：人工智能产业不仅吞噬了劳动力市场，还颠覆了许多领先企业的商业模式。

The essay is called the twenty twenty eight global intelligence crisis, and it basically sketched out a near future in which the AI industry eats not only the labor market, but also the business models of a number of leading companies.

里面举了很多例子。

There were lots of examples.

这是一篇非常长的论文。

It's a very long essay.

但基本上，这是这家机构试图说明：如果人工智能持续进步，未来几年可能会是什么样子。

But basically this was one firm's attempt to say, here's what the next few years could look like if AI progress continues.

而这家机构所说的未来图景，对凯文来说相当糟糕。

And what this firm says it will look like is pretty bad, Kevin.

对吧？

Right?

这里的观点是，人工智能代理会不断改进并接管经济。

The suggestion here is that AI agents improve and take over the economy.

因此，你会看到大规模的失业、股市大幅萎缩，以及这篇文章点名的许多公司，比如DoorDash，情况会特别糟糕。

And so as a result, you're gonna see massive job losses, like a huge contraction in the stock market and a lot of individual companies that it named in the piece like DoorDash was a big one.

这篇文章预测，这些公司将面临极其艰难的处境。

This essay predicts these companies are gonna have a really, really hard time.

是的。

Yeah.

我对Citrini研究的这篇论文并没有太大的印象。

And I was not that impressed by the Sattrini research essay.

我认为它做了一些我不会做的逻辑跳跃，但它产生了巨大影响。

I thought it made a number of logical jumps that I wouldn't make, but it had a big impact.

人们将这篇论文视为引发华尔街大规模抛售的导火索。

People are blaming this essay for triggering a massive Wall Street sell off.

像DoorDash、美国运通和黑石这样的公司，其股价在该文章发表后立即下跌了8%以上。

Companies like DoorDash, American Express, and Blackstone, all of their stock prices dropped more than 8% immediately after this essay was published.

我们现在正处于一种市场驱动的科幻时代，任何对人工智能如何影响经济持有鲜明且合理知情观点的人，只要他们的文章像这篇一样走红，就可能引发股市数十亿美元的损失。

We are now in the era of market moving science fiction where anyone with an opinionated and reasonably informed take on what AI is doing to the economy can now trigger billions of dollars in losses in the stock market if their essay kind of catches fire as this one did.

没错，凯文。

That's right, Kevin.

因此，我呼吁所有科幻作家向证券交易委员会注册。

And that's why I'm calling on all science fiction authors to register with the Securities and Exchange Commission.

你们的想法太过强大，必须受到监管。

Your ideas are too powerful, and they must be regulated.

是的。

Yeah.

我们不会在整个节目中都讨论这一篇文章，因为我认为它只是反映了当前正在发生的更大、更有趣的现象，即人们对这一切走向的经济不确定性——人工智能将去向何方，它将对劳动力市场、企业采用AI带来的生产率提升以及我们一些最大公司的商业模式产生何种影响。

So we're not gonna spend this whole episode talking about this one essay because I think it is symptomatic of something larger and more interesting that is happening right now, which is that economic uncertainty about where all of this is headed, where AI is going, what effects that's going to have on the labor market, on the productivity gains from companies that are implementing it, on the business models of some of our largest companies.

现在一切都显得非常不确定和脆弱。

It all feels really uncertain and tenuous right now.

我想，与其逐行分析这篇论文，不如请一位真正了解经济、并且比我们更早思考这些问题的人来参与讨论。

And I thought instead of just going line by line through this essay, we should actually bring in someone who knows the economy and has been thinking about this stuff for far longer than we have.

是的。

Yes.

尽管我们很想和大家分享一些大学一年级宏观经济学的知识，但我们觉得现在是时候请专家来解答了。

As much as we would like to share with you what we remember from freshman year macroeconomics, we thought this may be a time to call in the big guns.

对。

Yeah.

所以今天，我们为大家带来与安东·科拉内克的对话。

So today, we are bringing you a conversation with Anton Koranek.

安东是弗吉尼亚大学达德商学院经济学系的教授。

Anton is a professor in the department of economics in the Darden School of Business at the University of Virginia.

自去年四月以来，他也是Anthropic经济咨询委员会的成员。

He's also, since last April, a member of Anthropics Economic Advisory Council.

我早就期待能邀请他来参加我们的节目了。

And I've been really excited to get him on the show for a long time.

我一直很欣赏他的研究，我认为他是最早尝试厘清人工智能对经济影响的经济学家之一。

I have been a fan of his work, and I would say he's been at the forefront of economists who are trying to work out what effect AI will have on the economy.

他并不是最近才开始关注这个问题的。

He did not come to this question recently.

他在这个领域已经研究了十多年，也因此广为人知，因为他愿意考虑一些可能比其他经济学家更为极端的情景。

He's been working on this for more than a decade, and he has become well known as someone who is willing to consider maybe somewhat more extreme scenarios than many of his colleagues in economics.

正因如此，我觉得他非常有趣。

And for that reason, I think he's really interesting.

而且，凯文，我想我们都迫切希望现在就能得到清晰明确的答案：究竟发生了什么？大规模失业何时会开始？

And look, Kevin, I think we all want very simple clear answers right now to exactly what is going on, exactly when might massive job loss begin.

但事实是，我们并不知道。

And the truth is we don't know.

对吧？

Right?

我们还没有足够的数据。

We do not have the data.

我们对今天的AI能力还不够了解，更不用说明天的能力了，因此无法给出一个明确的答案来说明一切即将发生什么。

We don't understand the today's capabilities well enough, much less tomorrow's capabilities, and so we cannot give you one clear answer on everything is about to happen.

但我认为，市场仅仅因为几乎没有任何实质消息就能大幅波动，这凸显了当前人们的高度焦虑。

But I think the mere fact that the markets can move so much based on almost nothing underscores how high anxiety is right now.

所以，与一位密切关注此事并能明确告诉我们已知和未知信息的人交谈，是有帮助的。

So I think it's helpful to just talk to someone who follows this stuff very closely and is able to tell us in no uncertain terms what we know and what we don't know.

那我们请他进来吧。

So let's bring him in.

在请他进来之前，你本周已经做了新的披露，说你的未婚夫在Anthropic工作，而我也要披露一下，我为《纽约时报》工作，而《纽约时报》正在起诉OpenAI、微软和Perplexity，指控它们涉嫌侵犯版权。

And before we do that, you already made your updated disclosure this week that your fiance works at Anthropic, and I will make mine, which is that I work at the New York Times, which is suing OpenAI and Microsoft and Perplexity over alleged copyright violations.

好的。

Alright.

让我们请入安东·科纳克。

Let's bring in Anton Kornack.

安东·科纳克，欢迎来到《硬核》节目。

Anton Kornack, welcome to Hard Fork.

很高兴能和你们一起直播。

Great to be on air with you.

我非常期待能和你进行这场对话。

So I am very excited to have this conversation with you.

你是我一直想邀请上节目的嘉宾。

You're a guest I've been wanting to bring on the show for a long time.

而我们现在正处在一个关键时刻，整个经济似乎都依赖于这些具有承重作用的论文——这些所谓的推演或科幻作品，你怎么叫都行。

And we are finding you at a moment where the entire economy seems to be resting on these kind of load bearing essays, these works of, you know, extrapolation or science fiction, whatever you wanna call them.

本周，我们看到了一份来自Citrini的研究报告，题为《2028年全球智能危机》。

This week, we had this Citrini research report about the twenty twenty eight global intelligence crisis.

在此之前，还有一篇论文。

Before that, it was another essay.

所以我很好奇，作为一位多年来一直研究人工智能问题的经济学家，你怎么看待这个时刻——市场对感知的微小变化都如此敏感？

So I'm very curious what you, an economist who's been looking at the issue of AI for many years now, makes of this moment where markets seem so reactive to even small changes in perception.

是的，这确实是个有趣的时候，因为我已经研究这个问题十年了，一直在等待市场意识到即将降临我们头上的变化。

Yeah, you know, it's a funny moment because I have been studying this for a decade now and I have been kind of waiting and waiting to when markets are going to wake up to what's about to hit us.

但真正引发市场大幅反应的，往往是一些看似微小、近乎随机的小事。

And then it's kind of seemingly small, almost random little things that actually produce big market reactions.

所以，是的，市场是根据情绪波动的，我想这正是其中一个例子。

So, yeah, markets move according to emotions and I guess this is one of those instances.

但在背后，也有一些非常真实的进展，我想今天我们就是要讨论这些。

But in the background, there are also some very real developments, and I guess we're here to discuss those today.

是的。

Yeah.

没错。

That's right.

我们希望今天能稍微淡化一下对话中的情绪，转而聚焦于冷峻的事实。

We're hoping that today we can maybe drain a little bit of the emotion out of the conversation and get into the cold hard facts.

所以，安东，你能告诉我们当前的经济数据对这一时刻揭示了什么吗？

So, Anton, what can you tell us about what the current economic data tells us about this moment?

究竟正在发生什么？

What is actually happening?

有数据表明正在发生真正的转变吗？

Is there data that suggests something is really shifting?

还是说这仍然更多属于一种感觉层面？

Or is this still sort of more in the realm of vibes?

这仍然属于预期的范畴。

It's still in the realm of expectations.

所以，如果你查看实际数据，可以看到人工智能对就业市场、生产率增长等方面产生了一些相对较小的影响。

So if you look at the actual data, you can see some relatively small impacts of AI on things like the job market, things like productivity growth.

但这些影响首先仍处于非常微小的范围，比如百分之零点几，其次还存在争议。

But they're still, first of all, in the territory where they're very small, like fractions of a percent, and secondly, contested.

因此，目前有几篇经济学研究论文指出，我们在初级岗位的就业市场上确实能看到一些变化。

So at this point, there are like a couple of economic research papers that say, yes, we can see something in the job market for entry level jobs.

但也有人认为，这篇论文中存在这样或那样的问题，这些结果其实可以有其他解读方式。

But there are also people who still say, well, there's this and that that's wrong in this paper and we could actually interpret these results in a different light.

简而言之，目前还没有确凿的经济数据。

So in short, there is no really hard economic data yet.

我其实担心，即使到了我们都清楚地看到这种变化的时候，经济学研究仍然会有些争议。

I'm actually afraid that even by the time when all of us are going to see, yes, this is clearly visible now, the economic research is still going to be slightly contentious.

为什么会这样？

And why is that?

是因为收集这些数据需要时间，才能在生产率数据中显现出来吗？

Is that because it just takes a while to collect all the data for these things to start showing up in productivity numbers?

是滞后效应，还是AI对经济的变革方式超出了我们目前所收集的经济数据所能捕捉的范围？

Is it the lag or is there something about the way that AI is transforming the economy that is not able to be captured in the kinds of economic data we collect?

我认为两者都有。

I think it's a little bit of both.

我们的经济统计指标部分设计得非常全面，因此需要时间来汇总。

So our economic statistics there are designed in part to be very, very comprehensive and it takes time to compile them.

它们会被修订，因为初次估算未必是最终结果。

They get revised because the first take is not necessarily the final one.

所以如果你看生产率这类指标，时间滞后的影响就特别明显，你必须接受这样一个事实：直到数据实际产生一年后，我们才能获得一个完全清晰的图景。

So if you look at things like productivity, that's where the time lags really hit you and where you really have to just live with the fact that we won't have a fully clear picture until like a year after the data has actually materialised.

但第二点是，技术进步得太快了。

But the second thing is also that the technology is advancing so rapidly.

你今天使用的ChargeGPD与一年前的版本截然不同，尤其是在编程或白领工作方面，能力强得多。

The ChargeGPD that you work with today is very different from the one a year ago and can do much more, especially when it comes to things like coding or white collar work.

那我们来深入探讨其中一项研究吧。

So let's dig into one of these pieces of research.

本月早些时候，国家经济研究局发表了一篇题为《企业AI数据》的论文。

There was a paper at the National Bureau of Economic Research from earlier this month called Firm Data on AI.

他们调查了6000名高管。

They surveyed 6,000 executives.

研究发现，70%的公司使用了人工智能，但80%的企业表示，尚未在就业或生产率方面看到任何影响。

It found that 70% of their companies used AI, but, that 80% of the firms reported that they had seen no impact on employment or productivity.

我觉得我们经常看到这类调查。

I feel like we see these kinds of surveys a lot.

也就是说，这项技术已经被广泛部署了。

That's like, you know, this technology is being widely deployed.

我们无法判断它是否真的产生了效果。

We can't tell if it's doing anything.

作为相信人工智能最终会改变经济的人，你如何理解这类研究？

How do you, as someone who does believe that AI will eventually transform the economy, make sense of this kind of research?

是的，我认为在技术的前沿能力和实际日常应用之间存在着巨大的差距。

Yeah, I think there's a very big gap between the frontier of what's possible and what is actually used in daily use.

你刚刚提到的这篇论文告诉我们，就实际企业如何使用这些技术而言，截至几个月前，影响还并不显著。

And what the paper that you just mentioned tells us is that in the field when it comes to how actual corporations are using these technologies, as of a couple of months ago, there wasn't really that big of an impact yet.

我认为这与我与高管们交流时所看到和听到的一切相符。

And I think that corresponds to everything I'm seeing and hearing when I talk to executives.

因此，人们仍处于探索阶段：我们该如何真正有效地部署这些系统？如何从炫目的演示过渡到真正提升工作效率，实现更低成本、更可靠地完成任务，达到我们一贯的工作可靠性水平？

So people are still at the stage of where they are trying to figure out how do we actually deploy these systems productively, how do we go from, let's say, the shiny demo to having a productive impact on our work where we can do more, where we can do things more cheaply and in a reliable way with the same level of reliability that we've always worked.

明白了。

Got it.

在这篇2028年全球智能论文中，有一个引起广泛关注的概念，作者称之为‘幽灵GDP’。

One of the concepts in this 2028 global intelligence essay that got a lot of attention was something that the authors called ghost GDP.

这个观点认为，随着人工智能变得越来越强大并承担更多工作，我们将拥有越来越高效的公司，创造越来越多的收入和GDP，但这些增长并不会体现在工人的口袋里，因为机器在完成这些工作。

This idea that as AI kind of gets more capable and does more work, that we will have these increasingly productive firms creating increasing amounts of revenue and GDP, but that that will not be sort of showing up in the pockets of workers because machines are doing the work.

是的。

Yeah.

这与你所做的研究一致吗？

Does that track with any of the research you've been doing?

这个‘幽灵GDP’真的是一个值得我们担忧的真实概念吗？

Is this a real concept, this ghost GDP that we should be worried about?

我对此感到担忧。

I'm worried about it.

听起来这个说法确实比我在其他地方见到的术语更吓人。

Sounds very It's definitely a spookier term than what I have encountered this under.

但坦白说，如果技术达到类似通用人工智能或强大人工智能（无论你怎么称呼它）的水平，这个观点与普遍预期非常吻合。

But frankly, it does track very much with what the general expectation is if the technology reaches the level of something like AGI or powerful AI or whatever you want to call it.

所以某种程度上，情况甚至可能比这更糟。

So in some sense, can see it's even worse than that.

一方面，将会有大量GDP并非由人类直接参与生产。

So on the one hand, there's going to be a lot of GDP that is not going to be produced by humans in the loop.

这意味着没有任何工人能够从中获益。

So that means no worker is ever going to get the benefits of that.

但另一方面，还会有相当多的经济产出根本不会体现在GDP中，因为它们被算作中间产品。

But on the other hand, there's also going to be quite a significant amount of economic production that doesn't even show up in GDP because it gets counted as an intermediate good.

只有最终消费或用于资本积累、具有特定使用寿命的最终投资才会被计入GDP。

Things only show up in GDP when it is final consumption or final investment in things in capital that we can accumulate that has a useful life of a certain period.

而人工智能经济的许多部分都不会反映在GDP中。

And a lot of the parts of the AI economy are not going to be reflected in GDP.

我很感兴趣。

I'm curious.

我所接触的经济学家们之间正存在一场争论。

There's sort of this debate going on among economists that I talk to.

他们中的一些人会说，我们根本从未见过经济增速达到硅谷有些人所认为的那样快，比如20%的GDP增长率。

Some of them will say, you know, we just don't ever see really instances of the economy growing as quickly as some of the people in Silicon Valley think it might, you know, 20% GDP growth.

这在我们的历史上是前所未有的。

That's just like unprecedented in our history.

因此，他们认为人工智能会使经济增长变得缓慢得多，可能每年只有1%到2%，这在相对意义上已经算很大了，但并非湾区一些人所设想的那种超高速增长。

And so they're expecting that AI will make things grow much more slowly, maybe a percent or two a year, which would be big in in relative terms, but not the kind of hyper growth scenario that some people out here in the Bay Area are envisioning.

然后你有像花旗研究团队这样的人说，我们即将看到前所未有的情况。

Then you have people like the folks at Citi Research saying, we're about to see something we've never seen before.

我们即将看到整个经济脱离所有这些周期性模式。

We're about to see an entire economy sort of becoming unmoored from any of these cyclical patterns.

那么，在这个光谱上，你站在哪一边？

So where on that spectrum do you fall?

数据把你引向了哪个方向？是每年1%或2%的缓慢增长，还是10%或20%的超高速增长？

Where does the data lead you between the sort of slow growth 1% or 2% a year to the 10% or 20% a year hyper growth scenario?

是的，关于这一点，我想说两点。

Yeah, I'll say two things about that.

第一点是，这个故事尚未写就。

The first one is that the story has not been written yet.

如果我们以非常不负责任的方式开发这项技术，确实有可能出现自我复制并迅速扩张的情况，从人工智能的角度来看，这可能导致三位数的GDP增长。

And there is a possibility that if we develop this technology in a really irresponsible way, that we could actually see some self reproduction that takes off and that leads to triple digit GDP growth numbers if measured from the eyes of the AI.

但如果我们以让普通人受益的方式部署这项技术，我认为三位数的增长完全是不现实的。

But if we deploy the technology in a way that it makes the average person better off, then I think triple digit growth numbers are completely unrealistic.

那会导致过于剧烈的动荡。

They would lead to way too much disruption.

至于1%的增长，我也不确定，但从我的角度来看，这显然太低了，不可能实现。

And then I'm not quite sure, I think just 1% is definitely going to be too low to be realistic from my perspective.

在非常乐观的情景下，我认为我们有可能实现个位数的低双位数增长率；但我要说明的是，这预设的不仅是认知型AI，而是像OpenAI章程中所定义的完整AI——即高度自主、能够完成大多数具有经济价值工作的系统。

In really optimistic scenarios I think we could get to low double digit growth rates and I should say that presupposes not just cognitive AI but full AI in the way that is, for example, defined in the charter of OpenAI where they say systems that are highly autonomous and that can perform most economically valuable work.

因此，这还包括物理层面，即机器人技术的部分。

So that also includes a physical component that includes the robotics part.

否则，它对GDP的影响不会那么大，因为经济的大部分并不仅仅发生在人们坐在电脑前的时候。

Otherwise, it won't have that big of an effect of GDP because the majority of the economy isn't just sitting in front of a computer.

对。

Right.

我认为，现在许多关注股市和那些热门文章、试图理解这一切的人，都感到强烈的认知失调。

And I think a lot of people right now who are looking at the stock market and these viral essays and trying to make sense of this all are feeling a lot of cognitive dissonance.

因为一方面，一些看似非常聪明的人说，人工智能正在改变一切。

Because on one hand, we have people who seem very smart saying AI is transforming everything.

每家公司都在以不同于几个月前的方式运作。

Every company is doing things differently than it was a couple months ago.

我们正步入一片未知的领域。

We are headed into uncharted territory.

但当你环顾四周，我们依然保持着低于5%的失业率。

And then you look around and we're, you know, we're still below 5% unemployment.

我们还没有看到显著的生产率提升。

We still don't see a huge productivity boost.

大多数在工作中使用这些技术的人，只在使用旧模型，或者他们的IT部门不允许他们使用自主编码工具，因此，确实看起来，那些关注技术发展或预言未来的人，与我们周围可观察的现实之间，正出现越来越大的脱节。

Most people who are using this stuff at work are only using, you know, older models or their IT department won't let them use the agentic coding And and so it it does seem like we are seeing a a growing disconnect between what people who are looking at the technology or saying is going to happen and the observable reality around us.

那么，你如何看待这种脱节？人们应该如何看待这些关于快速变革的预测？

So what do you make of that disconnect and how should people be feeling about these projections of rapid change?

是的，我们之前提到过的一个方面是前沿能力与实际应用之间的差距。

Yeah, so the one part that we already touched upon is the gap between the frontier capabilities and the actual implementation.

这一部分是真实存在的，而且非常重要。

That part is real and that is very significant.

但这也是一种随着时间推移可能会逐渐消失的现象，对吧？

That's also something that is kind of bound to disappear over time, right?

但第二点是，我们听到的所有预测本质上都是外推，当人们看到人工智能系统在过去一年里的巨大进步时，反应却大不相同。

But the second part is that ultimately all the projections that we are hearing are extrapolations and people react very differently when they see how much the AI systems have improved, let's say over the past year.

有些人自然而然地得出结论：既然如此，那就继续外推吧，这些系统在短短几年内肯定会比任何人类都聪明得多。

Some people just naturally jump to the conclusion, well, let's extrapolate this and of course these systems are going to be way smarter than any human within just a small number of years.

而另一些人则认为，我们大脑的运作方式如此独特，机器在很长一段时间内都无法复制，这些机器的能力最终会趋于一个低于人类大脑水平的平台。

And then there is another camp that says, well, what our brains are doing is so special that machines won't be able to replicate it for a very long time and these machines are going to kind of asymptote to somewhere below our brains' capabilities.

坦白说，这两种观点都属于推测。

And frankly both is a speculative position.

我个人首先愿意接受其中的不确定性，我认为我们都应该如此；但如果你非要我做一个更让我安心的猜测，我会说，能力可能会继续提升。

I personally am first of all willing to embrace the uncertainty about it, and I think we all should, but if you ask me to make one guess that I feel more comfortable about, I would say capabilities are probably going to continue to increase.

我认为在近期我们面前并没有任何明确的限制。

And I don't think there is any clear limit in front of us in the near term.

因此，我预计将会产生非常重大的经济影响。

And so I do expect that there's going to be very significant economic impacts.

是的。

Yeah.

那么让我们再往更远的未来推演一下。

So let's extrapolate a little bit further into the future.

2017年，你合著了一篇论文，其中提出：'人工智能的进步更有可能取代人类劳动，甚至完全取代工人，而不是作为大多数工作的补充。'

In 2017, you co wrote a paper where you suggested that, quote, progress in AI is more likely to substitute for human labor or even to replace workers outright than it is to be a complimentary for most jobs.

当时你提出这个观点时，是走在了很前沿的位置。

At the time, you were way out on a limb when you wrote that.

我想，如今你比以往任何时候都更认同这一点。

I I I imagine you feel that today more than ever.

但是什么让你如此有信心？

But what is giving you that confidence?

你觉得我们到目前为止，在多大程度上已经感受到，这种情况比2017年时更接近现实了？

And to what degree do you feel like we've we've started to see it maybe feel more true than it did in 2017?

是的。

Yeah.

为了明确一下，我当时的预测始终是指那些达到或超越通用人工智能水平的AI系统，而不是2017年那些连狗和松饼都分不清的原始系统。

And just to be sure, that was always meant to be a prediction about AI systems that are essentially at the level of AGI or beyond, not for the literal systems we had in 2017 that could barely tell apart a dog and a muffin.

所以，我认为我的观点源于我对神经科学和计算机科学的研究；在某种程度上，一旦深度神经网络变得强大起来，我就觉得很难不得出这样一个结论：这些系统最终很可能能够完成人类大脑能做的几乎所有事情，而且它们所受的限制要宽松得多——比如，它们不需要被塞进一个小小的头骨里。

So I think ultimately where my perspective is coming from is that I have studied neuroscience, I have studied computer science and at some level, you know, once basically deep neural networks became powerful, I felt it is hard to not make the conclusion that, well, it looks like eventually these systems will be able to do pretty much anything that our brains can do, and they are subject to much, much more relaxed constraints, like they don't need to fit into a tiny human skull.

我们可以几乎无限制地扩展它们。

We can scale them almost without bounds.

某种程度上，过去十年我们所看到的正是如此，对吧？

In some sense, that's what we have seen over the past decade, right?

我们看到了持续不断的、海量的扩展。

We have seen lots and lots and lots of scaling.

如今这些系统消耗的能量相当于一座城市，而人类大脑的能耗却只相当于一个节能灯泡。

At this point these systems consume the energy of cities as opposed to what our brain does, which is the energy of an energy efficient light bulb.

但这还不是极限。

And that's still not the limit.

模型规模仍在扩大，能力也在不断提升，当然算法也在变得越来越好。

There's still increasing in size, increasing in capabilities, and of course the algorithms are getting better and better.

因此，基于这种观点，我看不出有什么自然的限制，更不用说有什么限制会低于我们人类的智力水平了。

So based on that perspective, I just don't see why there would be any natural limit and certainly not why there would be a limit that's below our human intellectual capabilities.

对。

Right.

那么问题来了，当这个世界到来时，工作会怎么样？

And I think the question then is, as this world arrives, what happens to the jobs?

在经济学中，有些听众可能还不熟悉所谓的‘劳动总量谬误’。

And, in economics, some of our listeners may not have, familiarized themselves yet with, what's called the lump of labor fallacy.

对吧？

Right?

这种观点认为，工作总量是固定的，任何被自动化取代的工作都将永远无法被替代。

The idea that there are a fixed number of jobs to be done and any job lost to automation will therefore never be replaced.

我们称之为谬论，因为自从经济学家开始追踪以来，自动化始终创造了更多的就业机会。

We call it a fallacy because ever since economists started tracking it, automation has always led to the creation of more jobs.

安东，你在另一次采访中提到，经济学家很难在这一点上转变立场，因为他们为此谬论争论了太久。

Anton, you mentioned in another interview that it's hard for economists to pivot on this because they fought this fallacy for so long.

作为一名经济学家，说‘这次真的不同了，工作正在真正消失’，这种感觉是怎样的？

What does it feel like to be an economist saying, actually, this time people should worry that the jobs are going away for real?

是的，这种感觉确实很奇怪。

Yeah, it does feel very strange.

在过去十年里，我从其他经济学家那里收到了不少批评。

And I have gotten a fair amount of flack from my fellow economists over the past decade.

不过，过去一两年里，许多同事都说：我仍然不完全认同你的观点，但我很高兴有人在思考这个问题，我不会完全排除这种可能性。

Although I'll say over the past year or two or so many of my colleagues have said, well, I still don't entirely buy your worldview, but I'm glad somebody is thinking about it and I wouldn't rule it entirely out.

每当经济中失去一份工作，就认为这个人将永远失业，这是一种谬论。

It is a fallacy that whenever a job is lost in the economy, that person is going to remain unemployed forever.

但我想我们真正应该关注的是对人类劳动力的总体需求。

But I think what we really want to look at is overall demand for human labour.

如果这种需求曲线因人工智能系统能够取代越来越多的人力而向下移动，那么最终的结果可能是就业岗位数量、工资水平，或两者同时收缩。

And if that demand curve shifts downwards because AI systems can supplant more and more of it, then what that's ultimately going to imply is that either the quantity of jobs or the wage levels or both may contract.

不过，我也应该指出，劳动市场也可能保持稳定，只是增长速度不如经济其他部分那么快。

Now, I should say there's also the possibility that labour continues to do okay and it just doesn't grow as fast as the rest of the economy.

换句话说，劳动在产出中的占比会下降，但至少我们在绝对水平上并未落后。

So in other words that the labour share of output is going to shrink, but at least we are not falling behind in absolute levels.

我们的经济理论告诉我们，是劳动市场仅相对萎缩，还是彻底失去优势，这两种结果中哪一种会实现，部分取决于自动化的速度。

Our economic theories tell us that whether that outcome or the one where labour just flat out loses, which one of those outcomes materialises depends in part on the speed of automation.

而且，为了我们所有人着想，我正双手合十，希望我们只会面临相对损失，而不会遭遇绝对损失。

And, you know, like for all of our sakes, I'm crossing my fingers and I'm hoping that we will only lose out in relative terms and not in absolute terms.

但目前，我认为我们还没有任何数据能以足够的确定性判断这两种结果中哪一种会发生。

But right now, I don't think we have any data that can tell us with any degree of certainty which of those outcomes is going to happen.

安东，我想回到你几个问题前提到的一点，你说你预计前沿人工智能能力与职场实际应用之间的差距会随着时间推移而缩小？

Anton, I want to return to something that you said a few questions ago, which was that you expect the gap between frontier AI capabilities and sort of workplace diffusion, how how workers are actually using this stuff to shrink over time?

我对这一点并不太确定。

I'm not so sure about that.

我花了很多时间与企业及教育机构的领导者交流，但我不认为他们的部署速度有明显提升。

I've spent a lot of time talking with leaders of businesses and educational institutions, And I would not say that their speed of deployment is increasing all that much.

你知道，他们担心安全和隐私问题，有太多理由让他们不愿意轻易把这些东西引入工作流程。

You know, they've got security fears and privacy fears, lots of reasons why they don't wanna just start throwing this stuff into their work.

所以你能帮我理解一下，为什么你认为这个差距可能会缩小吗？

So maybe help me understand why you believe that gap might shrink.

我可能表达得不够清楚，我的意思是，当前的能力最终会扩散到整个经济体系中。

I may have expressed myself a little bit unclearly, but what I meant to say is that the current capabilities are eventually going to diffuse to the economy.

当然，到那时，我完全同意你的观点，实际能力已经进一步提升了。

And of course, by that time, I'm very much with you, the actual capabilities are going to have advanced even further.

如果我们正处于能力飞速提升的轨迹上，那么这个差距本身可能确实会扩大，而不是缩小。

And if we are on this trajectory of skyrocketing capabilities, the gap itself may indeed go up rather than down.

我认为这可能是最有可能的结果。

I think that is the most plausible outcome probably.

但我真正想强调的是，我们目前拥有的能力最终会普及开来，并且首先会对生产力产生广泛影响，因为如今的AI系统在许多方面仍然与劳动者高度互补。

But what I really wanted to emphasise is that the capabilities that we currently have are eventually going to diffuse and are eventually going to have broad, for now at first, productivity effects, because right now AI systems are still in many ways very complementary to workers.

但一旦它们达到可以替代人类的水平，就会对劳动力市场产生一些负面影响。

But as soon as they reach the level where they become substitutes, there's also gonna be some adverse labor market effects.

我来告诉你我想要什么。

I'll tell you what I want.

对。

Yeah.

说吧。

Go ahead.

我想知道人们在工作中实际上是如何使用人工智能的。

I wanna know how people are actually using AI at work.

因为我们现有的数据主要是自我报告。

Because what we have, data that we have is largely self reports.

我认为一些公司夸大了他们使用人工智能的程度，因为他们想显得前沿、未来感十足，展示我们有多么大的转变。

And I think some firms have exaggerated how much they are doing with AI because they want to appear to be cutting edge and futuristic and look how transformed we are.

我认为一些人，尤其是员工，低估了他们使用人工智能的程度，因为他们对此感到尴尬，或者他们对。

And I think some people, especially workers, are downplaying how much they're using AI because they're embarrassed about it or they Yeah.

这违反了他们公司的IT政策，或者他们不确定自己是否被允许这么做。

It's against their company's IT policy or they're not, you know, they're not sure they're allowed to be doing it.

因此，我认为我们目前缺乏关于员工在工作中实际如何使用AI的细致数据，也无法确定AI是加快了还是拖慢了他们的工作效率。

And so I just don't think we have very good granular data about what people are actually doing with AI at work and whether it is speeding them up or slowing them down.

如果我能拥有一个水晶球，我想我不需要水晶球。

And if I could like have a crystal ball, I guess I what I wouldn't need is a I wouldn't need a crystal ball.

我需要的是一套监控系统。

I would need like a surveillance apparatus.

凯文想监控员工的电脑。

Kevin wants to spy on workers' computers.

但是

But

我们只是

I just We

我们确实有一点这样的数据。

we do have a little bit of that.

OpenAI 和 Anthropic 几乎实时地发布其系统实际使用情况的数据。

Both OpenAI and Anthropic publish data on how their systems are actually used almost in real time.

这让我们对当前状况有了一些了解，但信息毕竟有限。

And that gives us a bit of a picture of where we are, but it tells you only so much.

你能给我们的听众大致介绍一下吗？有没有两三个核心指标或报告，当你看到它们出来时，会立刻意识到：好了，关键信息来了。

Can you give our listeners a sense of like, are there two or three kind of core indicators or core reports that as you come out, you think, okay, here we go.

我终于能更新一下，看看我们是否正越来越接近一个大规模工作自动化的新未来。

I finally get to update and see if we're getting closer to a future of, you know, mass job automation.

有哪些指标在更新时，会改变你的认知？

What are those things that as they come in are updating your understanding?

能力的绝对水平可能是最重要的一个指标。

So the sheer level of capabilities is probably the most important one.

你可以追踪任何你感兴趣的基准测试，或者一些综合性的基准，它们能告诉我们 AI 系统在哪些方面仍显落后，在哪些方面已经表现得非常出色。

Like you can follow whatever benchmarks you want or some amalgamate of benchmarks that tells us where the AI systems were still lagging, where they are doing already pretty amazing well.

目前最大的一个短板是，这些系统无法像人类那样动态学习——当前的大型语言模型一旦训练完成，其权重就会被冻结，不再更新。

And one of the biggest shortcomings right now, but of course from the perspective of workers that's great because it makes us more complementary, is that these systems are not learning dynamically the way that current LLMs work as they are trained once and after that the weights are frozen in place.

这意味着在许多工作应用中，即使存在非常基本的错误，也必须一遍又一遍地重复同样的错误，因为它们从这些错误中能学到的东西非常有限。

And that means for a lot of work applications, even if there are very basic mistakes, have to go through the same mistake again and again and again and again because they can learn only so much from it.

所以，这是我正在期待的另一项突破。

So that's another sort of breakthrough that I'm looking for.

然后，我经常关注的第三个图表是关于AI能够自动化任务时长的图表。

And then maybe a third chart that I'm regularly following is this matter chart that looks at how long of a task AI can automate.

我认为他们通常每七个月就会发现这个时间范围翻倍。

And I think they usually find every seven months that timeframe doubles.

观察这一趋势如何持续发展，也有助于我们理解指数增长轨迹是否依然保持，甚至像最近看起来那样在加速，还是我们已经接近平台期。

And looking at how this is continuing is also quite helpful in understanding whether the exponential growth trajectory is intact or maybe even accelerating as it has seemed recently, or whether we are anywhere near plateauing.

这是支撑整个经济的图表。

This is the chart that's holding up the entire economy.

是的。

Yeah.

我们回来后，继续与安东·科纳克对话。

When we come back, more with Anton Kornack.

我是贾德森·琼斯。

I'm Judson Jones.

我是《纽约时报》的记者兼气象学家。

I'm a reporter and meteorologist at The New York Times.

近二十年来，我一直报道极端天气，由于气候变化，极端天气正变得越来越严重，及时、准确的天气信息也变得愈发重要。

For about two decades, I've been covering extreme weather, which is getting worse because of climate change, and it's becoming more important to get timely and accurate weather information.

因此，我们发送这些定制化的简报，提前最多三天告知您可能影响您或您关心的地区的极端天气。

That's why we send these customized newsletters letting you know up to three days in advance about extreme weather that could impact you or a place you care about.

在《纽约时报》，您可以确信我们发布的所有内容都基于我们所能获得的最准确的科学数据，因为我们希望您能够实时做出关于如何安排生活的决策。

At The Times, you can be confident that everything we publish is based off the most accurate scientific embedded information available to us because we want you to be able to make real time decisions about how to go about your life.

这种工作让订阅《纽约时报》变得如此有价值，也是您支持基于事实的独立新闻的方式。

This is the kind of work that makes subscribing to The New York Times so valuable, and it's how you can support fact based independent journalism.

如果您想订阅，请访问 nytimes.com/subscribe。

So if you'd like to subscribe, go to nytimes.com/subscribe.

安东，你提到十年前你刚开始撰写关于人工智能、自动化以及潜在失业和经济转型的文章时，你的经济学同行们非常怀疑。

Anton, you mentioned that when you started writing about AI and automation and potential job loss and economic transformation a decade ago, your colleagues in economics were very skeptical.

你在你的领域里被视为一个异类。

You were seen as something of an outlier in your field.

是的。

Yep.

我一位资深同事问我，你真的确定要为了这个放弃你的职业生涯吗？

One of my senior colleagues asked me, are you really sure you wanna throw away your career over this?

所以，显然，这种情况现在已经不再存在了。

So, obviously, that's no longer true.

现在有很多主流经济学家正在关注这些问题。

You now have many mainstream economists looking at these issues.

目前，有哪些观点让你在你的专业领域中处于边缘地位，而你的许多同事并不认同？

What are the ideas right now that you believe that put you on the fringes of your profession that many of your colleagues disagree with?

我确实觉得，在经济学界，认真对待人工智能通用化这一概念仍然是一种边缘观点。

So I do have the impression that taking the notion of something like artificial general intelligence really seriously is still a fringe perspective in the economics profession.

你说得对，确实有越来越多的人开始接受这一点，但这仍然是一个数量较少但声音越来越大的少数群体。

You're right that there are more people coming around to it, but it's still a small and increasingly loud minority.

我还相信，如果我们真正实现了通用人工智能，那将不是终点，而是经济发生重大变革的开端。

I also believe that if we seriously reach AGI, that's not going to be the end, but it's going to be the beginning of a really significant transformation of the economy.

在这方面，我可能比我的经济学家同行们更加边缘化。

And in that respect, I'm probably even more on the fringe of where my fellow economists are.

是的。

Yeah.

你曾写过关于这种超指数增长的可能性。

You've written about this possibility of hyperbolic growth.

基本上，这种情况发生在我们实现递归自我改进时，人工智能开始构建更先进的人工智能，进而建立机器人工厂，基本上创造自己的经济。

Basically, happens if we get recursive self improvement, the AI start building better AIs, they start building robot factories and basically create their own economy.

你实际上尝试模拟过，当人工智能达到这一关键转折点时，经济会发生什么变化。

And you actually tried to model what might happen in an economy where AI reached this critical inflection point.

你发现了什么？

What did you find?

是的，我们发现的第一件事是，会出现一系列相互强化的反馈回路。

Yeah, so the first thing that we found is there's going to be a whole bunch of feedback loops that will mutually reinforce each other.

假设我们在软件层面确实达到了递归自我改进的阶段。

So let's say we do reach this point of recursive self improvement on the software side.

能够实现这一点的AI系统将推动硬件方面的研究，加速该领域的技术进步。

AI systems that can do this are going to feed into the research process on the hardware side and are going to accelerate hardware research, the technological advances on that front.

此外，它们还将加速任何需要认知能力和智能工作的领域的研究。

Moreover, they are also going to accelerate research in anything else where cognitive work, smart things can be helpful.

比如，解锁更多廉价能源，如核聚变等，并制造更先进的机器人。

Let's say, for example, unlocking additional cheap energy sources like fusion and so on, and creating better robots.

所有这些方面相互促进，因为这些进步反过来又帮助AI进一步发展。

And all of these things feed into each other because those advances in turn help the AI advance more.

将所有这些因素综合起来，我们的模型中会出现远超指数级的增长。

And if you put it all together you can get vastly super exponential growth in our model.

这是一种通向奇点的双曲增长。

It is hyperbolic growth leading to a singularity.

物理学告诉我们，真正的奇点不可能发生，因为最终总会遇到某种资源限制，但我预计，现实世界中的这些反馈回路将推动巨大增长，直到我们尚未完全识别的新瓶颈出现。

Physics tells us that a literal singularity can't actually happen because there's going to be some resource limit at some point, but what I expect is that these feedback loops in the real world would lead to massive growth until some new bottleneck that maybe we haven't quite identified yet will be reached.

我很好奇，你知道，你过几分钟就得去给研究生上课了。

I'm curious, you know, you you have to go in a few minutes to teach your graduate students.

你所研究的内容，有没有改变你告诉学生该如何思考他们职业生涯的方式？

Has how has what you have studied changed what you tell your students about how they should think about their careers?

几年前，我决定干脆坦率地表达我对这个问题的看法。

You know, a couple years ago I've decided, well, I will just be blunt about my beliefs about this.

我告诉我的研究生，我不确定当他们毕业时，经济研究岗位是否还存在。

I am telling my graduate students that I'm not 100% sure if there will still be jobs for economic researchers by the time that they graduate.

我为他们默默祈祷，希望还有岗位，但目前我们不能对此抱有把握。

I'm crossing my fingers for them I hope that there will be, but I don't think we can count on it at this point.

我认为，我们都必须面对一个根本性的不确定性：几年后经济将走向何方。

And I think all of us have to face this fundamental uncertainty about where the economy is going to be in a couple of years.

这对你从研究生那里获得的课程评价有什么影响吗？

And how has that affected your course reviews that you get back from the grad students?

这是个非常好的问题。

That's a very good question.

我还没有进行系统的统计分析，数据点也不够多，无法确定AI是提高了还是降低了我的教学效率。

I have not done a systematic statistical analysis and there aren't enough data points to say for sure whether AI has increased or reduced my teaching productivity.

明白了。

Got it.

是的。

Yeah.

明白。

Got

它。

it.

说到效率，我想问问你关于我一直在研究的一个框架，这个框架用于思考AI可能如何改变经济。

Speaking of productivity, I wanna ask you about this framework that I've been working on for thinking about how AI might transform the economy.

基本上，在我看来，这里有三种可能的结果。

Basically, as I see it, there are three possible outcomes here.

一种是‘笨重的巨人’情景，即一些大公司主导经济，但它们太过缓慢且受监管过多，无法迅速采用所有新的AI技术。

One is kind of the lumbering giants outcome, where you have these big companies that dominate the economy, and they're just too slow and too regulated to really adopt all the new AI stuff quickly.

因此，经济只是缓慢地继续前行，也许每年增长百分之一到二，但没有任何根本性的变化。

And so the economy just kind of chugs along for a while, maybe growing at a percent or two a year, but nothing fundamentally changes.

第二种选择是‘飞奔的巨头’情景，即这些大公司真正行动起来，开始迅速前进。

The the second option is the sprinting giants outcome, which is where these big companies actually get their acts together and start moving really quickly.

也许它们裁员很多，也许它们创造了大量新职位，但它们的生产率大大提高，十年后的经济仍然由今天这些巨头公司主导。

Maybe they lay off a bunch of people, maybe they create a bunch more new jobs, but they're much more productive and the economy ten years from now is still dominated by the same giant companies we have today.

然后是第三种可能性，即‘死亡的巨头’情景，即如今所有主导市场的公司都会被使用人工智能的竞争对手击垮，这些对手所用的劳动力只有它们的百分之一甚至千分之一，旧经济将被这种全新的、由人工智能驱动的经济彻底吞没。

And then there's this sort of third option, which is the dead giants outcome, which is where you have basically every company that dominates today is going to be crushed by a competitor using AI with, you know, one hundredth or one thousandth of the labor that they have, and they were essentially going to see this this sort of swallowing of the old economy by this new AI powered one.

这些情景中，你认为哪一个更有可能？而且，这样思考可能的结果真的是正确的方向吗？

Those scenarios, is there one that you think is more plausible and is that even the right way to be thinking about the possible outcomes here?

我认为这些情景很值得思考，我最倾向于认为我们会看到第二种和第三种情景的混合：一些凭借现有优势的飞奔巨头会表现不错，同时也会有一些领域，新兴企业会压倒那些迟缓的巨头，用你的话来说就是这样。

I think those are interesting scenarios to think about and my best bet would be that we'll see a mix of the second and third scenario, that there's going to be some sprinting giants that are going to do okay given their incumbency advantages, and that there's also going to be some sectors where newcomers are going to overpower the slumbering giants, to use your analogies here.

是的，最终我认为这项技术会广泛传播。

Yeah, And ultimately, I do think that the technology will diffuse.

至于这种传播是通过现有公司还是新兴企业实现，主要取决于这些巨头行动的速度有多快。

And whether that's through the existing companies or through the newcomers, that depends largely on how fast the giants are gonna move.