巴菲特谈AI对人类社会的影响 — 文稿

M10:00

嗨，Warren。

Hi, Warren.

这个问题挺有意思的。

This one's a fun one.

Thomas Kemey 来了。

Thomas Kemey is here.

他是一位二十七岁的股东，来自 California 的 Kentfield。

He's a 27-year-old shareholder from Kentfield, California.

在问这个问题之前，我得先说明一下，他十七年前也来过这里。

And I should preface this question by saying that he was here 17 years ago.

那时候他十岁，在观众席里问过你一个问题，问你 internet 会不会伤害 Berkshire 的一些投资。

At 10 years old, asked you a question from the audience, asking you if the internet might hurt some of Berkshire's investments.

当时你说，你想先看看事情会怎么发展。

At the time, you said you wanted to see how things would play out.

现在他把这个问题更新了一下。

He's now updated the question.

对于 artificial intelligence，除了你已经谈过的 autonomous driving 和 GEICO 之外，你觉得它对 Berkshire 旗下其他业务会有什么影响？

What do you think about the implications of artificial intelligence on Berkshire's businesses beyond autonomous driving and GEICO, which you've talked about already?

你跟 Bill Gates 聊这些的时候，有没有想过还有哪些业务会受到最大的冲击？

In your conversations with Bill Gates, have you thought through which other businesses will be most impacted?

还有，你觉得再过十年，受 artificial intelligence 的影响，Berkshire 现在这些业务的员工人数会明显变多，还是明显变少？

And do you think Berkshire's current businesses will have significantly more or less employees a decade from now as a function of artificial intelligence?

好吧，我这是把几个问题揉到一块儿了。

Well, I've mixed a couple questions together.

对。

Yeah.

M20:54

我当然对 artificial intelligence 没有什么特别独到的见解，不过我敢说，未来几十年，甚至可能更短的时间里，这个领域一定会发生很多事。

I certainly have no special insights on artificial intelligence, but I will bet a lot of things happen in that field in the next couple of decades and probably a shorter time frame.

我当然觉得，它们应该会带来推动作用。

They should lead, I would certainly think.

不过再说一次，我在这个话题上也没太多可贡献的，但我确实认为，它肯定会让某些领域的就业大幅减少。

But again, I don't bring much to this party, but I would certainly think they would result in significantly less employment in certain areas.

但这对社会是好事。

But that's good for society.

这对某一家具体企业未必是好事，不过我们把这个问题推到极端来看。

And it may not be good for a given business, but let's take it to the extreme.

假设一个人只要按下一个按钮，基本上就能通过各种机器、robotics 以及各类技术，产出我们这个国家现在全部的产出。

Let's assume one person could push a button and essentially, through various machines, robotics, all kinds of things, turn out all of the output we have in this country.

也就是说，产出总量和我们现在一样多。

So everybody's— there's just as much output as we have.

这一切都是由一个人完成的，而不是现在这一亿五千多万就业人口来完成——那你说，这个世界到底是不是变得更好了呢？

It's all being done by, you know, instead of 150-some million people being employed, one person, you know, is the world better off or not?

嗯，显然这样一来，每周工作的时长肯定会更少，诸如此类。

Well, it certainly would work— involve less hours a week of work per week, and so on.

我的意思是，这本来会是件好事，但它也会要求人们彼此之间怎么相处、他们对政府有什么期待，等等这些方面，发生巨大的转变。

I mean, it would be a good thing, but it would require enormous transformation in how people relate to each other, what they expect of government, you know, all kinds of things.

当然，从现实角度来说，继续工作的人肯定不会只有一个。

And of course, as a practical matter, more than one person would keep working.

但如果这样去推演这个想法，你肯定会觉得，这就是 artificial intelligence 取得重大进展之后的后果之一。

But pushing the idea that way is one of the— you'd certainly think that's one of the consequences of making great progress in artificial intelligence.

而且从长远看，这对整个社会是非常有利的。

And that's enormously pro-social eventually.

但它在其他方面也会带来巨大的颠覆，而且对于民主制度以及民主制度该如何应对，也可能造成很大的问题。

It's enormously disruptive in other ways, and it can have huge problems in terms of a democracy and how it reacts to that.

这和我们在 trade 上遇到的问题有点类似。

It's similar to the problem we have in trade.

trade 对社会是有益的，但那些每天享受 trade 好处的人，在 Walmart 买袜子或者别的进口商品时，并不会看到价格标签上写着：你现在付的是 X，但如果这是国产的，你就得付 X 再加上这么几美分。

Where trade is beneficial to society, but the people that see the benefits day by day of trade don't see a price at Walmart on socks or whatever they're importing that says, you know, you're buying— you're, you're paying X, but you would pay X plus so many cents if you bought this domestically.

所以他们得到的是这些零零碎碎的好处，而且这些好处还是看不见的。

So they're getting these small benefits and invisible benefits.

而那个被它伤到的人，也就是自由贸易里的那个倒霉蛋，会非常切身地感受到这种冲击。

And the guy that gets hurt by it, who's the roadkill of free trade, feels it very specifically.

然后这就会反映到政治上。

And that translates into politics.

所以在我看来，如果生产率大幅提高，世界会怎么去适应，这件事就会变得非常不确定。

And so you can— it gets very uncertain as to how the world would adjust, in my view, to great increases in productivity.

所以虽然我对这东西其实并不了解，但我会觉得，artificial intelligence 会带来巨大的社会效益；不过如果它来得很快——而我觉得确实有可能——那它在政治上的影响就会非常难以预测。

And without knowing a thing about it, I would think that artificial intelligence would have that hugely beneficial social effect, but a very unpredictable political effect if it came in fast, which I think it could.

M33:45

Charlie？

Charlie?

嗯，你描绘的是一个很有意思的世界，在那个世界里，所有人都在做交易，交易内容就是我教你打高尔夫，你给我染头发。

Well, you're painting a very funny world where everybody's engaged in trade, and the trade is I give you golf lessons and you dye my hair.

那种世界有点像 Kuwait 的王室家族之类的。

And that would be a world kind of like the royal family of Kuwait or something.

而我，我不觉得如果所有东西都由一个人生产，剩下我们其他人就只是休闲度日，这对 America 会是件好事。

And I, I don't think it would be good for America to have everything produced by one person and the rest of us just engaged in leisure.

M24:10

那如果我们的生产率只是提高到两倍呢？

How about if we just got twice as productive?

M34:12

什么？

What?

M24:13

我是说，如果我们在很短时间里把生产率提高到两倍，让七千五百万人就能做完现在一亿五千万人在做的事呢？

How about if we got twice as productive in a short period of time so that 75 million people could do what 150 million people are doing now?

M34:20

我觉得你会惊讶于人们会多快对此作出反应。

I think you'd be amazed how quickly people would react to that.

M24:24

怎么个反应法？

In what way?

M34:24

正面的反应。

Favorably.

这其实就是当年那个时期发生的事，大家现在回忆起 Eisenhower 年代都特别有感情，那时候大概每年增长百分之五之类的。

And that's what happened during the period, which everybody remembers with such affection, back in the Eisenhower years, 5% a year or something.

人们很喜欢。

People loved it.

没人会抱怨说，自己现在有空调了，而以前没有。

Nobody complained that they were getting air conditioning and they didn't have it before.

也没人想回到美国南方那种又臭又闷、满身是汗的夜晚。

Nobody wanted to go back to stinking, sweating nights in the South.

M24:48

嗯，如果是把每个人的工时都砍掉一半，那是一回事。

Well, if you cut everybody's hours in half, it's one thing.

但如果是解雇一半的人，另一半人继续工作——我只是觉得那样事情就会变得非常难以预测。

But if you fire half the people and the other people keep working— I just think it gets very unpredictable.

我的意思是，我觉得这次选举里我们就已经看到了一些这种情况，因为我觉得——

I mean, I think we saw some of that in this election because I think that—.

M35:01

嗯，我们已经适应了非常大规模的这种变化。

Well, we've adjusted to an enormous amount of it.

只不过它是以每年百分之几的速度慢慢发生的。

It just came along a few percent per year.

M25:07

对，那问题就是，我们是不是得担心——

Well, and the question then is, do we have to worry about—.

M35:10

我不觉得你需要担心它会以每年百分之二十五的速度发生。

I don't think you have to worry about it coming on at 25% a year.

你知道，我觉得你真正该担心的是，它的增速会低于每年百分之二。

You know, I think you have to worry about it, you're gonna get less than 2% a year.

那才是让人担心的。

That's what's worrisome.

M25:22

好，我们继续。

Okay, we'll move on.

不过你知道，这绝对是个很吸引人的话题，看看这件事最后会怎么发展。只是如果在某种情况下，比如说，GEICO 现在雇了三万六千名员工，然后，呃，如果同样的职能——甚至几乎所有同样的职能——都能照样完成，但只需要五千人或者一万人就够了，那就会非常非常难预测。

But it will be, you know, it's an absolutely fascinating subject to see what happens with this, but it gets very, very hard to predict if in some way, you know, we've got 36,000 people, say, employed at GEICO, you know, and, and, uh, if some— if you could perform all the same functions, virtually all the same functions even, and do it with 5,000 or 10,000 people.

而且如果这种变化来得很快的话。

And it came on quickly.

而且同样的事情，也正在很多很多其他领域发生。

And the same thing was happening in a great many other areas.

你知道，我觉得我们以前从来没有经历过像那样的事情。

You know, I don't think we've ever experienced anything quite like that.

而且也许，我们未来也不会再经历那样的事情。

And maybe we won't experience anything like it in the future.

我对 AI 了解得没有那么多。

I don't know that much about AI.

M36:05

但是，但是，呃，我觉得，我觉得你不用担心那个。

But, but, uh, I don't think, I don't think you have to worry about that.

M26:08

嗯，那是因为我八十六岁了。

Well, that's because I'm 86.

M36:11

它不会来得那么快。

It's not going to come that quickly.

M26:14

好，Greg。

Okay, Greg.