How AI Will Answer Questions We Haven’t Thought to Ask | Aravind Srinivas | TED

翻译人员: Yip Yan Yeung 校对人员: Bruce Wang

There are a couple of ways I'm not a traditional tech founder.

我有很多地方不像一个 传统的科技公司创始人。

I never dropped out of college.

我从来没有从大学辍学过。

(Laughter)

（笑声）

In fact, I kept going.

而是一直读了下去。

I'm an academic, you could say.

可以说，我是一名学者。

And it’s OK to be proud that I have a PhD in AI from Berkeley,

我从湾区这里的伯克利取得了 AI 的博士学位也是值得骄傲的。

right here in the Bay Area.

(Applause)

（掌声）

But there's something interesting in AI

但是相比其他科技创始人， 我发现了 AI 中一些有趣的事。

that I've noticed,

compared to other tech founders.

Other stereotypes, at least.

或者算是一些刻板印象。

A lot of us hold PhDs.

很多创始人都有博士学位。

I mean, quite a lot.

真的很多。

11 out of 24 speakers just at this conference

仅在本次大会上， 24 位演讲者中就有 11 位

have PhDs,

拥有博士学位，

and over a third are assistant,

超过三分之一是知名大学的 助理教授、副教授或正教授。

associate or full professors with major universities.

Only time will tell

只有时间才能证明

if this is a new trend of seeing academics in technology startups.

这是否是让学者进入 科技创业公司的新趋势。

But I got pretty curious to find out if this is common or new.

但是我很好奇 这是个常见还是新奇的现象。

And it turns out this is somewhat new.

事实证明，它是新奇的。

Only over a year ago,

就在一年多以前，

researchers at the University of Maryland

马里兰大学的研究人员发现，

found a 38 percent decline at the rate of startup formation

在过去 20 年里， 由美国博士成立初创公司的比例

or share of employment by US PhDs over the past 20 years.

或美国博士的就业比重 降低了 38%。

Yet our attendance here today

然而，我们今天齐聚在此，

and the trend in AI technology broadly

加上 AI 技术的总体趋势，

does not seem to correlate with this finding.

似乎并不能印证这个发现。

As I said, only time and more data will tell.

正如我所说，只有时间 和更多的数据才能证明一切。

In the meantime, my curiosity led me to another question:

与此同时，我的好奇心 让我想到了另一个问题：

What was the last major technology company founded by academics?

最近一家由学者创立的 大型科技公司是哪一家？

Google.

谷歌。

At Perplexity, we get accused of trying to kill Google a lot.

我们 Perplexity 总是被人谴责想杀死谷歌。

(Laughter)

（笑声）

But trust me, we're not really trying to kill things.

但是相信我， 我们没想杀死什么。

We are motivated about building things.

我们想要创造。

The cofounders of Google would probably say the same.

谷歌的联合创始人 可能也会说同样的话。

Let's hear from Larry Page.

我们来听听拉里·佩奇 （Larry Page）说过些什么。

An interview of his from the year 2000.

这是他在 2000 年的采访。

(Video) Larry Page: AI would be the ultimate version of Google.

（视频）拉里·佩奇： AI 会是谷歌的终极版本。

So if we had the ultimate search engine, it would understand everything on the web.

如果我们拥有终极搜索引擎， 它就能理解网络上的所有内容。

It would understand, you know,

它会理解你到底想要什么，

exactly what you wanted,

and it would give you the right thing.

给你提供正确的东西。

And that's obviously artificial intelligence.

这显然就是人工智能。

It would be able to answer any question, basically,

它能够回答任何问题，

because almost everything is on the web, right?

因为几乎所有东西都在网络上，对吧？

Aravind Srinivas: Think about that.

阿拉温德·斯里尼瓦斯 （Aravind Srinivas）：想想吧。

Artificial intelligence

2000 年的人工智能。

in the year 2000.

I was only six back then.

那时我只有六岁。

(Laughter)

（笑声）

There are a few things interesting about this interview.

这次采访有些有趣的点。

One, Larry did accurately predict the future of search

第一，拉里在将近 25 年前 确实准确地预测了搜索的未来。

almost 25 years ago.

The future of search is artificial intelligence.

搜索的未来是人工智能。

That’s why I’m here, and we’re going to talk more about it.

这就是为什么我来到了这里， 我们将进一步讨论这个问题。

Second, it's very interesting how a common theme

其次，有趣的是 这样的采访、这样的大会

in interviews like those or events like these

都有一个共同的主题， 就是思考未来。

is us thinking about the future.

What is the future of search?

搜索的未来是什么？

What is the future of technology?

科技的未来是什么？

What is the future of AI?

AI 的未来是什么？

I'm sure a lot of you have lots of thoughts about these questions.

我敢肯定，在座很多人 对这些问题都有很多想法。

In some sense, that is the purpose of technology:

从某种意义上说， 这就是科技的使命：

to keep us thinking and to keep us evolving.

让我们不断思考、不断进步。

But people like Larry, or people like you or people like me,

但是像拉里这样的人， 或者像你我这样的人，

we are not building technology in a vacuum.

我们不是在与世隔绝地开发技术。

We are building technology for us, the people.

我们在为我们、为人民开发技术。

We are the people.

我们就是人民。

So when we come here to think about what is the future of technology

当我们思考科技的未来 或 AI 的未来是什么时，

or what is the future of AI,

let's ask ourselves this question:

让我们问自己一个问题：

What is the future of us, the people?

我们，也就是人民的未来是什么？

I believe that AI will make us even more human.

我相信人工智能 将使我们更加人性化。

Socrates, the Greek philosopher,

希腊哲学家苏格拉底

is famous for saying that wisdom comes from realizing how little we know,

有一句名言： 智慧来自于自知无知，

or that progress can only be made by asking better questions.

还有进步只会来自于提出更好的问题。

The Socratic method is essentially the practice of relentless questioning.

苏格拉底法其实就是不断地提问。

Relentless questioning is something academics do all the time.

不断提问是学者们一直在做的事情。

It has been core to the progress of human intellect

在过去的 1000 年中， 它一直是人类智力进步的核心。

over the past 1,000 years.

Relentless questioning is also a practice that can be done

借助 AI 力量，不断提问

orders of magnitude better with the power of AI.

可以得到几个数量级的提升。

And by the way, relentless questioning is something south Indian parents do

顺便说一句， 不断提问就是南印度家长

when you tell them you're leaving a good university or a stable job

在你告诉他们你要退出 一所好大学或者辞掉稳定的工作，

to go join a startup.

加入一家初创公司时会做的。

(Laughter)

（笑声）

So, jokes aside, relentless questioning is something fundamentally human.

抛开玩笑，不断提问就是人类的本性。

The physicist David Deutsch has proposed that we humans

物理学家戴维·多伊奇 （David Deutsch）提出，我们人类

are the only species who have curiosity for what is already familiar.

是唯一对已经熟悉的事物 怀有好奇心的物种。

We can know so much about the stars above us

我们可以对我们头顶的星星

or the machines in front of us

或眼前的机器了解很多，

and yet continue to have more questions about them.

却还有更多有关它们的问题。

It seems like for humans, every answer leads to a new set of questions.

似乎对于人类来说，每一个答案 都会引出一批新的问题。

Questions that we haven't even asked before.

我们以前从未问过的问题。

That, to me, is what the future of technology should be about.

对我来说， 科技的未来就差不多是这样。

And it's also how Perplexity was born.

这也是 Perplexity 诞生的方式。

I was raised as an academic in the comforting arms of universities.

我在大学的温暖怀抱中 走在学者的成长轨迹上。

So when I actually entered the real world and tried to do my own company,

当我真正进入现实世界 并尝试建立自己的公司时，

I had an endless set of questions.

我遇到了无穷无尽的问题。

SPVs, SAFE notes, health insurance.

SPV、SAFE 协议、健康保险。

I needed to figure all these things out.

我得把这些都搞明白。

And all these required to do a lot of research

这些都需要做大量的研究，

and needed actual answers.

需要实际的答案。

And traditional search engines left me lost.

而传统的搜索引擎让我一头雾水。

There was a ton of information

有大量的信息，

and very little time to evaluate any of it.

没什么时间去评估每一条。

And neither did I have access to all of the experts on all these topics.

我也无法找到这些话题的专家。

So I was actually truly in a state of perplexity.

所以我真的处于 困惑（perplexity）之中。

So that's when I thought,

我当时想，

maybe I could have an AI do this for me.

也许我可以让 AI 为我做这件事。

Maybe I could go ask an AI all these questions,

也许我可以向 AI 问出这些问题，

if it was able to pull information from the web

如果它能从网上找出信息，

and answer all my questions.

回答我所有的问题。

So my cofounders and I came together,

我和我的联合创始人们聚在一起

and we built a Sackbot

创建了一个 Slackbot，

where we could just ask our own questions.

我们可以提出自己的问题。

Once we began using it is when we realized

我们开始用它的时候才意识到

what we built was much bigger than ourselves.

我们创造的东西比我们自己大得多。

For the first time,

这是我第一次

I had the ability to go ask whatever question I wanted about any topic,

可以就任何话题提出任何我想问的问题，

no matter my level of expertise in it,

无论我在这方面的专业水平如何，

and get a well-researched answer from the web.

然后从网络上得到一个 经过充分调研的答案。

And it's not just about an answer.

而且这不仅仅是一个答案。

It's an answer that I can actually trust.

这是我真正可以信赖的答案。

In this case, every answer in Perplexity

Perplexity 中的每个答案都会 以引用的形式配以网络上的出处，

comes with sources from the web in the form of citations,

just like academics cite their sources.

就像学者引用出处一样。

Now this is pretty powerful

这非常强大，

because trust is not unique to animals or humans,

因为信任不是动物或人类所独有的，

but it empowers us pretty differently.

但是它会给我们带来很不一样的提升。

In the case of humans, an answer you could trust

就人类而言， 你可以信任的答案

allows you to ask better follow-up questions.

可以让你提出更好的后续问题。

More questions lead to more knowledge.

更多的问题会带来更多的知识。

That's the point of ensuring that you could always get an answer

这就是为什么要保证 你一直能得到有理有据的答案。

with well-cited sources.

And in Perplexity, ever since the beginning,

在 Perplexity 中，从一开始，

every answer has always come with sources

每个答案都会配有出处， 让你提出更多问题。

that allows you to ask more questions.

In my case, once I ask questions about SAFE notes or insurance,

以我为例，一旦我问了 有关 SAFE 协议或保险的问题，

I ask more questions.

我就会问更多的问题。

What areas outside of insurance could I benefit from

如果我可以在保险之外的什么领域

获取更好的答案，会有帮助呢？

having access to better answers?

Who else in the world benefits from having access to better answers?

世界上还有谁 能从获取更好的答案中受益？

Now the answer is basically all of us.

答案就是几乎我们所有人。

Every single person benefits from having access to better answers.

每个人都能从获取更好的答案中受益。

This is such a profound shift in human history.

这是人类历史中的重大转变。

Until recently, if you wanted the best answers,

直到最近， 如果你想要最好的答案，

you had to be someone who could afford it.

你得是有负担能力的人。

You had to be someone who had access to the greatest minds in the world

你得是能够接触到 世界上最伟大的人才

or the best materials, libraries, expertise.

或最好的资料、 图书馆和专业知识的人。

And now that's changing.

而现在情况发生了变化。

If a major achievement of the internet

如果互联网的一项重大成就

was to give everyone access to all of the world's information,

是让每个人都能访问世界上所有的信息，

a major achievement of AI would be to give everyone access

那么 AI 的一项重大成就 将是让每个人都能获得

to all of the world's answers.

世界上所有的答案。

It doesn't matter if you're a Harvard professor

无论你是哈佛教授

or an underserved student in a developing nation,

还是发展中国家条件不好的学生，

we all get access to the same answers.

我们都能获取相同的答案。

With AI that keeps getting better and better

随着 AI 在回答我们 各种问题方面越来越好，

at answering all our questions,

the marginal cost of research is rapidly approaching zero.

做研究的边际成本 正在迅速接近于零。

In that new era of humanity that AI is powering,

在 AI 推动的人类新时代中，

knowledge does not really care about who you are, where you’re from

知识并不在乎你是谁、 你来自哪里、你能接触到什么人。

or who you have access to.

Rather, what matters is the next question you're going to ask.

相反，关键在于 你接下来会问出什么问题。

When all of the world's answers

当世界上所有人 都能获取世界上所有答案时，

are available to all of the world's people,

one can only wonder:

人们只会思考：

What will the best questions be,

最好的问题会是什么，

and how many such questions will get asked?

能问出多少这样的问题？

This is again where David Deutsch argues that human potential is infinite.

这也是为什么戴维·多伊奇 认为人类的潜力是无限的。

As long as we keep engaging in relentless questioning

只要我们不断地提问，

and keep asking an interesting set of questions,

不断提出一些有趣的问题，

the sky is the limit in terms of what we can actually learn.

我们能学到的东西就会直上云霄。

For example, humans are always curious.

比如，人类总是很好奇。

You can see that in babies.

你能从婴儿身上看出来。

Even before they learn to crawl,

甚至在婴儿学会爬行之前，

they're pretty curious about what's around them.

他们就对周围的事物非常好奇。

That's a natural trait for all of us.

这是我们所有人的天性。

Take an example of the technologies that we are building.

举一个我们正在开发的技术的例子。

In the case of the bot that became Perplexity.

就拿之后发展成 Perplexity 的 机器人为例。

Once I got answers to something like health insurance,

我获取了健康保险相关的答案之后，

I could ask an infinite set of new questions,

我就可以问出无数的新问题，

ranging from very pointed ones,

从非常有针对性的问题，

like, what are concrete ways to improve the health care insurance industry,

比如：有什么改善 医疗保险行业的具体方法，

to very broad ones,

到非常广泛的问题，

like, who else would benefit from having access to such a technology?

比如：还有谁会 因可以使用这项科技而获益？

It seems to a curious species

对于这么一个好奇的物种，

every question and answer that you get is a lead to the next set of questions,

你得到的每一个问题和答案 都是通向下一批问题的线索，

and spawns several paths of curiosity,

引向多条好奇心之路，

more than any one person can keep track of.

超出了任何人的追踪能力。

So when we are here to wonder about what is the future of technology,

当我们思考科技的未来是什么、

or what is the future of AI,

AI 的未来是什么时，

we are merely talking about the outputs,

我们只考虑了结果，

the outputs of a much bigger question:

一个更大问题的结果：

What is the future of human curiosity?

人类好奇心的未来是什么？

It is my strong belief that in an age

我坚信，在这个时代，

where AI gets better and better at answering all our questions,

AI 越来越擅长回答我们的各种问题，

this human quality that makes us so human will become even more essential.

这种让我们之所以为人的人类品质 变得越来越重要。

Our innate curiosity and our relentless questioning.

我们与生俱来的好奇心和不断的提问。

With all of the world's answers available to us,

我们有着全世界的答案，

the tools we use to ask our questions,

有着用于提问的工具，

and the stuff that we build using those answers,

还有我们用这些答案创造出的东西，

those to me are the future of our technology.

都是我们科技的未来。

And more importantly,

更重要的是，

that is the future of us, the future of humans.

这是我们的未来、人类的未来。

We are all curious,

我们都是好奇的，

and when we are curious, we want answers.

当我们感到好奇时， 我们想要答案。

We really do.

我们真的想要。

But what we really want are those answers

但是我们真正想要的是

that lead us to the next set of questions.

那些引导我们提出下一批问题的答案。

And I, for one, can't wait to see what you will ask next.

比如我就很好奇你接下来会问些什么。

Thank you.

谢谢。

(Applause)

（掌声）