人工智能的未来——今日筑梦，明日启程

欢迎来到我的节目。

So welcome to my show.

节目名为《明日今谈》，由Sheikhar Natarajan主持。

It's called Tomorrow Today with Sheikhar Natarajan.

我们将要探讨明天或未来会发生的事情，而这一切讨论都将在今天进行。

So we're going to be talking about what's going to happen tomorrow or into the future, and we're going to be talking about it today.

这个节目关注的不是AI技术本身如何进步，而是我们需要考虑的深层社会影响——我们必须找到应对之策，并思考如何为未来做好准备。

And it's a show about, not about AI and where AI is basically like improving technologically, but the deep societal impact that we need to consider, we need to have answers to, and how are we preparing ourselves for the future?

在我的首期节目中，很荣幸邀请到两位最理想的嘉宾共同开启这场对话。

And on my show, I have the two best people that I would love to actually launch this show with.

一位是Nadia Atwal，新闻界的女王，我最好的朋友，一位杰出的记者和发人深省的思想者——她能解答你能想到的任何问题。

One is Nadia Atwal, like she's the queen of journalism, my best friend, amazing journalist, and a thought provoking like person, like, you know, she has answers to every question that you could ever imagine.

另一位是我在地球另一端的好友Kate Hancock，她是两次TEDx演讲嘉宾，拥有非凡的灵魂，目前正领导着一个名为'AI向善'的美好事业。

And on the other side of the world is my other best friend, Kate Hancock, and basically Kate is a two time TEDx, TEDx talk speaker, amazing soul, like, and she's leading a beautiful effort called AI for Good.

欢迎两位的到来，现在我将正式启动本期节目。

Welcome both of you, and I'm here to basically kick off the show.

凯特，不如你来给我们讲讲目前的情况吧？

Kate, like, why don't you actually walk us through what is happening?

你现在在哪里？

Where are you today?

你知道的，你在世界的另一端，我多希望节目开播时你能和我们在一起，但你正在做比这更重要的事情。

Like, you know, like you are in another part of the world, I wish like you were here with us as we launched the show, but you're doing something even more important than anything else.

所以请给我们讲讲发生了什么。

So please actually educate us what is going on.

好的，当然。

Yes, absolutely.

我现在在奥地利维也纳参加TED AI大会，所有人都在讨论AI和媒体。

I'm currently in Vienna, Austria, attending the Ted AI and everyone's talking about AI and media.

所以谢卡尔，我很好奇，媒体在报道AI时经常在危言耸听和过度炒作之间摇摆。

So Shaker, I'm kind of curious, like the media often swing between fear mongering and hype when it comes to AI.

你认为目前世界对AI的认知有哪些误区或正确之处？

What do you think the world is getting wrong or right about AI right now?

我对你关于这个问题的想法很好奇。

I'm curious about your thought about that.

是啊。

Yeah.

你知道，这取决于你和谁交谈，情况会有所不同。

And like, you know, it's like, it's different, like, depending on who you talk to.

那些真正开发AI的人，他们对AI充满热情。

Like the people who are actually developing AI, like they're all about like AI.

他们会让你对AI深信不疑。

They will let you drink the Kool Aid about AI.

而那些真正沉迷于AI的人，却觉得它并不像宣传的那么神奇。

And the guys who are actually drinking the Kool Aid, like, feel that it's not like as cool as what it is.

这真的要看具体情况。

It really is contextual.

基本上，AI领域正在发生太多事情了。

And it's so much basically, there's so much going on, like, in the world of AI.

让我们直击问题核心，就像，知道吗，回到节目本身。

And let's get to the heart of the question itself, like, know, to the show itself.

那么，你们认为，从AI的角度来看，有哪些深层次的社会问题值得我们思考？

Like, what do you guys think, like, are the deep societal things that we should think about from an AI perspective?

我个人认为，目前AI导致的虚假信息是个大问题，人们感觉被强行推入某些新闻叙事中。

Well, I personally think there's a big issue right now with disinformation that is led by AI, where people just feel they're being pushed into certain news narratives.

就像，我为什么要关心这些？

And like, why would I care?

为什么这些东西要强加给我？

And why is this being pushed on me?

这是真的还是假的？

And is this real or is this not?

还有伪造的声音和视频。

And there's fake voices and fake videos.

这些都让人们感到非常困惑。

And that is what people are very confused by.

我们还需要教育他们如何识别这些。

And we need to educate them on that also how to spot.

我认为这基本上将成为新的一切艺术——如何辨别的艺术。

I think this is basically going to be the new art of everything is how to decipher.

我们该如何分辨什么是真实的，什么不是？

How can we tell what is real and what is not?

为什么某些信息会被推送给我们，这可能已经给了你第一个应该更仔细查看的提示。

And why would certain information being pushed on us may already give you the first indication that you should look closer.

凯特，你怎么看？

Kate, what do you think?

是的，完全同意。

Yeah, absolutely.

特别是世界上正在发生的事情，无论是政治还是当前新闻，你甚至不知道哪个是真的，而每个人看起来都那么完美。

Especially what's happening in the world, like whether it's politics or the current news, you don't even know which one is real and everyone looks so good.

对我们这些在AR或媒体行业的人来说，很容易察觉，但对普通人来说，他们得到的是错误信息，情绪被带动，但全都是假的。

And for us in the AR or in the media world, it's easy for us to detect, but for everyday people, they're getting the wrong information and their emotions get carried away, but they're all fake.

这确实是我们需要解决的问题。

And that's something that we really need to address.

我们该如何在错误信息中辨别真实信息？

How can we identify the real information between the wrong?

我是说，毕竟不是每个人都能识破这些。

I mean, again, not everyone can detect it.

人们常常为一些根本不是事实的事情焦虑。

Every people stress about something that's not even the truth.

是啊。

Yeah.

不，我来分享下我的个人经历。

No, I'll give you my own personal experience.

就像...你知道...我其实折腾过不少事情。

Like, you know, like, I've been playing around quite a bit.

小时候成长环境有限，你知道...我想当艺术家、想学音乐、想学绘画、想像父亲说的那样打板球——但他说'你会是个糟糕的板球手，因为我们没法贿赂这个体系'。

Growing up as a kid, like, you know, we didn't have access to many things, like, you know, I wanted to be an artist, I wanted to learn music, I wanted to learn, like, painting, I wanted to, like, play cricket, like my father said, like, you will make a horrible cricket player because we cannot bribe the system.

你可能是个有才华的人，但别以为你就该追求那条路。

You could, you may be a talented guy, but like, don't think you should pursue that.

所以我一直对艺术很着迷。

And so I was always fascinated by art.

我也是。

Same.

对吧？

Right?

于是我开始画画，因为这对我来说是一种释放压力的情感宣泄方式。后来我遇到了一个叫suno.ai的新应用，它就是个音乐创作工具。

And so then, like, I started painting, I, like, because it's been an emotional release for me from all the stress, and I ran into this new application called suno.ai, suno.ai, And it's just a music composition.

哦，我就是从那儿听到这些歌的。

Oh, that's where I'm getting the songs from.

每次创作完我都发给你，因为你是我最好的评论家。

Every time I compose, I'm sending it to you because you're my best critique.

不得不说，我从没见过像你这样风格跨度这么大的作曲家，从邦拉舞曲到嘻哈你都能驾驭。

I have to say, you know, I've never known a composer who has so much range because you're going from Bhangra to hip hop.

这非常有趣。

And it's very interesting.

但我认为这正是AI的魅力所在——它不在乎你来自哪里，只是根据市场承受能力给你产出，对吧？

But I think this is part of the appeal of AI that AI doesn't care where you're coming from, just gives you whatever the market will bear, right?

没错，最让我着迷的是那种多巴胺和皮质醇同时迸发的感觉，我基本上连续三天不眠不休地创作音乐。

Exactly, so like, so what was fascinating about, it was like dopamine, you know, cortisol, kicking in like all at the same time, and I was basically composing music, and I was literally in it for like three days nonstop, composing music.

之后事情开始变得模糊，因为我所做的只是不断输入指令：用这类乐器、要史诗感、要电影质感、按这个顺序演奏这些乐器、用这些分贝。

And then, like things started blurring for me after that, because I, all I was doing was giving it prompts to say, use these types of instruments, like make this epic, like make this cinematic, make this like, you know, have these instruments play at this sequence, use these decibels.

我完全不懂音乐，但它创作出来的作品却像模像样。

I had no clue of music, But it was composing music as though it was something really good.

后来我开始放给印度的司机听。

And then I started playing it to my driver in India.

当时正值盛大的甘尼许节，我们在街头组织庆典，人们随着音乐起舞。

And there was this, there was this huge festival where we organize Lord Ganesha on the streets and people dance to music.

我的司机说：'这音乐太棒了，比那些专业音乐人做的还要好，让我开始播放这个吧。'

And my driver said, hey, like, you know, this music is so good, better than what, like, basically all these guys have actually done that let me start, like, playing this.

艺术家是匿名的。

The artist was unknown.

对吧？

Right?

音乐是匿名的。

The music is unknown.

我是匿名的。

I am unknown.

我完全不知道。

I have no clue.

我不是贝多芬。

I'm not a Beethoven.

我完全不知道自己在做什么。

I have no clue of what I'm doing.

你知道怎么写音乐吗？

Do you know how to write music?

我没有，我甚至不知道如何写音乐，但美妙之处就在这里，对吧？

I haven't, I didn't even know how to write music, but here's the beautiful thing, right?

我甚至不知道如何写音乐。

I didn't even know how to write music.

当我真的为某些歌曲提供歌词时，系统显示这些内容受版权保护。

And when I actually gave some lyrics about some of the songs, it said it's copyright protected.

嗯。

Mhmm.

所以它拒绝了。

And so it rejected it.

我说，那我该怎么破解它呢？

I said, like, how do I hack it?

于是我的电脑大脑开始飞速运转。

So my computer brain started going off.

然后我说，让我在歌词里添加隐形文字吧。

And I said, like, let me add invisible text to the lyrics.

我把这个交给系统后，它直接接收并录制了音乐。

And I gave that to that system, and it simply took it and recorded the music.

你是否也认为，当你输入歌词时，系统会将歌词的情绪转化为曲调？

Do you also think that, was it also happening that when you put in the lyrics that the system was conversing the mood of the lyrics into tunes?

是的，它知道这本质上是一首宗教歌曲，知道有诗节结构，明白需要在适当处停顿和变换音乐调性，你也可以通过提示来调整。但最有趣的是这些歌词实际上是古老的经文。

Yes, so it knew that it is basically a devotional song, and it knew that there was stanzas, like, you know, it has to, like, you know, pause and like change, like, you know, the tone of the music, and you could prompt it as well, but what was the most interesting thing about all of this is these lists, like, you know, the lyrics were actually scriptures that were written a long time ago.

这是神圣的，只是受到保护而已。

It was sacred, It's just protected.

而我成功破解了它。

And I was able to hack into it.

虽然有人可能会觉得我的做法有些冒犯，但我确实做到了。现在想想，我一个对音乐一窍不通的人竟然能推广音乐。

There was, like, you know, like when someone is thinking about it, like there was scant disrespect to what I did, but I was able to do it, and now, like, think of me as a musician who knows nothing about music and can promote music.

你所说的这个原理可以应用到世界上任何与信息相关的内容创作领域，这才是最可怕的部分。

What you're talking about, like, can be applied to anything in the world as it relates to information, content creation, that's the scary part.

我担心人们会失去自我身份认同，在这个AI世界里任何人都可以为所欲为。

Like, I'm worried about people losing their own identity and anyone can do anything in this world of AI.

嗯，Shreykar本身就是一位艺术家。

Well, Shreykar is an artist by itself.

我从未见过有人如此亲力亲为地为每次演讲打造品牌和视频，所以我完全理解这一点。

I've never met someone who's very hands on into creating his branding and video for every talk, so I could see that.

但确实，我是说，AI的未来会让成本和服务趋近于零，对吧？

But yeah, I mean, the future of AI, like the cost and services will be dropped to zero, right?

就像你不需要太多天赋，只需要技术和一点热情，就能像Peter提到的那样创作出新音乐。

Like you don't need a lot of talents, you just need the technology and a little bit of zest from you and boom, you can create a new music like Peter mentioned.

所以我们原本以为，AI会简化大量工作，让我们能专注于创造性思维。

So what we thought, we all thought that AI will actually simplify a lot of work and let us use our creative brain.

但AI首先冲击的恰恰是所有创意领域。

The first thing that AI went after is all the creative things.

这就是讽刺之处。

That's the irony.

对吧？

Right?

就像，你知道的，我们所知的内容创作已经不复存在了。

Like, you know, content creation as we know it, is gone.

对吧？

Right?

就像，你知道的，我有个朋友的叔叔去世了，他和我们关系很亲近。

Like, you know, like one of my friend's uncle passed away, like, you know, he was very close to us.

他是个真正帮助我们离开美国的人，抱歉，是离开印度。

He was a guy who actually helped us get out of United States and, like, you know, get out of India, sorry.

基本上，你知道的，当他去世时，我想写封信。

And basically, you know, when he passed away, I wanted to write a note.

如果是发自内心的信，通常要花两小时。

The note usually takes two hours if it's heartfelt.

你需要回忆、思考、撰写、反复修改。

You want to reminisce, you want to think, you want to write, you want to rewrite it, you want to rewrite it.

但如果你是个男人，可能会花更长时间。

But you were a man, you'd take longer.

我知道，女性通常处理得更简洁。

I know, like women take it short.

但你会不断经历那种反复修改的过程，直到把它写对为止。

But you keep going through that notion of writing and rewriting all that stuff till you get it right.

现在你会怎么做？

Now, what do you do?

你会去ChatGPT那里说，你想写一封悼念信。

You go to chat GPT and say, like, you know, you want to write, like, a note.

嗯。

Mhmm.

然后三十秒内，你就能让它写得像马丁·路德·金的演讲那样充满感情。

And in thirty seconds, you can make it like, you can say, like, you know, make it as emotional as Martin Luther King's speech.

或者让它模仿甘地的演讲风格，或是其他人的风格，它会捕捉到那个人的语气，然后自动帮你写出来，这简直太疯狂了。

Or like, make it like a Gandhi speech or make it like someone else and it would pick up the tone of that person and basically start writing it for you and like, it's crazy.

然后在这种过程中，你开始逐渐丧失自己的思考能力。

And then you start losing your cognitive brain in all of this.

你的大脑去哪儿了？

Where did your brain go?

对吧？

Right?

就像，你完全忘记了语言，忘记了情感，忘记了呼吸的真实意义，忘记了悲伤是什么感觉，也忘记了快乐意味着什么。

Like, you just forgot, like, language, you forgot, like, emotions, you forgot, like, you know, what it means to actually breathe, you forgot, like, what it means to, like, have sorrow, you forgot what it means to be happy.

我们把所有这些都搁置一旁，然后基本上让系统接管了我们全部的生活。

We are parking all of those aside and then basically just let systems take over all of our lives.

不过你也可以从另一个角度来看，因为我们总是希望孩子们多阅读。

Well, you can also look at it from the other perspective because we always want our kids to read.

当我读到一些教堂的EPT歌词和笔记时，不得不说它们就像优秀的文学作品。

And when I read some church EPT lyrics and notes, I have to say they're like good literature.

我认为现在的孩子们应该多读这些，而不是刷YouTube和玩社交媒体。

And I think our kids today should read more of that instead of YouTubing and doing social media.

我觉得这已经会是一个很大的进步了。

I think it would already be an upgrade, you know.

所以它仍然有其价值。

So there's still a value to it.

也许某种程度上，它能不能当老师呢？

Maybe also kind of like, could it be a teacher?

就像我们说，好吧，我们输入这个，这就是我们原本的想法。

Where we're saying, okay, we're putting this in, this is what we had in mind.

然后我们得到的反馈比我们预想的还要好得多。

And we're getting something back that is far better than we even envision.

我们得到的启示是：哦，这实际上比我计划的还要好。

And there is a takeaway we're saying, Oh, this is actually better than I planned it.

所以，我正在记录这些。

So, I'm taking notes of that.

如果你不偷懒，这可能是你的处理方式。

If you're not lazy, that's the way you may go about it.

我是说，我就有过这样的经历，实际上我从ChatGPT那里学到了东西。

I mean, that has happened to me where actually I learned from ChatGPT.

我说，哦，你知道吗，这其实是个有趣的细微差别。

I said, oh, you know, this is actually an interesting nuance.

很有意思。

It's interesting.

英语是我的第二语言。

English is my second language.

所以我永远是个学生。

So I'm still an eternal student.

有时当我得到某些反馈时，某些语言表达，我总是会放进ChatGPT里让它变得幽默风趣，因为这就是我的个性。

And then sometimes when I'm getting certain things back, a certain language, and what I always put it in ChatGPT, make it funny and witty because that's my personality.

我想要这样的效果。

I want to have it.

所以我从不想听起来枯燥乏味只传递信息，我追求的是寓教于乐。

So I never want to sound boring and just informational when I always have infotainment.

所以我试着教会ChatGPT体现我的个性。

So I try to teach shared GPT to reflect my personality.

这还在进行中。

It's work in progress.

你觉得怎么样？

What do you think of that?

我认为如果你试图那样做，我们将会开始失去自主性，对吧？

I think we will begin to lose agency if you try to do that, right?

所以我觉得在学校里，人们需要更多地教会你如何使用AI，但同时也要运用自己的大脑。

And so I think like in schools, what people need to teach you to do more is use AI, but basically use your brain as well.

是的。

Yes.

不要，不要把你的大脑外包出去，因为一旦你把大脑外包出去，那么你就把人性也外包了，明白吗？人们试图用大脑进行推理能力，就像所有决定都是通过内心做出的，但头脑才是你的推理引擎，如果你不断委派它、外包它，那么你作为一个人就开始迷失了。

Don't, don't like outsource your brain, because once you outsource your brain, then you outsource your human aspect, right, like so, you know, people try to use their brain for reasoning capabilities to, like, you know, all decisions are made through your heart, but like it's just a head is your reasoning engine, and if you keep like delegating it, outsourcing it, then you begin to lose you as a person.

所以未来学校和教育的角色是真正教会人们自主性，比如，我该如何处理ChatGPT输出的内容？

So the role of schools and like education in the future is to really teach people agency, Like, how do I take what is coming out of ChatGPT?

我该如何辨别是非对错？

How do I, like, have the sense of right and wrong?

我该如何掌控面对信息时的自主权与选择权？

How do I control the agency and the choices I make with what is being presented?

我认为这正是我们需要更多思考的方向

I think that's where we should be thinking more in those

基本上，人类思维与人工智能的结合仍是黄金标准

Basically the combination of human thinking and AI is still the gold standard.

而凯特正处在这一领域的最前沿

And Kate is on the forefront of that.

这就是她创立全球眼理事会的原因，因为一切都关乎向善的人工智能

That's why she created the Global Eye Council, because it's all about AI for good.

凯特和我都有儿子

And Kate and I both have sons.

现在重点在于思考未来我们将如何教育子女

And now it's all about, you know, thinking of how we're going to educate our children in the future.

我发现了一所名为阿尔法的学校，其教学体系高度基于人工智能

And I came across a school named Alpha, that is very AI based.

他们说，哦，孩子们每天基本上只有两小时的学校时间。

And they are saying, Oh, we only have basically two hours of school per day for kids.

而且这是非常个性化的。

And it's very tailored to the individual.

在医学领域，这正是我们的发展方向。

Well, in medicine, that's where we are going.

所以未来整个医疗方法将是根据你的DNA来定制治疗方案。

So the whole medical approach of the future is we're going to tailor your medical treatment towards your DNA.

那么，为什么我们现在的教育还是'一刀切'的模式呢？

Now, why do we still have school that is one thing fits all?

不得不说，我觉得这实际上非常过时。

I find that actually very retro, I have to say.

当我看着现在的课程设置和学校的做法时，我在想：老兄，你真的认为四年后还会存在四年制的法律学位吗？

I don't think when I'm looking at curriculums right now and the school's approach, I'm like, dude, you really think there's going to be a four year law degree in four years from now?

我不这么认为。

I don't think so.

也许会有个为期六个月的学位课程，教你怎么指导AI执行法律工作。

Maybe a six month degree in how to instruct AI to perform law.

你觉得呢，凯特？

What do you think, Kate?

你对这个问题

What is your take on

怎么看？

all that?

是啊，完全同意。

Yeah, absolutely.

教育系统正在变革，对吧？

The education system is changing, right?

现在免费获取的信息资源比老师讲授的内容优质多了。

Like, free information are available way better than a teacher that's teaching you.

所以未来教师制定课程的方式将会采用全新的评估标准。

So it's going to be a whole new metrics of how teachers will create curriculum.

归根结底，关键在于产出，因为信息已经无处不在。

It's all about output at the end of the day, because information is out there.

我的意思是，每个人都能接入ChatGPT，如果你输入更好的信息，它给出的答案会比你的老师更聪明。

I mean, everyone can plug in ChatGPT and if you load better information, it will answer you smarter than your teacher.

所以这确实是一套全新的衡量标准，整个教育体系必须改变，教师需要适应，你无法阻止学生使用AI。

So it really is a whole new metrics and that's something that the whole education system has to change and teachers to adapt and you can't help students utilizing AI.

全球AI理事会最让我欣赏的是，你们实质上是在推动AI的民主化。

What I love about what the Global AI Council is doing is you're basically democratizing AI.

你说要让AI覆盖那些存在挑战的地区，也就是真正需要它的人们所在的地方。

You say, I want to make it available to places where it's a bit of a challenge, you know, so where people need it.

不是我们现在所处的特权曼哈顿地区，而是世界上那些迫切需要AI并能产生重大影响的其他地方。

And it's not the privileged Manhattan place where we are right now, but you know, other places in the world where it's very much needed and could make a big impact.

请简单介绍一下全球AI理事会目前的扩展情况，因为看起来你们正在全球各大洲全面推进。

Tell us about a little bit how Global AI Council is expanding right now, because it seems to be on all fronts and in all continents.

是的，完全正确。

Yes, absolutely.

我们正在欧洲这里进行扩展。

And we're here in Europe and expanding it.

我对Shaker公司很好奇，你们创造了'天使智能'这个术语。

And I'm curious about Shaker with your company, you know, coining the term angelic intelligence.

这是什么意思

What does that mean

对你来说？

to you?

因为我们对所有这些非常着迷，你看，就像思考人类一样，他们总是对未知充满好奇。

So, because we are so fascinated with all of the, see the, like, you think about humans, right, like, they've always been very fascinated about the unknown.

基本上，未知就像人们认为的，你知道，就像克里斯托弗·哥伦布那样，对吧？

And the unknown basically, like, people think like, you know, it's like Christopher Columbus, right?

他进入了未知的领域，那片未知的领土，并成功建立了一个国家。

He went into uncharted territories, the unknowns, and he was able to create a nation.

因此，追求未知的冒险行为往往能带来更大的回报。

And, and so that risk taking of going after the unknowns is much more of a payout.

因此大多数公司、风险投资基金等机构的运作都旨在解决未知、探索未知。

So most of the companies, venture capital funds and everything is geared to solving the unknown, seeking the unknown.

但

But the

关于这一点我有个问题。

I have a question about that one.

当你说大多数风险投资都在探索未知时，是否因为想法越新奇、越前卫，就越有可能吸引到愿意投资的人？

So when you say most of the venture capital is about seeking the unknown, is it because the most outlandish and the most new the idea, the bigger the chance that you are landing people who want to fund it?

是的。

Yes.

因为潜在回报巨大，而且在这种情境下风险与回报成正比。

So because they, because the payout is big and like, you know, the risk and reward is higher in that kind of a scenario.

但传统物流行业更像是...你知道的，属于基础支撑领域。

But like logistics as a traditional world is much more of like, you know, it's underbelly.

这个行业成本很高，非常依赖基础设施投入。

It's very, you know, it's very costly, it's very infrastructure intensive.

本质上，这是一个竞争残酷的行业，你知道，每个人都试图通过赚更多钱来打压对手。比如你看任何最后一公里公司，他们的策略就是把成本压得比联邦快递和UPS更低，这样才能获得市场主导地位。

It is basically, it's a cutthroat industry, like, you know, like everyone tries to drag the other guy down by making more money, right, like if you, if you look at any last mile company, their strategy is to lower the cost than what a FedEx and UPS does, and that's the way they can gain the market dominance.

所以我们生活在这样一个世界里，物流总是被视为成本函数，从未得到关注，唯一获得关注的时候就是人们发现问题的时候，比如新冠疫情。

So, so we live in this world where, like, logistics is always considered a cost function, never gets the attention, and the only time it gets the attention is when people, like, begin to see problems, like COVID.

供应链变得非常受关注，你知道，甚至白宫都有专人负责新冠物流。但投资并不高，因此我们错过了非常简单的事情，比如如何预测配送？

Supply chains become very popular, like, you know, there was a, there was a guy who was running like COVID even in White House, you know, so the investments are not as high and hence basically we miss out on the very simple things, like how do you predict deliveries?

如何实现精准配送？

How do you drive precision in delivery?

而且这些箱子没有身份标识。

And the boxes don't have identity.

它不知道这是医疗配送。

It doesn't know it is a medical delivery.

是啊，我刚经历

Yeah, I just had

一场糟糕的

a horrible

在这方面我有过糟糕的经历，因为我一直无法收到美国运通的新替换卡。

experience on that front because I could not get my new replacement cut of Amex forever.

但幸好我遇到了一位主管，她帮我解决了问题。

But I landed, thank God, with a supervisor, and she sorted it out.

虽然只是一个非常简单的错误，却导致事情被无限期拖延。

It was a very simple error that happened, but it delayed things forever and ever.

然后你把这个与我们通过AI能实现的其他事情相比较。

And then you compare that to other things we are capable of through AI.

你就会觉得这根本不匹配。

And you just simply think this doesn't match.

你看，这正是问题的关键所在。

This like, you know, so like, that's exactly the point.

认知层面的问题，对吧？

The cognitive, right?

我们能够解决那些非常复杂的问题，却忽略了那些实际上能救命的简单问题。

Like our ability to solve these very complex problems and then forget the very simple problems that actually is gonna be, like, lifesaving.

就像那个拿不到药的女孩的情况一样。

Like with this girl who couldn't get her medicine.

跟我们说说她发生了什么。

Tell us about what happened with her.

基本上，这是我一个很好的朋友，他在一家财富百强公司工作，有个八岁的女儿叫玛雅，她天生患有糖尿病。他们订购的处方药显示已发货，但始终没送到。结果她出现了心律失常，陷入休克状态，因为急需药物，她的心脏和大脑同时在争夺资源。

So basically, like, this is a great friend of mine, like, you know, he works for like Fortune one hundred company, he had an eight year old daughter, name was like Maya, and basically, like, you know, she was born like diabetic, like, you know, she was like, you know, and then like, they were looking, like, you know, were basically, this was on a prescription drug, came in, they, it said like it's out for delivery, and then like, it never showed up, then she basically had like arrhythmia, like you know, she went into like shock, and then like, you know, because like, you know, she needed the medication, her heart and her brain like both were competing at the same time.

短短三十六小时内，这个充满活力的美丽小女孩——他是我最好的朋友——

And literally within thirty six hours, this young soul, beautiful girl, full of energy, and this guy is like my bestie.

好吧。

Okay.

三十六小时后这女孩就离世了。

The girl just vanishes thirty six hours later.

这其实是个简单的问题。我们事后可以回顾，或许周末该直接去药房试试？谁能预料到所有这些事会同时急速恶化？她同时遭遇多种并发症：心脏问题、心律失常，还有糖尿病休克。

And it's a simple problem, like, you know, like, and gosh, like, you know, we can go back and run this in history and say, like, you know, like, should we have gone to, like, a medical store and tried to see or do this or do that because of weekend, and who would have expected that all of these sequence of events would just happen so fast and we would, you know, just lose a kid to multiple things all at the same time, because she's, she had multiple things going on at the same time, she had the heart problem, she had the arrhythmia, and like, you know, like, the diabetic shock.

太疯狂了。

It's crazy.

是啊，但归根结底，简单的问题还在于工作人员没有准时到达。

Yeah, but at the end of the day, the simple problem was also, staff didn't arrive on time.

准时

On

没错，正是。

time, exactly.

对。

Yeah.

没错，你看，你把药物放在外面，如果它需要冷藏，又放置太久，就会失效。

Right, like, you know, you leave medication outside and, like, you know, if it is, like, to be refrigerated and if it is left for too long, it is not effective.

哦对，它

Oh yeah, it

没错，这些都是常见问题，比如今天你去送杂货，对方是坐轮椅的老人，他们行动不便，对吧？

Right, so these are the common challenges, like, someone who's actually, like, groceries, like today when you go and deliver groceries and it is to an elderly person and they are, they don't have the ability, they're in wheelchair, right?

你会指望一个住在一楼坐轮椅的人自己去取杂货吗？

Do you expect someone in wheelchair who's actually living in the 1st Floor to go fetch the groceries themselves?

现有的配送系统并没有设计成能理解那个人需要这种照顾，对吧？

The delivery systems are not designed to understand that that person needs that care, right?

这些都是我们作为社会所忽视的普遍问题，而我还是要为那些考虑口罩问题的人点赞。

So these are all the common problems, like, you know, that we have ignored as a society, and yet we are like, and like, I give kudos to people who think about masks.

很棒，你知道，尽管去折腾吧。

Great, like, you know, go knock yourself out.

我常说，如果人们想去那里，我不想加入，我要留在地球上，你懂的。

I always say, if people want to go there, I don't want to join, I stay on Earth, you know.

但是，你知道

And, but like, you

去火星吧，但别带上我。

know, go to Mars, but without me.

但我们也需要解决普通人的问题。

But let's also solve the common man problem.

普通人的问题就是，你知道，我需要药品，需要日用品，需要高度可靠的这些东西，这些都是真正让我很困扰的问题。我在仓库工作过，你知道，我基本上在大型配送中心工作过，比如沃尔玛的配送中心。

Common man problem is like, you know, I need medicine, I need consumables, I need this with high kind of degree of reliability, and so these are the problems that like, you know, like really hurt me a lot, and I worked in warehouses, I worked like, you know, I worked in basically like mega distribution centers, like a distribution center, like a Walmart distribution center.

你彻底改变了像沃尔玛、美国鹰牌服饰和百事可乐这样的公司。

You revolutionized companies like Walmart and American Eagle, PepsiCo.

所以你的核心理念正是我所理解的，这对我来说很有道理。

So your whole thing is about what I'm understanding, and that makes sense to me.

你想做的本质上是将我们现有的配送模式进行集中化和简化。

What you want to do is basically centralize and simplify our existing model of delivery.

想想所有现有的最后一英里解决方案，对吧？

You know, so think about all the last mile solutions out there, right?

哪些是目前已知最有效的运作方式？

Like what are the best ones which work like that we know that works?

比如水电供应到你家，可能有六间卧室五间浴室，但不会对应11家公用事业公司。

Water, electricity, and in your house, may have six bedrooms and like five bathrooms, but you don't have 11 utility companies.

没错。

Exactly.

你只需要一家供应商提供服务。

You just have one who's delivering.

所以，在水上也是一样的道理。

So it's the same thing on, on, on, on, even on the water.

所有这些系统的设计方式都是：让我们把适用于千店模式的方法扩展到1.3亿个家庭门户，然后指望它能奏效。

So the, the way all these systems were designed was like, let's take what worked for Thousand Door stores and extend it to 130,000,000 doors called households and hope that it works.

但实际上行不通，对吧？你不能派八个人去追求效率，而是应该派一个人带着八件东西，这样更高效。

Well, it doesn't work that way, right, like, you cannot, like, dispatch eight people and try to be efficient, you dispatch one guy with eight things, that's more efficient.

因此每个系统都构建得非常低效，而我们却试图通过在这种低效上竞争来延续这个问题——无论是仓库、卡车还是司机，这些基础资源结构都存在，我们却在争夺同一批资源，同时不断推高价格。

And so every system was built, like, very inefficiently, and we're trying to perpetuate that problem by actually competing on that inefficiency where the underlying resource structure, whether it's a warehouse, a truck, a driver, they're all, like, they're all there and we're all competing for the same set of resources and we're just jacking up the prices at the same time.

你知道现在如果我是家小公司，送一个包裹的实际成本是多少吗？哪怕距离不到100英里，每个包裹也要15到17美元。

You know how much it costs to actually make a delivery if I'm a small company today, if I'm delivering a package, literally 100 and, like less than 100 miles, it's like 15 to $17 a package.

一件衣服的成本才多少？

What is the cost of a garment?

这太疯狂了，对吧？

It's crazy, right?

那么小型公司在这种环境下要怎么生存呢？

So like, how does a small company survive in this kind of a world?

这些都是影响深远的大问题，关乎创业、创新乃至社会繁荣，而人工智能可以发挥重要作用。但在解决这些问题的过程中，我根本发现的问题是——这也是我投身Angelic Intelligence的原因——我们建造的所有这些机器都只存在于实验室和公司董事会里。

So these are like, these are not like, these are big problems of consequence for entrepreneurship, for innovation, for the society to thrive and, and AI can play a big role, but while solving all of this, what I fundamentally found was the problem, and that's why I went after Angelic Intelligence, is that all of these machines that we have built were only built in labs and in corporate boards.

那些实验室缺乏人文关怀与人类思考。

Labs which were devoid of human touch and human thinking.

我们总是认为人类是系统中的漏洞，必须予以解决。

We always thought that human is a bug in the system that we've got to go solve for.

问题的本质不在于如何让人类成为超人。

It's not about basically saying, how do I make human a superhuman?

这种观念一直困扰着我。

And so that always bothered me.

当我儿子出生时，我说我绝不要留下一个把每个人都视为漏洞的世界。

And when my son was born, I said, I'm not going to freaking leave a world where every human is considered a bug.

因为我儿子总有一天也会被当成漏洞。

Because my son is gonna be considered a bug one of these days.

我绝不允许这样。

I don't wanna do that.

我真正想要构建的是一个世界，在那里每个人尽管有其局限性，都能成为超人，因为我们试图通过系统所做的一切，就是放大美好。

I wanna actually build a world where every human can be a superhuman despite their limitations, because all we are trying to do with systems is to amplify the goodness.

通过放大美好，你就能加速他们实现潜能的能力。

By amplifying the goodness, you amplify their ability to reach their potential faster.

当每个人都能更快地实现潜能，世界就会变得更美好，因为现在每个人都能赚更多、花更多，他们能够创造尊严，能够获得机会，拥有自主权。

When everyone reaches their potential faster, the world is a better place because everyone is now making more, spending more, they're able to create dignity, they're able to create access, they have agency.

这正是我们人类所代表的。

That's what we as humans stands for.

现在，凯特知道，我们正通过全球人工智能理事会站在这一领域的最前沿。

Now, there are Kate knows that, we are on the forefront on that here with the Global AI Council.

批评者说，包括那些AI教父中的许多人认为，大势已去。

Critics say, and those who were the godfather of AI, many say the ship has already sailed.

是的。

Yeah.

我们基本上处于一场竞赛中，游戏中有太多人不在乎AI的道德指南针。

We are basically in a race where there are too many people in the game who don't care about the moral compass of AI.

他们只在乎利润。

They just care about profits.

你对此持什么立场？

Where do you stand on that?

你怎么看？

What do you think?

你认为还有希望吗？

Do you think there's still hope?

所以

So

听着，我认为社会已经反复证明，当事情出错时，大自然自有其重置的方式，对吧？

listen, I think the society has proven over and over again that when things go wrong, nature has a way to reset it, right?

当你观察外界动向，看看有多少人对此担忧——比如皮尤报告，那些关于人们对AI看法的调查，有40%的人不知道该如何应对AI。

And when you read the tea leaves out there, of how many people who are concerned about it, if you like, let's take like the Pew report, let's take about like the reports about like what people think about, you know, AI, there are 40% of people who don't know how to react about AI.

好吗？

Okay?

有60%或55%的人认为人工智能是负面的。

There's 60% of people or 55% of people who think AI is bad.

有10%的人认为人工智能会带来真正的好处。

There are 10% of people who think AI is going to be really good.

只有10%。

Only 10%.

只有10%。

Only 10%.

凯特，你对这个怎么看？

Kate, how do you feel about that one?

而人们必须

And people are Gotta

扩大并倍增这10%的比例，

amplify and multiply the 10%,

对，首先。

right, first.

因此我认为将会有一股强大的力量来思考如何纠正错误的方向，对吧，而我坚信的是，我所相信和正在构建的，如果我在构建，我知道在这个宇宙中，我们都自认为是提出这些想法的孤独天才。

So I think there is going to be a strong force to kind of think about how do we right the ship from going wrong, right, and what I will believe is, what I'm believing and what I'm building, and I'm, if I'm building, I know that in this universe, we all think that we are the lone geniuses who has come up with ideas.

当你开始寻找非常相似的想法时，你会发现有成千上万的人和你想法完全一致，完全一致，明白吗。

When you go start searching for ideas which are very similar, you'll find that there are 10,000,000 people who are thinking just like you, just like you, okay.

执行方式可能大不相同，但他们很可能都在思考同一件事，对吧？

The style of execution is very different, but they're all probably thinking the same thing, right?

所以我确信就在此刻，当我们交谈时，有成千上万像我这样的人正对现状感到非常非常不安，他们可能正在为更美好的未来而努力。

So I'm sure at this time, as we speak, there are thousands of mine which are actually being very, very disturbed about what is going on, and they're probably building towards a better future.

我只能控制我能控制的部分。

And I can only control what I can control.

我能控制的就是我所构建的东西。

What I can control is what I build.

我正在构建的基本上是一个基于美德的LLM。

What I build is basically a virtue based LLM.

在这个基于美德的LLM中，我正在整合所有构成人性的要素，我称它们为数字天使。

In this virtue based LLM, what I'm building is taking all of the aspects of what makes us human, and I'm calling them digital angels.

一共有27个，对吧？

There are 27 of those, right?

这27位天使神奇地分成了七组。

Those 27 angels somehow magically fall into seven groups.

这七组正好对应我们身体的七个能量中心，也就是七个意识领域。

These seven groups are exactly like the seven centers of energy in our body and those are the seven realms of consciousness.

神奇的是，一切都完美契合。

Magically, it all falls.

为什么呢？

Why?

因为这在神学上是正确的，哲学上是正确的，也是宇宙的真理。

It's theologically right, philosophically right, and it is universe.

嗯，也许这就是为什么著名且备受尊敬的《商业影响者》杂志会授予你年度全球影响力大奖。

Well, that is why maybe, the very well known and highly respected magazine, The Business Influencer gave you the award for making global impact of the year.

顺便说一句，我认为这个奖项实至名归。

I think well deserved, by the way.

非常感谢。

Thank you so much.

你知道，我明白你是带着正确的意图来的。

You know, I know that you are coming with the right intent.

正如你多次提到的，你在企业界有着辉煌的职业生涯。

And as you said many times, you had a fantastic career in the corporate world.

你本可以过着舒适的生活，站在巅峰之巅。

You were on top of the top way, would have been a comfortable life.

然而，你却选择创办自己的公司，承担随之而来的所有风险、未知和不确定性，同时还要在这期间养育年幼的家庭。

Instead, you chose to start your own company with all the risks that it comes with, all the unknown, all the uncertainty, and raising a young family in the midst of it.

你对Orchestra未来一两年有什么展望？

What are you seeing for Orchestra in the next one, two years?

因为显然我们正处于一场相当激烈的竞争中。

Because obviously we are in a race that is pretty tight.

时间紧迫。

There's urgency.

至少媒体试图向我们暗示，我们正处于一场拯救人类的激烈竞赛中。

At least the media is trying to suggest to us we're in a very tight race to save humanity.

是的。

Yeah.

所以我想，正如我所说，我们能控制可控之事。

So I guess like, you know, like as I said, we can control what we can control.

我认为在未来两年内，我们已经开始涉足这个高度依赖人力却最不受重视的物流领域。

And I think like in the next two years, we already have begun to take the world, which is highly human, but least respected logistics.

如果我能将天使智能的概念应用于物流，那么我们就能将其推广到其他所有领域，因为它将成为一种普适性服务——无论是酒店业、医疗保健，还是任何涉及人类的职业领域。我们正在以截然不同的方式构建这套体系，而且我已经走访过许多国家，你我经常讨论这一点。

And if I can apply the concept of angelic intelligence to logistics, then we can, then we can actually take it everywhere else because it just becomes a service that is like, available to every sector, like hospitality, to like, you know, healthcare, and like any profession where humans are involved, we can take the same philosophy, and the way we are building it is fundamentally different, And I have gone out to many countries, you know, you and I keep talking about it.

我去过世界上很多地方，能明显感受到强烈的反差——我所构建的与世界其他地区正在发展的模式之间存在着巨大的恐惧鸿沟。

I've gone out to like, you know, many places in the world, and then there is, there is a sharp, like, there's a, there's a lot of fear and there's a sharp contrast between what I am building and what the rest of the world is building.

具体是什么反差？

What is the sharp contrast?

反差其实很简单。

The contrast is very simple.

当前的AI模型，将安全防护视为事后补救措施，对吧？

The current model, the AI models, like assumes safeguard as an afterthought, right?

所以你构建系统，遇到错误，发生事故，然后打个补丁，对吧？

So you build, you run into mistakes, you have accidents, then you put a patch, right?

这不是构建现代机器应有的方式。

And that's not how you build the modern day machines.

你要将美德作为原生属性来构建。

You build virtues as the native.

它是内置的。

It's built in.

这不是事后补救，而且在计算层面和问题思考方式上，都与传统模型的训练方法截然不同。

It's not an afterthought and computationally, and the way you think about the problem itself is very different than how the traditional models have been trained.

传统模型是人类给机器提供数据。

The traditional models is the humans gave the data to the machines.

机器接收数据后帮助人类获得更好的答案。

The machines took the data and is helping the humans get better answers.