Mark Zuckerberg: Meta, Facebook, Instagram, and the Metaverse
心理与人性技术与编程商业与创业音乐与艺术政治与社会
🤖
AI 智能总结
马克·扎克伯格谈Meta、Facebook与元宇宙
这是 Lex Fridman 与 Meta CEO 马克·扎克伯格的深度对话。扎克伯格坦诚讨论了 Facebook 的内容审核困境、元宇宙的愿景、对 AI 的投入，以及他作为一个父亲和公众人物的个人思考。
MetaFacebook元宇宙言论自由AI社交媒体领导力
马克·扎克伯格是 Meta（原 Facebook）的联合创始人兼 CEO，在哈佛大学宿舍创建了 Facebook，现在管理着拥有 Facebook、Instagram、WhatsApp 和 Threads 的全球最大社交媒体帝国。
📌 核心观点
- 言论自由与内容审核的困境：扎克伯格承认这是 Meta 面临的最难问题——如何在保护用户免受有害内容侵害的同时，不成为言论的仲裁者。他认为平台不应该决定什么是真相，但也不能对有害内容袖手旁观。
- 元宇宙愿景：扎克伯格解释了为什么他将公司更名为 Meta 并押注元宇宙——他认为下一个计算平台将是沉浸式的虚拟现实/增强现实，就像移动互联网取代桌面互联网一样，元宇宙将取代当前的互联网。
- 对 AI 的战略投入：扎克伯格认为 AI 是 Meta 最重要的技术投资，从推荐算法到内容审核再到 Llama 系列开源模型，AI 正在重塑 Meta 的每一个产品。
- 关于青少年和社交媒体：扎克伯格承认社交媒体对青少年心理健康的影响是一个严肃的问题，Meta 正在开发更多工具来保护未成年用户，但他也认为问题被过度简化了。
- 个人成长与领导力：扎克伯格分享了他从一个 20 岁的创业者成长为管理数万名员工的 CEO 的历程，他认为最重要的是保持学习心态和对使命的坚持。
✨ 金句摘录
扎克伯格：你不会建立这样一家公司，除非你相信人们表达自己是一件好事。
扎克伯格：我非常关心人们使用我们产品时的感受——我不想建立让人愤怒的产品。
扎克伯格：元宇宙将是下一个计算平台，就像移动互联网取代桌面互联网一样。
📋 章节目录
暂无章节信息
🔑 关键词
donstuffablesocialtryingcompanymetaversecontentmediadoinghumanbuildingspeechconversationrealtalkinghelppersongoinginteresting
💬 精彩语录
暂无语录
🎙️ 完整对话（2723 条）
Lex Fridman (00:00.000)
Let's talk about free speech and censorship.
我们来谈谈言论自由和审查制度。
Lex Fridman (00:02.560)
You don't build a company like this
你不会建立这样的公司
Lex Fridman (00:04.040)
unless you believe that people expressing themselves
除非你相信人们会表达自己
Lex Fridman (00:06.440)
is a good thing.
是一件好事。
Lex Fridman (00:07.280)
Let me ask you as a father,
作为一个父亲我想问你，
Mark Zuckerberg (00:08.640)
there's a weight heavy on you
你身上的担子很重
Lex Fridman (00:09.760)
that people get bullied on social networks.
人们在社交网络上受到欺凌。
Mark Zuckerberg (00:13.000)
I care a lot about how people feel
我非常关心人们的感受
Lex Fridman (00:14.440)
when they use our products
当他们使用我们的产品时
Lex Fridman (00:15.460)
and I don't want to build products that make people angry.
我不想制造让人生气的产品。
Lex Fridman (00:19.320)
Why do you think so many people dislike you?
你觉得为什么那么多人不喜欢你？
Mark Zuckerberg (00:23.440)
Some even hate you.
有些人甚至讨厌你。
Lex Fridman (00:25.580)
And how do you regain their trust and support?
如何重新获得他们的信任和支持？
Mark Zuckerberg (00:30.160)
The following is a conversation with Mark Zuckerberg,
以下是与马克·扎克伯格的对话，
Lex Fridman (00:32.720)
CEO of Facebook, now called Meta.
Facebook 首席执行官，现名为 Meta。
Mark Zuckerberg (00:36.760)
Please allow me to say a few words
请允许我说几句话
Lex Fridman (00:38.680)
about this conversation with Mark Zuckerberg,
关于与马克·扎克伯格的对话，
Mark Zuckerberg (00:41.300)
about social media,
关于社交媒体，
Lex Fridman (00:42.680)
and about what troubles me in the world today,
以及当今世界困扰我的事情，
Lex Fridman (00:45.440)
and what gives me hope.
是什么给了我希望。
Lex Fridman (00:47.760)
If this is not interesting to you,
Mark Zuckerberg (00:49.480)
I understand, please skip.
Lex Fridman (00:51.540)
I believe that at its best,
Mark Zuckerberg (00:54.620)
social media puts a mirror to humanity
Lex Fridman (00:57.800)
and reveals the full complexity of our world,
Mark Zuckerberg (01:01.060)
shining a light on the dark aspects of human nature
Lex Fridman (01:04.080)
and giving us hope, a way out,
Mark Zuckerberg (01:06.700)
through compassionate but tense chaos of conversation
Lex Fridman (01:09.780)
that eventually can turn into understanding,
Mark Zuckerberg (01:12.780)
friendship, and even love.
Lex Fridman (01:15.600)
But this is not simple.
Mark Zuckerberg (01:17.460)
Our world is not simple.
Lex Fridman (01:19.580)
It is full of human suffering.
Mark Zuckerberg (01:22.340)
I think about the hundreds of millions of people
Lex Fridman (01:24.460)
who are starving and who live in extreme poverty,
Mark Zuckerberg (01:28.460)
the one million people who take their own life every year,
Lex Fridman (01:31.600)
the 20 million people that attempt it,
Lex Fridman (01:33.940)
and the many, many more millions who suffer quietly
Lex Fridman (01:37.380)
in ways that numbers can never know.
Mark Zuckerberg (01:40.740)
I'm troubled by the cruelty and pain of war.
Lex Fridman (01:44.660)
Today, my heart goes out to the people of Ukraine.
Mark Zuckerberg (01:48.540)
My grandfather spilled his blood on this land,
Lex Fridman (01:52.220)
held the line as a machine gunner
Mark Zuckerberg (01:54.020)
against the Nazi invasion, surviving impossible odds.
Lex Fridman (01:59.180)
I am nothing without him.
Mark Zuckerberg (02:01.380)
His blood runs in my blood.
Lex Fridman (02:04.980)
My words are useless here.
Mark Zuckerberg (02:07.560)
I send my love.
Lex Fridman (02:09.020)
It's all I have.
Mark Zuckerberg (02:11.160)
I hope to travel to Russia and Ukraine soon.
Lex Fridman (02:14.140)
I will speak to citizens and leaders,
Mark Zuckerberg (02:16.780)
including Vladimir Putin.
Lex Fridman (02:19.980)
As I've said in the past, I don't care about access,
Mark Zuckerberg (02:22.700)
fame, money, or power, and I'm afraid of nothing.
Lex Fridman (02:27.660)
But I am who I am, and my goal in conversation
Mark Zuckerberg (02:31.100)
is to understand the human being before me,
Lex Fridman (02:33.460)
no matter who they are, no matter their position.
Lex Fridman (02:36.540)
And I do believe the line between good and evil
Lex Fridman (02:40.060)
runs through the heart of every man.
Lex Fridman (02:43.540)
So this is it.
Lex Fridman (02:45.380)
This is our world.
Mark Zuckerberg (02:47.420)
It is full of hate, violence, and destruction.
Lex Fridman (02:51.720)
But it is also full of love, beauty,
Lex Fridman (02:55.300)
and the insatiable desire to help each other.
Lex Fridman (02:59.140)
The people who run the social networks
Mark Zuckerberg (03:01.320)
that show this world, that show us to ourselves,
Lex Fridman (03:05.380)
have the greatest of responsibilities.
Mark Zuckerberg (03:08.500)
In a time of war, pandemic, atrocity,
Lex Fridman (03:11.740)
we turn to social networks to share real human insights
Lex Fridman (03:14.460)
and experiences, to organize protests and celebrations,
Lex Fridman (03:18.620)
to learn and to challenge our understanding of the world,
Mark Zuckerberg (03:21.940)
of our history and of our future,
Lex Fridman (03:24.300)
and above all, to be reminded of our common humanity.
Mark Zuckerberg (03:28.500)
When the social networks fail,
Lex Fridman (03:30.500)
they have the power to cause immense suffering.
Lex Fridman (03:33.640)
And when they succeed,
Lex Fridman (03:35.020)
they have the power to lessen that suffering.
Mark Zuckerberg (03:37.820)
This is hard.
Lex Fridman (03:39.340)
It's a responsibility, perhaps,
Mark Zuckerberg (03:41.080)
almost unlike any other in history.
Lex Fridman (03:44.080)
This podcast conversation attempts to understand the man
Lex Fridman (03:47.260)
and the company who take this responsibility on,
Lex Fridman (03:50.680)
where they fail and where they hope to succeed.
Mark Zuckerberg (03:54.260)
Mark Zuckerberg's feet are often held to the fire,
Lex Fridman (03:57.920)
as they should be, and this actually gives me hope.
Mark Zuckerberg (04:01.540)
The power of innovation and engineering,
Lex Fridman (04:03.900)
coupled with the freedom of speech
Mark Zuckerberg (04:05.560)
in the form of its highest ideal,
Lex Fridman (04:07.700)
I believe can solve any problem in the world.
Lex Fridman (04:11.140)
But that's just it, both are necessary,
Lex Fridman (04:14.780)
the engineer and the critic.
Mark Zuckerberg (04:17.620)
I believe that criticism is essential, but cynicism is not.
Lex Fridman (04:23.260)
And I worry that in our public discourse,
Mark Zuckerberg (04:25.660)
cynicism too easily masquerades as wisdom, as truth,
Lex Fridman (04:30.380)
becomes viral and takes over,
Lex Fridman (04:32.220)
and worse, suffocates the dreams of young minds
Lex Fridman (04:35.460)
who want to build solutions to the problems of the world.
Mark Zuckerberg (04:39.260)
We need to inspire those young minds.
Lex Fridman (04:41.540)
At least for me, they give me hope.
Lex Fridman (04:44.580)
And one small way I'm trying to contribute
Lex Fridman (04:47.160)
is to have honest conversations like these
Mark Zuckerberg (04:49.520)
that don't just ride the viral wave of cynicism,
Lex Fridman (04:53.260)
but seek to understand the failures
Lex Fridman (04:54.900)
and successes of the past, the problems before us,
Lex Fridman (04:57.940)
and the possible solutions
Mark Zuckerberg (04:59.540)
in this very complicated world of ours.
Lex Fridman (05:02.560)
I'm sure I will fail often,
Lex Fridman (05:05.820)
and I count on the critic to point it out when I do.
Lex Fridman (05:10.180)
But I ask for one thing,
Lex Fridman (05:12.540)
and that is to fuel the fire of optimism,
Lex Fridman (05:15.180)
especially in those who dream to build solutions,
Mark Zuckerberg (05:18.340)
because without that, we don't have a chance
Lex Fridman (05:21.780)
on this too fragile, tiny planet of ours.
Mark Zuckerberg (05:25.820)
This is the Lex Friedman podcast.
Lex Fridman (05:28.020)
To support it, please check out our sponsors
Mark Zuckerberg (05:30.260)
in the description.
Lex Fridman (05:31.620)
And now, dear friends, here's Mark Zuckerberg.
Lex Fridman (05:40.980)
Can you circle all the traffic lights, please?
Lex Fridman (05:53.880)
You actually did it.
Mark Zuckerberg (05:54.860)
That is very impressive performance.
Lex Fridman (05:56.820)
Okay, now we can initiate the interview procedure.
Mark Zuckerberg (06:00.020)
Is it possible that this conversation is happening
Lex Fridman (06:02.980)
inside a metaverse created by you,
Mark Zuckerberg (06:05.380)
by Meta many years from now,
Lex Fridman (06:07.180)
and we're doing a memory replay experience?
Mark Zuckerberg (06:10.180)
I don't know the answer to that.
Lex Fridman (06:11.260)
Then I'd be some computer construct
Lex Fridman (06:15.220)
and not the person who created that Meta company.
Lex Fridman (06:18.980)
But that would truly be Meta.
Mark Zuckerberg (06:21.380)
Right, so this could be somebody else
Lex Fridman (06:23.360)
using the Mark Zuckerberg avatar
Mark Zuckerberg (06:26.420)
who can do the Mark and the Lex conversation replay
Lex Fridman (06:29.200)
from four decades ago when Meta, it was first sort of.
Mark Zuckerberg (06:33.340)
I mean, it's not gonna be four decades
Lex Fridman (06:34.780)
before we have photorealistic avatars like this.
Lex Fridman (06:38.020)
So I think we're much closer to that.
Lex Fridman (06:40.100)
Well, that's something you talk about
Mark Zuckerberg (06:41.380)
is how passionate you are about the idea
Lex Fridman (06:43.500)
of the avatar representing who you are in the metaverse.
Lex Fridman (06:46.580)
So I do these podcasts in person.
Lex Fridman (06:51.100)
You know, I'm a stickler for that,
Mark Zuckerberg (06:52.260)
because there's a magic to the in person conversation.
Lex Fridman (06:55.780)
How long do you think it'll be before
Mark Zuckerberg (06:58.140)
you can have the same kind of magic in the metaverse,
Lex Fridman (07:00.620)
the same kind of intimacy in the chemistry,
Lex Fridman (07:02.620)
whatever the heck is there when we're talking in person?
Lex Fridman (07:06.060)
How difficult is it?
Lex Fridman (07:07.100)
How long before we have it in the metaverse?
Lex Fridman (07:10.340)
Well, I think this is like the key question, right?
Mark Zuckerberg (07:12.940)
Because the thing that's different about virtual
Lex Fridman (07:17.500)
and hopefully augmented reality
Mark Zuckerberg (07:19.140)
compared to all other forms of digital platforms before
Lex Fridman (07:22.400)
is this feeling of presence, right?
Mark Zuckerberg (07:24.420)
The feeling that you're right,
Lex Fridman (07:25.860)
that you're in an experience
Lex Fridman (07:27.000)
and that you're there with other people or in another place.
Lex Fridman (07:29.700)
And that's just different from all of the other screens
Lex Fridman (07:32.300)
that we have today, right?
Lex Fridman (07:33.900)
Phones, TVs, all the stuff.
Mark Zuckerberg (07:36.780)
They're trying to, in some cases, deliver experiences
Lex Fridman (07:39.500)
that feel high fidelity,
Lex Fridman (07:43.040)
but at no point do you actually feel like you're in it, right?
Lex Fridman (07:46.380)
At some level, your content is trying to sort of convince you
Mark Zuckerberg (07:49.660)
that this is a realistic thing that's happening,
Lex Fridman (07:51.980)
but all of the kind of subtle signals are telling you,
Mark Zuckerberg (07:54.860)
no, you're looking at a screen.
Lex Fridman (07:56.500)
So the question about how you develop these systems is like,
Lex Fridman (08:00.700)
what are all of the things that make the physical world
Lex Fridman (08:03.860)
all the different cues?
Lex Fridman (08:04.900)
So I think on visual presence and spatial audio,
Lex Fridman (08:13.080)
we're making reasonable progress.
Mark Zuckerberg (08:15.440)
Spatial audio makes a huge deal.
Lex Fridman (08:16.940)
I don't know if you've tried this experience,
Mark Zuckerberg (08:19.620)
workrooms that we launched where you have meetings.
Lex Fridman (08:21.960)
And I basically made a rule for all of the top,
Mark Zuckerberg (08:26.380)
you know, management folks at the company
Lex Fridman (08:27.760)
that they need to be doing standing meetings
Lex Fridman (08:29.520)
in workrooms already, right?
Lex Fridman (08:31.720)
I feel like we got to dog food this,
Mark Zuckerberg (08:33.260)
you know, this is how people are gonna work in the future.
Lex Fridman (08:35.760)
So we have to adopt this now.
Lex Fridman (08:38.800)
And there were already a lot of things
Lex Fridman (08:40.780)
that I think feel significantly better
Mark Zuckerberg (08:42.280)
than like typical Zoom meetings,
Lex Fridman (08:44.740)
even though the avatars are a lot lower fidelity.
Mark Zuckerberg (08:48.620)
You know, the idea that you have spatial audio,
Lex Fridman (08:50.780)
you're around a table in VR with people.
Mark Zuckerberg (08:53.940)
If someone's talking from over there,
Lex Fridman (08:55.260)
it sounds like it's talking from over there.
Mark Zuckerberg (08:56.860)
You can see, you know, the arm gestures
Lex Fridman (08:59.960)
and stuff feel more natural.
Mark Zuckerberg (09:01.760)
You can have side conversations,
Lex Fridman (09:03.080)
which is something that you can't really do in Zoom.
Mark Zuckerberg (09:04.860)
I mean, I guess you can text someone over,
Lex Fridman (09:06.940)
like out of band,
Lex Fridman (09:08.740)
but if you're actually sitting around a table with people,
Lex Fridman (09:12.940)
you know, you can lean over
Lex Fridman (09:14.200)
and whisper to the person next to you
Lex Fridman (09:15.580)
and like have a conversation that you can't,
Mark Zuckerberg (09:17.760)
you know, that you can't really do
Lex Fridman (09:19.100)
with in just video communication.
Lex Fridman (09:23.500)
So I think it's interesting in what ways
Lex Fridman (09:27.500)
some of these things already feel more real
Mark Zuckerberg (09:29.820)
than a lot of the technology that we have,
Lex Fridman (09:32.580)
even when the visual fidelity isn't quite there,
Lex Fridman (09:35.040)
but I think it'll get there over the next few years.
Lex Fridman (09:37.180)
Now, I mean, you were asking about comparing that
Mark Zuckerberg (09:38.740)
to the true physical world,
Lex Fridman (09:40.500)
not Zoom or something like that.
Lex Fridman (09:42.740)
And there, I mean, I think you have feelings
Lex Fridman (09:45.700)
of like temperature, you know, olfactory,
Mark Zuckerberg (09:50.700)
obviously touch, right, we're working on haptic gloves,
Lex Fridman (09:54.380)
you know, the sense that you wanna be able to,
Mark Zuckerberg (09:56.300)
you know, put your hands down
Lex Fridman (09:57.220)
and feel some pressure from the table.
Mark Zuckerberg (09:59.740)
You know, all of these things
Lex Fridman (10:00.580)
I think are gonna be really critical
Mark Zuckerberg (10:01.740)
to be able to keep up this illusion
Lex Fridman (10:04.660)
that you're in a world
Lex Fridman (10:06.820)
and that you're fully present in this world.
Lex Fridman (10:08.820)
But I don't know,
Mark Zuckerberg (10:09.740)
I think we're gonna have a lot of these building blocks
Lex Fridman (10:11.640)
within, you know, the next 10 years or so.
Lex Fridman (10:14.220)
And even before that, I think it's amazing
Lex Fridman (10:15.920)
how much you're just gonna be able to build with software
Mark Zuckerberg (10:18.140)
that sort of masks some of these things.
Lex Fridman (10:21.380)
I realize I'm going long,
Lex Fridman (10:22.780)
but I was told we have a few hours here.
Lex Fridman (10:25.300)
So it's a...
Mark Zuckerberg (10:26.140)
We're here for five to six hours.
Lex Fridman (10:27.220)
Yeah, so I mean, it's, look,
Mark Zuckerberg (10:28.740)
I mean, that's on the shorter end
Lex Fridman (10:30.460)
of the congressional testimonies I've done.
Lex Fridman (10:32.540)
But it's, but, you know, one of the things
Lex Fridman (10:36.300)
that we found with hand presence, right?
Lex Fridman (10:39.540)
So the earliest VR, you just have the headset
Lex Fridman (10:42.000)
and then, and that was cool, you could look around,
Mark Zuckerberg (10:44.440)
you feel like you're in a place,
Lex Fridman (10:45.380)
but you don't feel like you're really able to interact with it
Mark Zuckerberg (10:47.380)
until you have hands.
Lex Fridman (10:48.540)
And then there was this big question
Mark Zuckerberg (10:49.620)
where once you got hands,
Lex Fridman (10:51.340)
what's the right way to represent them?
Lex Fridman (10:53.460)
And initially, all of our assumptions was, okay,
Lex Fridman (10:58.560)
when I look down and see my hands in the physical world,
Mark Zuckerberg (11:00.420)
I see an arm and it's gonna be super weird
Lex Fridman (11:02.780)
if you see, you know, just your hand.
Lex Fridman (11:06.460)
But it turned out to not be the case
Lex Fridman (11:08.060)
because there's this issue with your arms,
Lex Fridman (11:09.860)
which is like, what's your elbow angle?
Lex Fridman (11:11.540)
And if the elbow angle that we're kind of interpolating
Mark Zuckerberg (11:14.780)
based on where your hand is and where your headset is
Lex Fridman (11:18.660)
actually isn't accurate,
Mark Zuckerberg (11:19.940)
it creates this very uncomfortable feeling
Lex Fridman (11:21.900)
where it's like, oh, like my arm is actually out like this,
Lex Fridman (11:24.440)
but it's like showing it in here.
Lex Fridman (11:25.940)
And that actually broke the feeling of presence a lot more.
Mark Zuckerberg (11:29.540)
Whereas it turns out that if you just show the hands
Lex Fridman (11:31.900)
and you don't show the arms,
Mark Zuckerberg (11:34.840)
it actually is fine for people.
Lex Fridman (11:36.220)
So I think that there's a bunch
Mark Zuckerberg (11:38.420)
of these interesting psychological cues
Lex Fridman (11:41.060)
where it'll be more about getting the right details right.
Lex Fridman (11:44.980)
And I think a lot of that will be possible
Lex Fridman (11:46.940)
even over a few year period or a five year period.
Lex Fridman (11:49.780)
And we won't need like every single thing to be solved
Lex Fridman (11:52.100)
to deliver this like full sense of presence.
Mark Zuckerberg (11:54.620)
Yeah, it's a fascinating psychology question
Lex Fridman (11:56.500)
of what is the essence
Lex Fridman (11:59.960)
that makes in person conversation special?
Lex Fridman (12:04.260)
It's like emojis are able to convey emotion really well,
Mark Zuckerberg (12:08.020)
even though they're obviously not photorealistic.
Lex Fridman (12:10.580)
And so in that same way, Jessica, you're saying,
Mark Zuckerberg (12:12.460)
just showing the hands is able
Lex Fridman (12:14.260)
to create a comfortable expression with your hands.
Lex Fridman (12:18.180)
So I wonder what that is.
Lex Fridman (12:19.520)
People in the world wars used to write letters
Lex Fridman (12:21.920)
and you can fall in love with just writing letters.
Lex Fridman (12:24.380)
You don't need to see each other in person.
Mark Zuckerberg (12:26.640)
You can convey emotion.
Lex Fridman (12:27.700)
You can be depth of experience with just words.
Lex Fridman (12:32.740)
So that's, I think, a fascinating place
Lex Fridman (12:35.740)
to explore psychology of like,
Lex Fridman (12:37.460)
how do you find that intimacy?
Lex Fridman (12:39.220)
Yeah, and the way that I come to all of this stuff is,
Mark Zuckerberg (12:42.700)
I basically studied psychology and computer science.
Lex Fridman (12:45.060)
So all of the work that I do
Mark Zuckerberg (12:47.900)
is sort of at the intersection of those things.
Lex Fridman (12:49.900)
I think most of the other big tech companies
Mark Zuckerberg (12:52.180)
are building technology for you to interact with.
Lex Fridman (12:55.060)
What I care about is building technology
Mark Zuckerberg (12:56.680)
to help people interact with each other.
Lex Fridman (12:58.100)
So I think it's a somewhat different approach
Mark Zuckerberg (12:59.900)
than most of the other tech entrepreneurs
Lex Fridman (13:02.300)
and big companies come at this from.
Lex Fridman (13:04.340)
And a lot of the lessons
Lex Fridman (13:08.980)
in terms of how I think about designing products
Lex Fridman (13:10.980)
come from some just basic elements of psychology, right?
Lex Fridman (13:15.980)
In terms of our brains,
Mark Zuckerberg (13:19.060)
you can compare it to the brains of other animals.
Lex Fridman (13:22.080)
We're very wired to specific things, facial expressions.
Lex Fridman (13:25.560)
I mean, we're very visual, right?
Lex Fridman (13:28.200)
So compared to other animals,
Mark Zuckerberg (13:29.340)
I mean, that's clearly the main sense
Lex Fridman (13:31.820)
that most people have.
Lex Fridman (13:32.980)
But there's a whole part of your brain
Lex Fridman (13:35.260)
that's just kind of focused on reading facial cues.
Lex Fridman (13:38.560)
So when we're designing the next version of Quest
Lex Fridman (13:42.180)
or the VR headset, a big focus for us is face tracking
Lex Fridman (13:45.820)
and basically eye tracking so you can make eye contact,
Lex Fridman (13:48.860)
which again, isn't really something
Mark Zuckerberg (13:50.100)
that you can do over a video conference.
Lex Fridman (13:51.500)
It's sort of amazing how far video conferencing
Lex Fridman (13:55.500)
has gotten without the ability to make eye contact, right?
Lex Fridman (13:58.620)
It's sort of a bizarre thing if you think about it.
Mark Zuckerberg (14:00.540)
You're looking at someone's face,
Lex Fridman (14:03.140)
sometimes for an hour when you're in a meeting
Lex Fridman (14:05.620)
and you looking at their eyes to them
Lex Fridman (14:08.960)
doesn't look like you're looking at their eyes.
Mark Zuckerberg (14:11.780)
You're always looking past each other, I guess.
Lex Fridman (14:15.020)
I guess you're right.
Mark Zuckerberg (14:15.860)
You're not sending that signal.
Lex Fridman (14:16.680)
Well, you're trying to.
Mark Zuckerberg (14:17.520)
Right, you're trying to.
Lex Fridman (14:18.360)
A lot of times, or at least I find myself,
Mark Zuckerberg (14:19.760)
I'm trying to look into the other person's eyes.
Lex Fridman (14:21.420)
But they don't feel like you're looking to their eyes.
Lex Fridman (14:23.060)
So then the question is,
Lex Fridman (14:23.900)
all right, am I supposed to look at the camera
Lex Fridman (14:25.220)
so that way you can have a sensation
Lex Fridman (14:27.800)
that I'm looking at you?
Mark Zuckerberg (14:28.640)
I think that that's an interesting question.
Lex Fridman (14:30.140)
And then with VR today,
Mark Zuckerberg (14:33.820)
even without eye tracking
Lex Fridman (14:35.660)
and knowing what your eyes are actually looking at,
Lex Fridman (14:37.500)
you can fake it reasonably well, right?
Lex Fridman (14:39.380)
So you can look at where the head pose is.
Lex Fridman (14:42.240)
And if it looks like I'm kind of looking
Lex Fridman (14:43.720)
in your general direction,
Mark Zuckerberg (14:44.680)
then you can sort of assume
Lex Fridman (14:46.500)
that maybe there's some eye contact intended
Lex Fridman (14:48.640)
and you can do it in a way where it's like,
Lex Fridman (14:50.740)
okay, maybe it's not a fixated stare,
Lex Fridman (14:54.300)
but it's somewhat natural.
Lex Fridman (14:56.880)
But once you have actual eye tracking,
Mark Zuckerberg (14:58.740)
you can do it for real.
Lex Fridman (15:00.180)
And I think that that's really important stuff.
Lex Fridman (15:02.140)
So when I think about Meta's contribution to this field,
Lex Fridman (15:05.300)
I have to say it's not clear to me
Mark Zuckerberg (15:06.640)
that any of the other companies
Lex Fridman (15:08.700)
that are focused on the Metaverse
Mark Zuckerberg (15:11.180)
or on virtual and augmented reality
Lex Fridman (15:13.340)
are gonna prioritize putting these features in the hardware
Lex Fridman (15:15.820)
because like everything, they're trade offs, right?
Lex Fridman (15:18.260)
I mean, it adds some weight to the device.
Mark Zuckerberg (15:21.500)
Maybe it adds some thickness.
Lex Fridman (15:22.740)
You could totally see another company taking the approach
Mark Zuckerberg (15:24.840)
of let's just make the lightest and thinnest thing possible.
Lex Fridman (15:27.600)
But I want us to design the most human thing possible
Mark Zuckerberg (15:31.380)
that creates the richest sense of presence
Lex Fridman (15:33.340)
and cause so much of human emotion and expression
Mark Zuckerberg (15:37.900)
comes from these like micro movements.
Lex Fridman (15:39.500)
If I like move my eyebrow millimeter,
Mark Zuckerberg (15:41.800)
you will notice and that like means something.
Lex Fridman (15:44.640)
So the fact that we're losing these signals
Lex Fridman (15:46.840)
and a lot of communication I think is a loss.
Lex Fridman (15:49.940)
So it's not like, okay, there's one feature
Lex Fridman (15:51.700)
and you add this, then it all of a sudden
Lex Fridman (15:53.300)
is gonna feel like we have real presence.
Mark Zuckerberg (15:55.100)
You can sort of look at how the human brain works
Lex Fridman (15:57.820)
and how we express and kind of read emotions
Lex Fridman (16:01.820)
and you can just build a roadmap of that,
Lex Fridman (16:04.700)
of just what are the most important things
Mark Zuckerberg (16:06.460)
to try to unlock over a five to 10 year period
Lex Fridman (16:08.520)
and just try to make the experience
Mark Zuckerberg (16:10.040)
more and more human and social.
Lex Fridman (16:12.780)
When do you think would be a moment,
Mark Zuckerberg (16:16.640)
like a singularity moment for the Metaverse
Lex Fridman (16:19.340)
where there's a lot of ways to ask this question,
Lex Fridman (16:22.300)
but people will have many or most
Lex Fridman (16:26.580)
of their meaningful experiences
Mark Zuckerberg (16:28.820)
in the Metaverse versus the real world.
Lex Fridman (16:31.340)
And actually it's interesting to think about
Mark Zuckerberg (16:33.060)
the fact that a lot of people are having
Lex Fridman (16:35.560)
the most important moments of their life
Mark Zuckerberg (16:37.280)
happen in the digital sphere,
Lex Fridman (16:39.100)
especially not during COVID,
Mark Zuckerberg (16:41.680)
like even falling in love or meeting friends
Lex Fridman (16:45.060)
or getting excited about stuff
Mark Zuckerberg (16:46.420)
that is happening on the 2D digital plane.
Lex Fridman (16:49.660)
When do you think the Metaverse
Mark Zuckerberg (16:50.860)
will provide those experiences for a large number,
Lex Fridman (16:54.100)
like a majority of the population?
Mark Zuckerberg (16:54.940)
Yeah, I think it's a really good question.
Lex Fridman (16:57.240)
There was someone, I read this piece
Mark Zuckerberg (17:00.260)
that framed this as a lot of people think
Lex Fridman (17:03.740)
that the Metaverse is about a place,
Lex Fridman (17:06.040)
but one definition of this is it's about a time
Lex Fridman (17:10.380)
when basically immersive digital worlds
Mark Zuckerberg (17:12.900)
become the primary way that we live our lives
Lex Fridman (17:17.100)
and spend our time.
Mark Zuckerberg (17:18.720)
I think that that's a reasonable construct.
Lex Fridman (17:20.160)
And from that perspective,
Mark Zuckerberg (17:21.900)
I think you also just wanna look at this as a continuation
Lex Fridman (17:25.540)
because it's not like, okay,
Mark Zuckerberg (17:27.140)
we are building digital worlds,
Lex Fridman (17:28.940)
but we don't have that today.
Mark Zuckerberg (17:29.820)
I think you and I probably already live
Lex Fridman (17:32.340)
a very large part of our life in digital worlds.
Mark Zuckerberg (17:34.660)
They're just not 3D immersive virtual reality,
Lex Fridman (17:37.260)
but I do a lot of meetings over video
Mark Zuckerberg (17:39.820)
or I spend a lot of time writing things over email
Lex Fridman (17:42.300)
or WhatsApp or whatever.
Lex Fridman (17:44.540)
So what is it gonna take to get there
Lex Fridman (17:46.060)
for kind of the immersive presence version of this,
Mark Zuckerberg (17:48.700)
which I think is what you're asking.
Lex Fridman (17:51.060)
And for that, I think that there's just a bunch
Mark Zuckerberg (17:52.940)
of different use cases.
Lex Fridman (17:55.640)
And I think when you're building technology,
Mark Zuckerberg (18:00.460)
I think a lot of it is just you're managing this duality
Lex Fridman (18:05.820)
where on the one hand,
Mark Zuckerberg (18:06.980)
you wanna build these elegant things that can scale
Lex Fridman (18:10.100)
and have billions of people use them
Lex Fridman (18:12.140)
and get value from them.
Lex Fridman (18:13.340)
And then on the other hand,
Mark Zuckerberg (18:14.540)
you're fighting this kind of ground game
Lex Fridman (18:17.020)
where there are just a lot of different use cases
Lex Fridman (18:19.680)
and people do different things
Lex Fridman (18:20.880)
and you wanna be able to unlock them.
Lex Fridman (18:22.260)
So the first ones that we basically went after
Lex Fridman (18:25.940)
were gaming with Quest and social experiences.
Lex Fridman (18:30.340)
And it goes back to when we started working
Lex Fridman (18:32.540)
on virtual reality.
Mark Zuckerberg (18:33.380)
My theory at the time was basically
Lex Fridman (18:37.400)
people thought about it as gaming,
Lex Fridman (18:39.440)
but if you look at all computing platforms up to that point,
Lex Fridman (18:44.160)
gaming is a huge part, it was a huge part of PCs,
Mark Zuckerberg (18:47.440)
it was a huge part of mobile,
Lex Fridman (18:49.460)
but it was also very decentralized.
Mark Zuckerberg (18:51.960)
There wasn't, for the most part,
Lex Fridman (18:54.240)
one or two gaming companies.
Mark Zuckerberg (18:55.660)
There were a lot of gaming companies
Lex Fridman (18:57.440)
and gaming is somewhat hits based.
Mark Zuckerberg (18:58.700)
I mean, we're getting some games that have more longevity,
Lex Fridman (19:01.440)
but in general, there were a lot of different games
Mark Zuckerberg (19:05.340)
out there.
Lex Fridman (19:06.560)
But on PC and on mobile,
Mark Zuckerberg (19:10.740)
the companies that focused on communication
Lex Fridman (19:13.700)
and social interaction,
Mark Zuckerberg (19:15.100)
there tended to be a smaller number of those
Lex Fridman (19:17.260)
and that ended up being just as important of a thing
Mark Zuckerberg (19:19.160)
as all of the games that you did combined.
Lex Fridman (19:21.580)
I think productivity is another area.
Mark Zuckerberg (19:23.140)
That's obviously something
Lex Fridman (19:23.980)
that we've historically been less focused on,
Lex Fridman (19:26.020)
but I think it's gonna be really important for us.
Lex Fridman (19:27.220)
With workroom, do you mean productivity
Lex Fridman (19:29.580)
in the collaborative aspect?
Lex Fridman (19:30.860)
Yeah, I think that there's a workroom's aspect of this,
Mark Zuckerberg (19:34.360)
like a meeting aspect,
Lex Fridman (19:35.360)
and then I think that there's like a Word, Excel,
Mark Zuckerberg (19:39.500)
productivity, either you're working or coding
Lex Fridman (19:42.940)
or knowledge work as opposed to just meetings.
Lex Fridman (19:46.760)
So you can kind of go through all these different use cases.
Lex Fridman (19:49.620)
Gaming, I think we're well on our way.
Mark Zuckerberg (19:51.280)
Social, I think we're just the kind of preeminent company
Lex Fridman (19:56.080)
that focuses on this.
Lex Fridman (19:57.040)
And I think that that's already on Quest becoming the,
Lex Fridman (1:00:01.640)
if you really care about what you're doing.
Lex Fridman (1:00:04.240)
And yeah, so, I mean, it's not great, but like,
Lex Fridman (1:00:08.000)
but look, I think that at some level,
Mark Zuckerberg (1:00:11.560)
whether 25% of people dislike you
Lex Fridman (1:00:14.960)
or 75% of people dislike you,
Mark Zuckerberg (1:00:18.080)
your experience as a public figure is gonna be
Lex Fridman (1:00:21.040)
that there's a lot of people who dislike you, right?
Mark Zuckerberg (1:00:23.360)
So, I actually am not sure how different it is.
Lex Fridman (1:00:28.680)
You know, certainly, you know,
Mark Zuckerberg (1:00:31.200)
the country's gotten more polarized
Lex Fridman (1:00:32.760)
and we in particular have gotten, you know,
Mark Zuckerberg (1:00:35.040)
more controversial over the last five or years or so.
Lex Fridman (1:00:39.080)
But, I don't know, I kind of think like as a public figure
Lex Fridman (1:00:45.240)
and leader of one of these enterprises.
Lex Fridman (1:00:48.560)
Comes with the job.
Mark Zuckerberg (1:00:49.400)
Yeah, part of what you do is like,
Lex Fridman (1:00:51.440)
and look, the answer can't just be ignore it, right?
Mark Zuckerberg (1:00:54.640)
Because like a huge part of the job
Lex Fridman (1:00:56.680)
is like you need to be getting feedback
Lex Fridman (1:00:58.200)
and internalizing feedback on how you can do better.
Lex Fridman (1:01:00.760)
But I think increasingly what you need to do
Mark Zuckerberg (1:01:02.520)
is be able to figure out, you know,
Lex Fridman (1:01:04.560)
who are the kind of good faith critics
Mark Zuckerberg (1:01:08.000)
who are criticizing you because
Lex Fridman (1:01:10.640)
they're trying to help you do a better job
Mark Zuckerberg (1:01:12.520)
rather than tear you down.
Lex Fridman (1:01:13.960)
And those are the people I just think you have to cherish
Lex Fridman (1:01:16.440)
and like, and listen very closely
Lex Fridman (1:01:19.120)
to the things that they're saying,
Mark Zuckerberg (1:01:20.280)
because, you know, I think it's just as dangerous
Lex Fridman (1:01:23.040)
to tune out everyone who says anything negative
Lex Fridman (1:01:26.840)
and just listen to the people who are kind of positive
Lex Fridman (1:01:29.320)
and support you, you know,
Mark Zuckerberg (1:01:31.240)
as it would be psychologically to pay attention
Lex Fridman (1:01:33.760)
trying to make people who are never gonna like you like you.
Lex Fridman (1:01:36.600)
So I think that that's just kind of a dance
Lex Fridman (1:01:38.880)
that people have to do.
Lex Fridman (1:01:40.080)
But I mean, I, you know,
Lex Fridman (1:01:41.760)
so you kind of develop more of a feel for like,
Mark Zuckerberg (1:01:44.720)
who actually is trying to accomplish
Lex Fridman (1:01:46.280)
the same types of things in the world
Lex Fridman (1:01:48.400)
and who has different ideas about how to do that
Lex Fridman (1:01:51.440)
and how can I learn from those people?
Lex Fridman (1:01:52.880)
And like, yeah, we get stuff wrong.
Lex Fridman (1:01:54.800)
And when the people whose opinions I respect
Mark Zuckerberg (1:01:57.760)
call me out on getting stuff wrong,
Lex Fridman (1:01:59.840)
that hurts and makes me wanna do better.
Lex Fridman (1:02:02.120)
But I think at this point, I'm pretty tuned to just,
Lex Fridman (1:02:04.680)
all right, if someone, if I know they're,
Mark Zuckerberg (1:02:06.360)
they're kind of like operating in bad faith
Lex Fridman (1:02:08.080)
and they're not really trying to help,
Mark Zuckerberg (1:02:10.800)
then, you know, I don't know, it's not, it's, it doesn't,
Lex Fridman (1:02:13.000)
you know, I think over time,
Mark Zuckerberg (1:02:13.840)
it just doesn't bother you that much.
Lex Fridman (1:02:15.280)
But you are surrounded by people that believe in the mission
Mark Zuckerberg (1:02:18.720)
that love you.
Lex Fridman (1:02:21.240)
Are there friends or colleagues in your inner circle
Mark Zuckerberg (1:02:23.600)
you trust that call you out on your bullshit
Lex Fridman (1:02:26.560)
whenever your thinking may be misguided
Lex Fridman (1:02:28.600)
as it is for leaders at times?
Lex Fridman (1:02:30.920)
I think we have a famously open company culture
Mark Zuckerberg (1:02:34.840)
where we sort of encourage that kind of dissent internally,
Lex Fridman (1:02:39.440)
which is, you know, why there's so much material
Mark Zuckerberg (1:02:41.760)
internally that can leak out
Lex Fridman (1:02:43.120)
with people sort of disagreeing
Mark Zuckerberg (1:02:44.520)
is because that's sort of the culture.
Lex Fridman (1:02:47.480)
You know, our management team, I think it's a lot of people,
Mark Zuckerberg (1:02:50.200)
you know, there are some newer folks who come in,
Lex Fridman (1:02:52.000)
there are some folks who've kind of been there for a while,
Lex Fridman (1:02:54.800)
but there's a very high level of trust.
Lex Fridman (1:02:56.840)
And I would say it is a relatively confrontational
Mark Zuckerberg (1:02:59.840)
group of people.
Lex Fridman (1:03:01.080)
And my friends and family, I think, will push me on this.
Lex Fridman (1:03:04.560)
But look, it's not just,
Lex Fridman (1:03:06.400)
but I think you need some diversity, right?
Mark Zuckerberg (1:03:09.280)
It can't just be, you know,
Lex Fridman (1:03:12.080)
people who are your friends and family.
Mark Zuckerberg (1:03:13.760)
It's also, you know, I mean, there are journalists
Lex Fridman (1:03:16.920)
or analysts or, you know,
Mark Zuckerberg (1:03:19.520)
peer executives at other companies
Lex Fridman (1:03:23.240)
or, you know, other people who sort of are insightful
Mark Zuckerberg (1:03:27.600)
about thinking about the world,
Lex Fridman (1:03:28.760)
you know, certain politicians
Mark Zuckerberg (1:03:30.800)
or people kind of in that sphere
Lex Fridman (1:03:32.680)
who I just think have like very insightful perspectives
Mark Zuckerberg (1:03:36.160)
who even if they would,
Lex Fridman (1:03:39.800)
they come at the world from a different perspective,
Mark Zuckerberg (1:03:41.600)
which is sort of what makes the perspective so valuable.
Lex Fridman (1:03:44.360)
But, you know, I think fundamentally
Mark Zuckerberg (1:03:46.200)
we're trying to get to the same place
Lex Fridman (1:03:47.560)
in terms of, you know, helping people connect more,
Mark Zuckerberg (1:03:50.680)
helping the whole world function better,
Lex Fridman (1:03:53.480)
not just, you know, one place or another.
Lex Fridman (1:03:57.120)
And I don't know, I mean,
Lex Fridman (1:03:58.920)
those are the people whose opinions really matter to me.
Lex Fridman (1:04:02.880)
And I just, it's, you know,
Lex Fridman (1:04:04.240)
that's how I learn on a day to day basis.
Mark Zuckerberg (1:04:05.640)
People are constantly sending me comments on stuff
Lex Fridman (1:04:07.880)
or links to things they found interesting.
Lex Fridman (1:04:10.160)
And I don't know, it's kind of constantly evolving
Lex Fridman (1:04:13.400)
this model of the world
Lex Fridman (1:04:14.480)
and kind of what we should be aspiring to be.
Lex Fridman (1:04:16.840)
You've talked about, you have a famously open culture
Mark Zuckerberg (1:04:20.880)
which comes with the criticism
Lex Fridman (1:04:25.440)
and the painful experiences.
Lex Fridman (1:04:27.280)
So let me ask you another difficult question.
Lex Fridman (1:04:30.960)
Frances Haugen, the Facebook whistleblower,
Mark Zuckerberg (1:04:33.440)
leaked the internal Instagram research
Lex Fridman (1:04:35.800)
into teenagers and wellbeing.
Mark Zuckerberg (1:04:38.040)
Her claim is that Instagram is choosing profit
Lex Fridman (1:04:41.240)
over wellbeing of teenage girls.
Lex Fridman (1:04:43.080)
So Instagram is quote, toxic for them.
Lex Fridman (1:04:46.720)
Your response titled,
Lex Fridman (1:04:48.120)
what our research really says about teen wellbeing
Lex Fridman (1:04:52.440)
and Instagram says, no, Instagram research shows
Mark Zuckerberg (1:04:55.520)
that 11 of 12 wellbeing issues,
Lex Fridman (1:04:58.800)
teenage girls who said they struggle
Mark Zuckerberg (1:05:02.720)
with those difficult issues also said
Lex Fridman (1:05:04.400)
that Instagram made them better rather than worse.
Mark Zuckerberg (1:05:07.600)
Again, can you steal man and defend the point
Lex Fridman (1:05:11.000)
and Frances Haugen's characterization of the study
Lex Fridman (1:05:14.800)
and then help me understand the positive
Lex Fridman (1:05:17.080)
and negative effects of Instagram
Lex Fridman (1:05:19.000)
and Facebook on young people?
Lex Fridman (1:05:20.880)
So there are certainly questions around teen mental health
Mark Zuckerberg (1:05:25.840)
that are really important.
Lex Fridman (1:05:26.680)
It's hard to, as a parent, it's like hard to imagine
Mark Zuckerberg (1:05:29.480)
any set of questions that are sort of more important.
Lex Fridman (1:05:32.040)
I mean, I guess maybe other aspects of physical health
Mark Zuckerberg (1:05:34.080)
or wellbeing are probably come to that level,
Lex Fridman (1:05:37.240)
but like, these are really important questions, right?
Mark Zuckerberg (1:05:40.600)
Which is why we dedicate teams to studying them.
Lex Fridman (1:05:45.640)
I don't think the internet or social media are unique
Mark Zuckerberg (1:05:48.880)
in having these questions.
Lex Fridman (1:05:50.000)
I mean, I think people and there've been sort of magazines
Mark Zuckerberg (1:05:53.160)
with promoting certain body types for women
Lex Fridman (1:05:56.320)
and kids for decades,
Lex Fridman (1:05:58.520)
but we really care about this stuff.
Lex Fridman (1:06:01.440)
So we wanted to study it.
Lex Fridman (1:06:02.760)
And of course, we didn't expect
Lex Fridman (1:06:05.000)
that everything was gonna be positive all the time.
Mark Zuckerberg (1:06:07.000)
So, I mean, the reason why you study this stuff
Lex Fridman (1:06:08.520)
is to try to improve and get better.
Mark Zuckerberg (1:06:10.760)
So, I mean, look, the place where I disagree
Lex Fridman (1:06:13.200)
with the characterization first,
Mark Zuckerberg (1:06:15.320)
I thought some of the reporting and coverage of it
Lex Fridman (1:06:18.720)
just took the whole thing out of proportion
Lex Fridman (1:06:20.840)
and that it focused on, as you said,
Lex Fridman (1:06:22.640)
I think there were like 20 metrics in there
Lex Fridman (1:06:24.280)
and on 18 or 19, the effect of using Instagram
Lex Fridman (1:06:27.600)
was neutral or positive on the teen's wellbeing.
Lex Fridman (1:06:30.920)
And there was one area where I think it showed
Lex Fridman (1:06:34.560)
that we needed to improve
Lex Fridman (1:06:35.480)
and we took some steps to try to do that
Lex Fridman (1:06:37.720)
after doing the research.
Lex Fridman (1:06:38.800)
But I think having the coverage just focus on that one
Lex Fridman (1:06:41.680)
without focusing on the,
Mark Zuckerberg (1:06:43.040)
I mean, I think an accurate characterization
Lex Fridman (1:06:45.080)
would have been that kids using Instagram
Mark Zuckerberg (1:06:47.920)
or not kids, teens is generally positive
Lex Fridman (1:06:52.200)
for their mental health.
Lex Fridman (1:06:53.760)
But of course, that was not the narrative that came out.
Lex Fridman (1:06:55.560)
So I think it's hard to,
Mark Zuckerberg (1:06:56.720)
that's not a kind of logical thing to straw man,
Lex Fridman (1:06:59.200)
but I sort of disagree or steel man,
Lex Fridman (1:07:01.440)
but I sort of disagree with that overall characterization.
Lex Fridman (1:07:04.040)
I think anyone sort of looking at this objectively would,
Lex Fridman (1:07:09.960)
but then, I mean, there is this sort of intent critique
Lex Fridman (1:07:15.040)
that I think you were getting at before,
Lex Fridman (1:07:16.360)
which says, it assumes some sort of malevolence, right?
Lex Fridman (1:07:19.680)
It's like, which it's really hard for me
Mark Zuckerberg (1:07:23.160)
to really wrap my head around this
Lex Fridman (1:07:26.520)
because as far as I know,
Mark Zuckerberg (1:07:29.800)
it's not clear that any of the other tech companies
Lex Fridman (1:07:31.720)
are doing this kind of research.
Lex Fridman (1:07:33.320)
So why the narrative should form that we did research
Lex Fridman (1:07:37.840)
because we were studying an issue
Mark Zuckerberg (1:07:38.800)
because we wanted to understand it to improve
Lex Fridman (1:07:40.760)
and took steps after that to try to improve it,
Mark Zuckerberg (1:07:43.560)
that your interpretation of that would be
Lex Fridman (1:07:46.280)
that we did the research
Lex Fridman (1:07:47.880)
and tried to sweep it under the rug.
Lex Fridman (1:07:49.240)
It just, it sort of is like, I don't know,
Mark Zuckerberg (1:07:53.920)
it's beyond credibility to me
Lex Fridman (1:07:55.920)
that like that's the accurate description of the actions
Mark Zuckerberg (1:07:59.000)
that we've taken compared to the others in the industry.
Lex Fridman (1:08:01.160)
So I don't know, that's kind of, that's my view on it.
Mark Zuckerberg (1:08:05.280)
These are really important issues
Lex Fridman (1:08:06.600)
and there's a lot of stuff
Mark Zuckerberg (1:08:07.960)
that I think we're gonna be working on
Lex Fridman (1:08:09.120)
related to teen mental health for a long time,
Mark Zuckerberg (1:08:11.400)
including trying to understand this better.
Lex Fridman (1:08:14.240)
And I would encourage everyone else
Mark Zuckerberg (1:08:15.280)
in the industry to do this too.
Lex Fridman (1:08:18.400)
Yeah, I would love there to be open conversations
Lex Fridman (1:08:21.880)
and a lot of great research being released internally
Lex Fridman (1:08:25.680)
and then also externally.
Mark Zuckerberg (1:08:27.960)
It doesn't make me feel good
Lex Fridman (1:08:31.040)
to see press obviously get way more clicks
Mark Zuckerberg (1:08:35.000)
when they say negative things about social media.
Lex Fridman (1:08:39.280)
Objectively speaking, I can just tell
Mark Zuckerberg (1:08:42.400)
that there's hunger to say negative things
Lex Fridman (1:08:44.400)
about social media.
Lex Fridman (1:08:46.040)
And I don't understand how that's supposed to lead
Lex Fridman (1:08:50.720)
to an open conversation about the positives
Lex Fridman (1:08:53.000)
and the negatives, the concerns about social media,
Lex Fridman (1:08:56.000)
especially when you're doing that kind of research.
Mark Zuckerberg (1:08:59.200)
I mean, I don't know what to do with that,
Lex Fridman (1:09:01.720)
but let me ask you as a father,
Mark Zuckerberg (1:09:05.560)
there's a weight heavy on you
Lex Fridman (1:09:06.720)
that people get bullied on social networks.
Lex Fridman (1:09:10.280)
So people get bullied in their private life.
Lex Fridman (1:09:13.520)
But now because so much of our life is in the digital world,
Mark Zuckerberg (1:09:17.080)
the bullying moves from the physical world
Lex Fridman (1:09:19.640)
to the digital world.
Lex Fridman (1:09:21.200)
So you're now creating a platform
Lex Fridman (1:09:24.520)
on which bullying happens.
Lex Fridman (1:09:26.520)
And some of that bullying can lead to damage
Lex Fridman (1:09:30.440)
to mental health.
Lex Fridman (1:09:31.840)
And some of that bullying can lead to depression,
Lex Fridman (1:09:35.120)
even suicide.
Mark Zuckerberg (1:09:37.640)
There's a weight heavy on you
Lex Fridman (1:09:38.760)
that people have committed suicide
Mark Zuckerberg (1:09:43.240)
or will commit suicide based on the bullying
Lex Fridman (1:09:46.120)
that happens on social media.
Mark Zuckerberg (1:09:48.080)
Yeah, I mean, there's a set of harms
Lex Fridman (1:09:51.560)
that we basically track and build systems to fight against.
Lex Fridman (1:09:55.400)
And bullying and self harm are,
Lex Fridman (1:10:01.600)
these are some of the biggest things
Mark Zuckerberg (1:10:03.240)
that we are most focused on.
Lex Fridman (1:10:10.920)
For bullying, like you say, it's gonna be,
Mark Zuckerberg (1:10:16.200)
while this predates the internet,
Lex Fridman (1:10:18.240)
then it's probably impossible to get rid of all of it.
Mark Zuckerberg (1:10:22.000)
You wanna give people tools to fight it
Lex Fridman (1:10:24.160)
and you wanna fight it yourself.
Lex Fridman (1:10:27.200)
And you also wanna make sure that people have the tools
Lex Fridman (1:10:28.800)
to get help when they need it.
Lex Fridman (1:10:30.160)
So I think this isn't like a question of,
Lex Fridman (1:10:33.200)
can you get rid of all bullying?
Mark Zuckerberg (1:10:34.640)
I mean, it's like, all right, I mean, I have two daughters
Lex Fridman (1:10:39.080)
and they fight and push each other around and stuff too.
Lex Fridman (1:10:43.840)
And the question is just,
Lex Fridman (1:10:44.680)
how do you handle that situation?
Lex Fridman (1:10:47.080)
And there's a handful of things that I think you can do.
Lex Fridman (1:10:51.480)
We talked a little bit before around some of the AI tools
Mark Zuckerberg (1:10:55.000)
that you can build to identify
Lex Fridman (1:10:56.800)
when something harmful is happening.
Mark Zuckerberg (1:10:59.160)
It's actually, it's very hard in bullying
Lex Fridman (1:11:00.480)
because a lot of bullying is very context specific.
Mark Zuckerberg (1:11:02.720)
It's not like you're trying to fit a formula of like,
Lex Fridman (1:11:06.320)
if like looking at the different harms,
Mark Zuckerberg (1:11:09.480)
someone promoting a terrorist group is like,
Lex Fridman (1:11:12.280)
probably one of the simpler things to generally find
Mark Zuckerberg (1:11:14.360)
because things promoting that group are gonna look
Lex Fridman (1:11:17.000)
at a certain way or feel a certain way.
Mark Zuckerberg (1:11:19.240)
Bullying could just be, you know,
Lex Fridman (1:11:21.840)
someone making some subtle comment about someone's appearance
Mark Zuckerberg (1:11:24.880)
that's idiosyncratic to them.
Lex Fridman (1:11:26.800)
And it could look at just like humor.
Lex Fridman (1:11:28.680)
So humor to one person can be destructive
Lex Fridman (1:11:31.000)
to another human being, yeah.
Lex Fridman (1:11:32.280)
So with bullying, I think there are certain things
Lex Fridman (1:11:36.400)
that you can find through AI systems,
Lex Fridman (1:11:40.240)
but I think it is increasingly important
Lex Fridman (1:11:42.640)
to just give people more agency themselves.
Lex Fridman (1:11:44.800)
So we've done things like making it
Lex Fridman (1:11:46.480)
so people can turn off comments
Mark Zuckerberg (1:11:47.800)
or take a break from hearing from a specific person
Lex Fridman (1:11:52.240)
without having to signal at all
Mark Zuckerberg (1:11:54.120)
that they're gonna stop following them
Lex Fridman (1:11:55.600)
or kind of make some stand that,
Mark Zuckerberg (1:11:58.200)
okay, I'm not friends with you anymore.
Lex Fridman (1:11:59.400)
I'm not following you.
Mark Zuckerberg (1:12:00.440)
I just like, I just don't wanna hear about this,
Lex Fridman (1:12:01.880)
but I also don't wanna signal at all publicly
Mark Zuckerberg (1:12:05.560)
that or to them that there's been an issue.
Lex Fridman (1:12:10.880)
And then you get to some of the more extreme cases
Mark Zuckerberg (1:12:14.040)
like you're talking about
Lex Fridman (1:12:14.880)
where someone is thinking about self harm or suicide.
Lex Fridman (1:12:19.160)
And there we've found that that is a place
Lex Fridman (1:12:24.080)
where AI can identify a lot
Mark Zuckerberg (1:12:26.400)
as well as people flagging things.
Lex Fridman (1:12:28.560)
If people are expressing something
Mark Zuckerberg (1:12:31.080)
that is potentially they're thinking of hurting themselves,
Lex Fridman (1:12:35.000)
those are cues that you can build systems
Lex Fridman (1:12:37.520)
and hundreds of languages around the world
Lex Fridman (1:12:39.560)
to be able to identify that.
Lex Fridman (1:12:41.080)
And one of the things that I'm actually quite proud of
Lex Fridman (1:12:45.320)
is we've built these systems
Mark Zuckerberg (1:12:47.400)
that I think are clearly leading at this point
Lex Fridman (1:12:50.960)
that not only identify that,
Lex Fridman (1:12:53.000)
but then connect with local first responders
Lex Fridman (1:12:57.040)
and have been able to save, I think at this point,
Mark Zuckerberg (1:12:59.680)
it's in thousands of cases,
Lex Fridman (1:13:01.960)
be able to get first responders to people
Mark Zuckerberg (1:13:04.560)
through these systems who really need them
Lex Fridman (1:13:07.800)
because of specific plumbing that we've done
Mark Zuckerberg (1:13:09.600)
between the AI work and being able to communicate
Lex Fridman (1:13:11.680)
with local first responder organizations.
Mark Zuckerberg (1:13:13.800)
We're rolling that out in more places around the world.
Lex Fridman (1:13:15.800)
And I think the team that worked on that
Mark Zuckerberg (1:13:18.160)
just did awesome stuff.
Lex Fridman (1:13:19.360)
So I think that that's a long way of saying,
Mark Zuckerberg (1:13:22.760)
yeah, I mean, this is a heavy topic
Lex Fridman (1:13:25.480)
and you want to attack it in a bunch of different ways
Lex Fridman (1:13:30.000)
and also kind of understand that some of nature
Lex Fridman (1:13:33.240)
is for people to do this to each other,
Mark Zuckerberg (1:13:36.360)
which is unfortunate,
Lex Fridman (1:13:37.320)
but you can give people tools and build things that help.
Mark Zuckerberg (1:13:40.600)
It's still one hell of a burden though.
Lex Fridman (1:13:43.840)
A platform that allows people
Mark Zuckerberg (1:13:46.160)
to fall in love with each other
Lex Fridman (1:13:48.600)
is also by nature going to be a platform
Mark Zuckerberg (1:13:51.000)
that allows people to hurt each other.
Lex Fridman (1:13:52.880)
And when you're managing such a platform, it's difficult.
Lex Fridman (1:13:57.120)
And I think you spoke to it,
Lex Fridman (1:13:58.200)
but the psychology of that, of being a leader in that space,
Mark Zuckerberg (1:14:01.280)
of creating technology that's playing in this space,
Lex Fridman (1:14:05.280)
like you mentioned, psychology is really damn difficult.
Lex Fridman (1:14:10.280)
And I mean, the burden of that is just great.
Lex Fridman (1:14:13.120)
I just wanted to hear you speak to that point.
Mark Zuckerberg (1:14:18.720)
I have to ask about the thing you've brought up a few times,
Lex Fridman (1:14:23.160)
which is making controversial decisions.
Mark Zuckerberg (1:14:26.520)
Let's talk about free speech and censorship.
Lex Fridman (1:14:29.440)
So there are two groups of people pressuring Meta on this.
Mark Zuckerberg (1:14:33.920)
One group is upset that Facebook, the social network,
Lex Fridman (1:14:37.240)
allows misinformation in quotes to be spread on the platform.
Mark Zuckerberg (1:14:41.480)
The other group are concerned that Facebook censors speech
Lex Fridman (1:14:44.800)
by calling it misinformation.
Lex Fridman (1:14:46.560)
So you're getting it from both sides.
Lex Fridman (1:14:48.840)
You, in 2019, October at Georgetown University,
Mark Zuckerberg (1:14:54.600)
eloquently defended the importance of free speech,
Lex Fridman (1:14:58.240)
but then COVID came and the 2020 election came.
Lex Fridman (1:15:04.360)
Do you worry that outside pressures
Lex Fridman (1:15:06.440)
from advertisers, politicians, the public,
Mark Zuckerberg (1:15:08.840)
have forced Meta to damage the ideal of free speech
Lex Fridman (1:15:11.840)
that you spoke highly of?
Mark Zuckerberg (1:15:14.000)
Just to say some obvious things upfront,
Lex Fridman (1:15:16.880)
I don't think pressure from advertisers
Mark Zuckerberg (1:15:18.840)
or politicians directly in any way
Lex Fridman (1:15:21.440)
affects how we think about this.
Mark Zuckerberg (1:15:22.680)
I think these are just hard topics.
Lex Fridman (1:15:25.080)
So let me just take you through our evolution
Mark Zuckerberg (1:15:26.880)
from kind of the beginning of the company
Lex Fridman (1:15:28.240)
to where we are now.
Mark Zuckerberg (1:15:30.240)
You don't build a company like this
Lex Fridman (1:15:31.720)
unless you believe that people expressing themselves
Lex Fridman (1:15:34.120)
is a good thing, right?
Lex Fridman (1:15:35.720)
So that's sort of the foundational thing.
Mark Zuckerberg (1:15:38.160)
You can kind of think about our company as a formula
Lex Fridman (1:15:41.880)
where we think giving people voice
Lex Fridman (1:15:44.240)
and helping people connect creates opportunity, right?
Lex Fridman (1:15:47.440)
So those are the two things that we're always focused on
Mark Zuckerberg (1:15:49.640)
are sort of helping people connect.
Lex Fridman (1:15:50.800)
We talked about that a lot,
Lex Fridman (1:15:52.080)
but also giving people voice
Lex Fridman (1:15:53.880)
and ability to express themselves.
Mark Zuckerberg (1:15:55.800)
Then by the way, most of the time
Lex Fridman (1:15:56.960)
when people express themselves,
Mark Zuckerberg (1:15:58.120)
that's not like politically controversial content.
Lex Fridman (1:16:00.840)
It's like expressing something about their identity
Mark Zuckerberg (1:16:04.040)
that's more related to the avatar conversation
Lex Fridman (1:16:06.520)
we had earlier in terms of expressing some facet,
Lex Fridman (1:16:08.600)
but that's what's important to people on a day to day basis.
Lex Fridman (1:16:11.240)
And sometimes when people feel strongly enough
Mark Zuckerberg (1:16:13.480)
about something, it kind of becomes a political topic.
Lex Fridman (1:16:16.360)
That's sort of always been a thing that we've focused on.
Mark Zuckerberg (1:16:19.120)
There's always been the question of safety in this,
Lex Fridman (1:16:22.320)
which if you're building a community,
Mark Zuckerberg (1:16:24.360)
I think you have to focus on safety.
Lex Fridman (1:16:26.040)
We've had these community standards from early on,
Lex Fridman (1:16:28.320)
and there are about 20 different kinds of harm
Lex Fridman (1:16:32.640)
that we track and try to fight actively.
Mark Zuckerberg (1:16:34.840)
We've talked about some of them already.
Lex Fridman (1:16:36.400)
So it includes things like bullying and harassment.
Mark Zuckerberg (1:16:40.880)
It includes things like terrorism or promoting terrorism,
Lex Fridman (1:16:46.000)
inciting violence, intellectual property theft.
Lex Fridman (1:16:49.320)
And in general, I think call it about 18 out of 20 of those.
Lex Fridman (1:16:53.760)
There's not really a particularly polarized definition
Mark Zuckerberg (1:16:57.200)
of that.
Lex Fridman (1:16:59.160)
I think you're not really gonna find many people
Mark Zuckerberg (1:17:01.440)
in the country or in the world
Lex Fridman (1:17:03.760)
who are trying to say we should be
Mark Zuckerberg (1:17:07.040)
fighting terrorist content less.
Lex Fridman (1:17:09.320)
I think the content where there are a couple of areas
Mark Zuckerberg (1:17:12.200)
where I think that this has gotten more controversial
Lex Fridman (1:17:14.000)
recently, which I'll talk about.
Lex Fridman (1:17:16.320)
And you're right, the misinformation is basically is up there.
Lex Fridman (1:17:20.000)
And I think sometimes the definition of hate speech
Mark Zuckerberg (1:17:21.920)
is up there too.
Lex Fridman (1:17:22.760)
But I think in general, most of the content
Mark Zuckerberg (1:17:25.760)
that I think we're working on for safety
Lex Fridman (1:17:29.560)
is not actually, people don't kind of have these questions.
Lex Fridman (1:17:32.560)
So it's sort of this subset.
Lex Fridman (1:17:35.280)
But if you go back to the beginning of the company,
Mark Zuckerberg (1:17:37.400)
this was sort of pre deep learning days.
Lex Fridman (1:17:42.000)
And therefore, it was me and my roommate Dustin join me.
Lex Fridman (1:17:47.000)
And if someone posted something bad,
Lex Fridman (1:17:54.040)
it was the AI technology did not exist yet
Mark Zuckerberg (1:17:57.880)
to be able to go basically look at all the content.
Lex Fridman (1:18:02.480)
And we were a small enough outfit
Mark Zuckerberg (1:18:06.120)
that no one would expect that we could review it all.
Lex Fridman (1:18:08.800)
Even if someone reported it to us,
Lex Fridman (1:18:10.400)
we basically did our best, right?
Lex Fridman (1:18:11.720)
It's like someone would report it
Lex Fridman (1:18:12.680)
and we try to look at stuff and deal with stuff.
Lex Fridman (1:18:16.880)
And for call it the first seven or eight years
Mark Zuckerberg (1:18:22.360)
of the company, we weren't that big of a company.
Lex Fridman (1:18:26.600)
For a lot of that period, we weren't even really profitable.
Mark Zuckerberg (1:18:28.760)
The AI didn't really exist to be able to do
Lex Fridman (1:18:30.520)
the kind of moderation that we do today.
Lex Fridman (1:18:32.720)
And then at some point in kind of the middle
Lex Fridman (1:18:35.760)
of the last decade, that started to flip.
Lex Fridman (1:18:38.160)
And we got to the point where we were sort of a larger
Lex Fridman (1:18:44.160)
and more profitable company.
Lex Fridman (1:18:45.240)
And the AI was starting to come online
Lex Fridman (1:18:48.000)
to be able to proactively detect
Mark Zuckerberg (1:18:50.480)
some of the simpler forms of this.
Lex Fridman (1:18:52.840)
So things like pornography,
Mark Zuckerberg (1:18:54.800)
you could train an image classifier
Lex Fridman (1:18:57.600)
to identify what a nipple was,
Mark Zuckerberg (1:18:59.520)
or you can fight against terrorist content.
Lex Fridman (1:19:01.320)
You still could.
Mark Zuckerberg (1:19:02.160)
There's actually papers on this, it's great.
Lex Fridman (1:19:03.440)
Oh, of course there are.
Mark Zuckerberg (1:19:04.280)
Technical papers.
Lex Fridman (1:19:05.120)
Of course there are.
Mark Zuckerberg (1:19:06.480)
Those are relatively easier things to train AI to do
Lex Fridman (1:19:09.280)
than for example, understand the nuances
Mark Zuckerberg (1:19:12.440)
of what is inciting violence
Lex Fridman (1:19:14.000)
in a hundred languages around the world
Lex Fridman (1:19:15.800)
and not have the false positives of like,
Lex Fridman (1:19:20.200)
okay, are you posting about this thing
Mark Zuckerberg (1:19:22.360)
that might be inciting violence
Lex Fridman (1:19:24.040)
because you're actually trying to denounce it?
Mark Zuckerberg (1:19:26.360)
In which case we probably shouldn't take that down.
Lex Fridman (1:19:28.280)
Where if you're trying to denounce something
Mark Zuckerberg (1:19:29.520)
that's inciting violence in some kind of dialect
Lex Fridman (1:19:33.920)
in a corner of India, as opposed to,
Mark Zuckerberg (1:19:37.200)
okay, actually you're posting this thing
Lex Fridman (1:19:38.440)
because you're trying to incite violence.
Mark Zuckerberg (1:19:39.600)
Okay, building an AI that can basically get
Lex Fridman (1:19:42.360)
to that level of nuance and all the languages
Mark Zuckerberg (1:19:44.400)
that we serve is something that I think
Lex Fridman (1:19:47.120)
is only really becoming possible now,
Mark Zuckerberg (1:19:49.680)
not towards the middle of the last decade.
Lex Fridman (1:19:51.920)
But there's been this evolution,
Lex Fridman (1:19:54.880)
and I think what happened,
Lex Fridman (1:19:57.560)
people sort of woke up after 2016
Lex Fridman (1:20:00.120)
and a lot of people are like,
Lex Fridman (1:20:02.560)
okay, the country is a lot more polarized
Lex Fridman (1:20:05.000)
and there's a lot more stuff here than we realized.
Lex Fridman (1:20:08.080)
Why weren't these internet companies on top of this?
Lex Fridman (1:20:11.800)
And I think at that point it was reasonable feedback
Lex Fridman (1:20:18.760)
that some of this technology had started becoming possible.
Lex Fridman (1:20:22.400)
And at that point, I really did feel like
Lex Fridman (1:20:25.320)
we needed to make a substantially larger investment.
Mark Zuckerberg (1:20:27.920)
We'd already worked on this stuff a lot,
Lex Fridman (1:20:29.680)
on AI and on these integrity problems,
Lex Fridman (1:20:32.400)
but that we should basically invest,
Lex Fridman (1:20:35.360)
have a thousand or more engineers
Mark Zuckerberg (1:20:37.080)
basically work on building these AI systems
Lex Fridman (1:20:39.160)
to be able to go and proactively identify the stuff
Mark Zuckerberg (1:20:41.600)
across all these different areas.
Lex Fridman (1:20:43.680)
Okay, so we went and did that.
Mark Zuckerberg (1:20:45.360)
Now we've built the tools to be able to do that.
Lex Fridman (1:20:48.000)
And now I think it's actually a much more complicated
Mark Zuckerberg (1:20:50.560)
set of philosophical rather than technical questions,
Lex Fridman (1:20:53.400)
which is the exact policies, which are okay.
Mark Zuckerberg (1:20:56.960)
Now, the way that we basically hold ourselves accountable
Lex Fridman (1:21:01.960)
is we issue these transparency reports every quarter
Lex Fridman (1:21:04.400)
and the metric that we track is for each of these
Lex Fridman (1:21:06.320)
20 types of harmful content.
Lex Fridman (1:21:10.240)
How much of that content are we taking down
Lex Fridman (1:21:12.400)
before someone even has to report it to us?
Lex Fridman (1:21:14.320)
So how effective is our AI at doing this?
Lex Fridman (1:21:17.000)
But that basically creates this big question,
Mark Zuckerberg (1:21:19.440)
which is okay, now we need to really be careful
Lex Fridman (1:21:22.960)
about how proactive we set the AI
Lex Fridman (1:21:25.440)
and where the exact policy lines are
Lex Fridman (1:21:28.000)
around what we're taking down.
Mark Zuckerberg (1:21:30.200)
It's certainly at a point now where I felt like
Lex Fridman (1:21:35.200)
at the beginning of that journey
Mark Zuckerberg (1:21:37.720)
of building those AI systems, there was a lot of push.
Lex Fridman (1:21:43.160)
There's saying, okay, you've got to do more.
Mark Zuckerberg (1:21:44.360)
There's clearly a lot more bad content
Lex Fridman (1:21:46.320)
that people aren't reporting or that you're not getting to
Lex Fridman (1:21:49.880)
and you need to get more effective at that.
Lex Fridman (1:21:51.200)
And I was pretty sympathetic to that.
Lex Fridman (1:21:52.920)
But then I think at some point along the way,
Lex Fridman (1:21:54.800)
there started to be almost equal issues on both sides
Mark Zuckerberg (1:21:58.960)
of, okay, actually you're kind of taking down
Lex Fridman (1:22:00.960)
too much stuff, right?
Mark Zuckerberg (1:22:02.080)
Or some of the stuff is borderline
Lex Fridman (1:22:05.560)
and it wasn't really bothering anyone
Lex Fridman (1:22:07.560)
and they didn't report it.
Lex Fridman (1:22:09.640)
So is that really an issue that you need to take down?
Mark Zuckerberg (1:22:13.000)
Whereas we still have the critique on the other side too
Lex Fridman (1:22:15.440)
where a lot of people think we're not doing enough.
Lex Fridman (1:22:18.560)
So it's become, as we built the technical capacity,
Lex Fridman (1:22:21.840)
I think it becomes more philosophically interesting almost
Mark Zuckerberg (1:22:25.960)
where you wanna be on the line.
Lex Fridman (1:22:27.520)
And I just think you don't want one person
Mark Zuckerberg (1:22:31.160)
making those decisions.
Lex Fridman (1:22:32.440)
So we've also tried to innovate
Mark Zuckerberg (1:22:33.760)
in terms of building out this independent oversight board,
Lex Fridman (1:22:36.520)
which has people who are dedicated to free expression
Lex Fridman (1:22:39.520)
but from around the world who people can appeal cases to.
Lex Fridman (1:22:43.640)
So a lot of the most controversial cases basically go to them
Lex Fridman (1:22:46.200)
and they make the final binding decision
Lex Fridman (1:22:47.640)
on how we should handle that.
Lex Fridman (1:22:49.080)
And then of course, their decisions,
Lex Fridman (1:22:50.680)
we then try to figure out what the principles are
Mark Zuckerberg (1:22:53.000)
behind those and encode them into the algorithms.
Lex Fridman (1:22:55.760)
And how are those people chosen, which, you know,
Mark Zuckerberg (1:22:58.080)
you're outsourcing a difficult decision.
Lex Fridman (1:23:00.200)
Yeah, the initial people,
Mark Zuckerberg (1:23:02.560)
we chose a handful of chairs for the group
Lex Fridman (1:23:09.040)
and we basically chose the people
Mark Zuckerberg (1:23:12.480)
for a commitment to free expression
Lex Fridman (1:23:16.480)
and like a broad understanding of human rights
Lex Fridman (1:23:19.640)
and the trade offs around free expression.
Lex Fridman (1:23:21.520)
So they fundamentally people
Mark Zuckerberg (1:23:22.720)
who are gonna lean towards free expression.
Lex Fridman (1:23:24.880)
Towards freedom of speech.
Mark Zuckerberg (1:23:26.040)
Okay, so there's also this idea of fact checkers.
Lex Fridman (1:23:28.560)
So jumping around to the misinformation questions,
Mark Zuckerberg (1:23:31.600)
especially during COVID,
Lex Fridman (1:23:33.120)
which is an exceptionally speaking of polarization.
Lex Fridman (1:23:36.080)
Can I speak to the COVID thing?
Lex Fridman (1:23:38.240)
I mean, I think one of the hardest set of questions
Mark Zuckerberg (1:23:40.240)
around free expression,
Lex Fridman (1:23:41.200)
because you asked about Georgetown
Lex Fridman (1:23:42.240)
has my stance fundamentally changed?
Lex Fridman (1:23:43.840)
And the answer to that is no, my stance has not changed.
Mark Zuckerberg (1:23:48.400)
It is fundamentally the same as when I was talking
Lex Fridman (1:23:52.040)
at Georgetown from a philosophical perspective.
Mark Zuckerberg (1:23:56.480)
The challenge with free speech is that everyone agrees
Lex Fridman (1:24:01.840)
that there is a line where if you're actually
Mark Zuckerberg (1:24:05.440)
about to do physical harm to people
Lex Fridman (1:24:08.160)
that there should be restrictions.
Mark Zuckerberg (1:24:10.560)
So, I mean, there's the famous Supreme Court
Lex Fridman (1:24:13.960)
historical example of like,
Mark Zuckerberg (1:24:15.120)
you can't yell fire in a crowded theater.
Lex Fridman (1:24:18.040)
The thing that everyone disagrees on
Lex Fridman (1:24:20.360)
is what is the definition of real harm?
Lex Fridman (1:24:22.680)
Where I think some people think,
Mark Zuckerberg (1:24:24.560)
okay, this should only be a very literal,
Lex Fridman (1:24:27.920)
I mean, take it back to the bullying conversation
Mark Zuckerberg (1:24:29.840)
we were just having, where is it just harm
Lex Fridman (1:24:32.760)
if the person is about to hurt themselves
Lex Fridman (1:24:34.800)
because they've been bullied so hard?
Lex Fridman (1:24:36.640)
Or is it actually harm like as they're being bullied?
Lex Fridman (1:24:39.880)
And kind of at what point in the spectrum is that?
Lex Fridman (1:24:42.160)
And that's the part that there's not agreement on.
Lex Fridman (1:24:44.480)
But I think what people agree on pretty broadly
Lex Fridman (1:24:47.000)
is that when there is an acute threat
Mark Zuckerberg (1:24:49.440)
that it does make sense from a societal perspective
Lex Fridman (1:24:52.960)
to tolerate less speech.
Mark Zuckerberg (1:24:57.120)
That could be potentially harmful in that acute situation.
Lex Fridman (1:24:59.560)
So I think where COVID got very difficult is,
Mark Zuckerberg (1:25:02.840)
I don't think anyone expected this to be going on for years.
Lex Fridman (1:25:06.000)
But if you'd kind of asked now a priori,
Mark Zuckerberg (1:25:10.360)
would a global pandemic where a lot of people are dying
Lex Fridman (1:25:14.880)
and catching this, is that an emergency
Mark Zuckerberg (1:25:19.040)
that where you'd kind of consider it
Lex Fridman (1:25:21.560)
that it's problematic to basically yell fire
Lex Fridman (1:25:25.600)
in a crowded theater?
Lex Fridman (1:25:26.840)
I think that that probably passes that test.
Lex Fridman (1:25:29.000)
So I think that it's a very tricky situation,
Lex Fridman (1:25:32.320)
but I think the fundamental commitment
Mark Zuckerberg (1:25:35.240)
to free expression is there.
Lex Fridman (1:25:38.200)
And that's what I believe.
Lex Fridman (1:25:39.840)
And again, I don't think you start this company
Lex Fridman (1:25:41.440)
unless you care about people being able
Mark Zuckerberg (1:25:42.720)
to express themselves as much as possible.
Lex Fridman (1:25:44.800)
But I think that that's the question,
Mark Zuckerberg (1:25:48.880)
is how do you define what the harm is
Lex Fridman (1:25:50.480)
and how acute that is?
Lex Fridman (1:25:52.440)
And what are the institutions that define that harm?
Lex Fridman (1:25:55.440)
A lot of the criticism is that the CDC, the WHO,
Mark Zuckerberg (1:25:59.720)
the institutions we've come to trust as a civilization
Lex Fridman (1:26:03.800)
to give the line of what is and isn't harm
Mark Zuckerberg (1:26:07.760)
in terms of health policy have failed in many ways,
Lex Fridman (1:26:11.640)
in small ways and in big ways, depending on who you ask.
Lex Fridman (1:26:14.320)
And then the perspective of meta and Facebook is like,
Lex Fridman (1:26:17.120)
well, where the hell do I get the information
Lex Fridman (1:26:20.160)
of what is and isn't misinformation?
Lex Fridman (1:26:22.400)
So it's a really difficult place to be in,
Lex Fridman (1:26:25.160)
but it's great to hear that you're leaning
Lex Fridman (1:26:26.720)
towards freedom of speech on this aspect.
Lex Fridman (1:26:30.140)
And again, I think this actually calls to the fact
Lex Fridman (1:26:33.000)
that we need to reform institutions
Mark Zuckerberg (1:26:35.320)
that help keep an open mind
Lex Fridman (1:26:36.800)
of what is and isn't misinformation.
Lex Fridman (1:26:39.880)
And misinformation has been used to bully on the internet.
Lex Fridman (1:26:44.600)
I mean, I just have, I'm friends with Joe Rogan
Lex Fridman (1:26:46.920)
and he is called as a,
Lex Fridman (1:26:49.280)
I remember hanging out with him in Vegas
Lex Fridman (1:26:51.280)
and somebody yelled, stop spreading misinformation.
Lex Fridman (1:26:54.660)
I mean, and there's a lot of people that follow him
Mark Zuckerberg (1:26:57.640)
that believe he's not spreading misinformation.
Lex Fridman (1:26:59.880)
Like you can't just not acknowledge the fact
Mark Zuckerberg (1:27:02.900)
that there's a large number of people
Lex Fridman (1:27:05.720)
that have a different definition of misinformation.
Lex Fridman (1:27:08.840)
And that's such a tough place to be.
Lex Fridman (1:27:10.760)
Like who do you listen to?
Lex Fridman (1:27:11.840)
Do you listen to quote unquote experts who gets,
Lex Fridman (1:27:15.320)
as a person who has a PhD, I gotta say,
Mark Zuckerberg (1:27:17.600)
I mean, I'm not sure I know what defines an expert,
Lex Fridman (1:27:21.120)
especially in a new,
Mark Zuckerberg (1:27:24.080)
in a totally new pandemic or a new catastrophic event,
Lex Fridman (1:27:29.160)
especially when politics is involved
Lex Fridman (1:27:31.520)
and especially when the news are,
Lex Fridman (1:27:33.320)
the media involved that can propagate
Mark Zuckerberg (1:27:37.440)
sort of outrageous narratives
Lex Fridman (1:27:39.520)
and thereby make a lot of money.
Lex Fridman (1:27:40.720)
Like what the hell?
Lex Fridman (1:27:41.800)
Where's the source of truth?
Lex Fridman (1:27:43.200)
And then everybody turns to Facebook.
Lex Fridman (1:27:45.480)
It's like, please tell me what the source of truth is.
Mark Zuckerberg (1:27:49.040)
Well, I mean, well, how would you handle this
Lex Fridman (1:27:50.740)
if you were in my position?
Mark Zuckerberg (1:27:52.680)
Is very, very, very, very difficult.
Lex Fridman (1:27:55.160)
I would say,
Mark Zuckerberg (1:27:59.400)
I would more speak about how difficult the choices are
Lex Fridman (1:28:02.680)
and be transparent about like,
Lex Fridman (1:28:04.040)
what the hell do you do with this?
Lex Fridman (1:28:05.360)
Like here, you got exactly,
Mark Zuckerberg (1:28:07.080)
ask the exact question you just asked me,
Lex Fridman (1:28:08.800)
but to the broader public, like, okay, yeah,
Mark Zuckerberg (1:28:10.840)
you guys tell me what to do.
Lex Fridman (1:28:12.400)
So like crowdsource it.
Lex Fridman (1:28:14.200)
And then the other aspect is when you spoke really eloquently
Lex Fridman (1:28:19.800)
about the fact that there's this going back and forth
Lex Fridman (1:28:23.440)
and now there's a feeling like you're censoring
Lex Fridman (1:28:25.240)
a little bit too much.
Lex Fridman (1:28:26.720)
So I would lean, I would try to be ahead of that feeling.
Lex Fridman (1:28:30.240)
I would now lean towards freedom of speech and say,
Mark Zuckerberg (1:28:33.080)
we're not the ones that are going to define misinformation.
Lex Fridman (1:28:36.240)
Let it be a public debate, let the idea stand.
Lex Fridman (1:28:40.040)
And I actually place, this idea of misinformation,
Lex Fridman (1:28:44.280)
I place the responsibility
Mark Zuckerberg (1:28:46.360)
on the poor communication skills of scientists.
Lex Fridman (1:28:50.020)
They should be in the battlefield of ideas
Lex Fridman (1:28:52.560)
and everybody who is spreading information
Lex Fridman (1:28:57.400)
against the vaccine, they should not be censored.
Mark Zuckerberg (1:29:00.400)
They should be talked with and you should show the data,
Lex Fridman (1:29:03.040)
you should have open discussion
Mark Zuckerberg (1:29:04.800)
as opposed to rolling your eyes and saying,
Lex Fridman (1:29:07.080)
I'm the expert, I know what I'm talking about.
Mark Zuckerberg (1:29:09.840)
No, you need to convince people, it's a battle of ideas.
Lex Fridman (1:29:13.240)
So that's the whole point of freedom of speech.
Mark Zuckerberg (1:29:15.360)
It's the way to defeat bad ideas
Lex Fridman (1:29:17.120)
is with good ideas, with speech.
Lex Fridman (1:29:20.080)
So like the responsibility here falls
Lex Fridman (1:29:22.080)
on the poor communication skills of scientists.
Mark Zuckerberg (1:29:26.560)
Thanks to social media, scientists are not communicators.
Lex Fridman (1:29:32.180)
They have the power to communicate.
Mark Zuckerberg (1:29:34.040)
Some of the best stuff I've seen about COVID
Lex Fridman (1:29:36.800)
from doctors is on social media.
Mark Zuckerberg (1:29:38.840)
It's a way to learn to respond really quickly,
Lex Fridman (1:29:41.520)
to go faster than the peer review process.
Lex Fridman (1:29:43.800)
And so they just need to get way better
Lex Fridman (1:29:45.460)
at that communication.
Lex Fridman (1:29:46.480)
And also by better, I don't mean just convincing,
Lex Fridman (1:29:50.060)
I also mean speak with humility,
Mark Zuckerberg (1:29:51.800)
don't talk down to people, all those kinds of things.
Lex Fridman (1:29:54.280)
And as a platform, I would say,
Mark Zuckerberg (1:29:56.860)
I would step back a little bit.
Lex Fridman (1:29:59.800)
Not all the way, of course,
Mark Zuckerberg (1:30:00.800)
because there's a lot of stuff that can cause real harm
Lex Fridman (1:30:03.520)
as we've talked about,
Lex Fridman (1:30:04.440)
but you lean more towards freedom of speech
Lex Fridman (1:30:06.920)
because then people from a brand perspective
Mark Zuckerberg (1:30:09.560)
wouldn't be blaming you for the other ills of society,
Lex Fridman (1:30:13.760)
which there are many.
Mark Zuckerberg (1:30:14.600)
The institutions have flaws, the political divide,
Lex Fridman (1:30:19.840)
obviously politicians have flaws, that's news.
Mark Zuckerberg (1:30:23.400)
The media has flaws that they're all trying to work with.
Lex Fridman (1:30:28.040)
And because of the central place of Facebook in the world,
Mark Zuckerberg (1:30:31.080)
all of those flaws somehow kind of propagate to Facebook.
Lex Fridman (1:30:34.320)
And you're sitting there as Plato, the philosopher,
Mark Zuckerberg (1:30:38.160)
have to answer to some of the most difficult questions
Lex Fridman (1:30:40.720)
asking, being asked of human civilization.
Lex Fridman (1:30:43.960)
So I don't know, maybe this is an American answer though,
Lex Fridman (1:30:47.000)
to lean towards freedom of speech.
Mark Zuckerberg (1:30:48.420)
I don't know if that applies globally.
Lex Fridman (1:30:51.300)
So yeah, I don't know.
Lex Fridman (1:30:52.640)
But transparency and saying, I think as a technologist,
Lex Fridman (1:30:57.400)
one of the things I sense about Facebook and meta
Mark Zuckerberg (1:30:59.560)
when people talk about this company
Lex Fridman (1:31:02.360)
is they don't necessarily understand
Mark Zuckerberg (1:31:04.960)
fully how difficult the problem is.
Lex Fridman (1:31:06.880)
You talked about AI has to catch
Mark Zuckerberg (1:31:08.440)
a bunch of harmful stuff really quickly.
Lex Fridman (1:31:11.720)
Just the sea of data you have to deal with.
Mark Zuckerberg (1:31:14.600)
It's a really difficult problem.
Lex Fridman (1:31:16.700)
So like any of the critics,
Mark Zuckerberg (1:31:18.400)
if you just hand them the helm for a week,
Lex Fridman (1:31:22.840)
let's see how well you can do.
Mark Zuckerberg (1:31:25.400)
Like that, to me, that's definitely something
Lex Fridman (1:31:28.160)
that would wake people up to how difficult this problem is
Mark Zuckerberg (1:31:31.320)
if there's more transparency
Lex Fridman (1:31:32.640)
of saying how difficult this problem is.
Mark Zuckerberg (1:31:35.580)
Let me ask you about, on the AI front,
Lex Fridman (1:31:37.800)
just because you mentioned language and my ineloquence.
Mark Zuckerberg (1:31:41.600)
Translation is something I wanted to ask you about.
Lex Fridman (1:31:44.120)
And first, just to give a shout out to the supercomputer.
Mark Zuckerberg (1:31:47.760)
You've recently announced the AI research supercluster, RSC.
Lex Fridman (1:31:51.960)
Obviously, I'm somebody who loves the GPUs.
Mark Zuckerberg (1:31:54.680)
It currently has 6,000 GPUs.
Lex Fridman (1:31:57.120)
NVIDIA DGX A100 is the systems that have
Mark Zuckerberg (1:32:02.160)
in total 6,000 GPUs.
Lex Fridman (1:32:04.080)
And it will eventually, maybe this year,
Mark Zuckerberg (1:32:06.560)
maybe soon, will have 16,000 GPUs.
Lex Fridman (1:32:10.000)
So it can do a bunch of different kinds
Mark Zuckerberg (1:32:11.680)
of machine learning applications.
Lex Fridman (1:32:15.040)
There's a cool thing on the distributed storage aspect
Lex Fridman (1:32:18.560)
and all that kind of stuff.
Lex Fridman (1:32:19.680)
So one of the applications that I think is super exciting
Mark Zuckerberg (1:32:23.040)
is translation, real time translation.
Lex Fridman (1:32:26.320)
I mentioned to you that having a conversation,
Mark Zuckerberg (1:32:29.120)
I speak Russian fluently,
Lex Fridman (1:32:30.200)
I speak English somewhat fluently,
Lex Fridman (1:32:32.360)
and having a conversation with Vladimir Putin,
Lex Fridman (1:32:34.940)
say, as a use case.
Mark Zuckerberg (1:32:36.040)
Me, as a user, coming to you as a use case.
Lex Fridman (1:32:38.480)
We both speak each other's language.
Mark Zuckerberg (1:32:42.520)
I speak Russian, he speaks English.
Lex Fridman (1:32:45.040)
How can we have that communication go well
Lex Fridman (1:32:48.000)
with the help of AI?
Lex Fridman (1:32:49.400)
I think it's such a beautiful and a powerful application
Mark Zuckerberg (1:32:52.440)
of AI to connect the world,
Lex Fridman (1:32:54.720)
that bridge the gap, not necessarily between me and Putin,
Lex Fridman (1:32:57.560)
but people that don't have that shared language.
Lex Fridman (1:33:01.960)
Can you just speak about your vision with translation?
Mark Zuckerberg (1:33:04.120)
Because I think that's a really exciting application.
Lex Fridman (1:33:06.680)
If you're trying to help people connect
Mark Zuckerberg (1:33:08.000)
all around the world,
Lex Fridman (1:33:09.400)
a lot of content is produced in one language
Lex Fridman (1:33:11.600)
and people in all these other places are interested in it.
Lex Fridman (1:33:14.720)
So being able to translate that
Mark Zuckerberg (1:33:17.720)
just unlocks a lot of value on a day to day basis.
Lex Fridman (1:33:20.560)
I mean, so the kind of AI around translation is interesting
Mark Zuckerberg (1:33:24.400)
because it's gone through a bunch of iterations.
Lex Fridman (1:33:27.880)
But the basic state of the art
Mark Zuckerberg (1:33:29.680)
is that you don't wanna go through
Lex Fridman (1:33:33.560)
different kind of intermediate symbolic
Mark Zuckerberg (1:33:38.800)
representations of language or something like that.
Lex Fridman (1:33:42.520)
You basically wanna be able to map the concepts
Lex Fridman (1:33:46.920)
and basically go directly from one language to another.
Lex Fridman (1:33:49.300)
And you just can train bigger and bigger models
Mark Zuckerberg (1:33:53.040)
in order to be able to do that.
Lex Fridman (1:33:54.120)
And that's where the research supercluster comes in
Mark Zuckerberg (1:33:58.160)
is basically a lot of the trend in machine learning
Lex Fridman (1:34:01.080)
is just you're building bigger and bigger models
Lex Fridman (1:34:03.400)
and you just need a lot of computation to train them.
Lex Fridman (1:34:05.700)
So it's not that like the translation would run
Mark Zuckerberg (1:34:08.360)
on the supercomputer, the training of the model,
Lex Fridman (1:34:12.080)
which could have billions or trillions of examples
Mark Zuckerberg (1:34:15.800)
of just basically that.
Lex Fridman (1:34:19.080)
You're training models on this supercluster
Mark Zuckerberg (1:34:22.360)
in days or weeks that might take a much longer period of time
Lex Fridman (1:34:27.120)
on a smaller cluster.
Lex Fridman (1:34:28.120)
So it just wouldn't be practical for most teams to do.
Lex Fridman (1:34:30.200)
But the translation work,
Mark Zuckerberg (1:34:34.560)
we're basically getting from being able to go
Lex Fridman (1:34:38.160)
between about a hundred languages seamlessly today
Mark Zuckerberg (1:34:42.280)
to being able to go to about 300 languages in the near term.
Lex Fridman (1:34:46.740)
So from any language to any other language.
Mark Zuckerberg (1:34:48.720)
Yeah.
Lex Fridman (1:34:49.560)
And part of the issue when you get closer to more languages
Mark Zuckerberg (1:34:53.880)
is some of these get to be pretty,
Lex Fridman (1:34:59.840)
not very popular languages, right?
Mark Zuckerberg (1:35:01.920)
Where there isn't that much content in them.
Lex Fridman (1:35:04.280)
So you end up having less data
Lex Fridman (1:35:07.280)
and you need to kind of use a model that you've built up
Lex Fridman (1:35:10.960)
around other examples.
Lex Fridman (1:35:12.080)
And this is one of the big questions around AI
Lex Fridman (1:35:14.040)
is like how generalizable can things be?
Lex Fridman (1:35:16.680)
And that I think is one of the things
Lex Fridman (1:35:18.760)
that's just kind of exciting here
Mark Zuckerberg (1:35:19.800)
from a technical perspective.
Lex Fridman (1:35:21.300)
But capturing, we talked about this with the metaverse,
Mark Zuckerberg (1:35:23.800)
capturing the magic of human to human interaction.
Lex Fridman (1:35:26.440)
So me and Putin, okay.
Mark Zuckerberg (1:35:29.180)
Again, this is therapy session.
Lex Fridman (1:35:30.020)
I mean, it's a tough example
Mark Zuckerberg (1:35:31.080)
because you actually both speak Russian and English.
Lex Fridman (1:35:33.360)
No, but that's.
Lex Fridman (1:35:34.200)
But in the future.
Lex Fridman (1:35:35.020)
I see it as a touring test of a kind
Mark Zuckerberg (1:35:37.740)
because we would both like to have an AI that improves
Lex Fridman (1:35:40.440)
because I don't speak Russian that well.
Mark Zuckerberg (1:35:42.240)
He doesn't speak English that well.
Lex Fridman (1:35:44.200)
Yeah.
Mark Zuckerberg (1:35:45.040)
It would be nice to outperform our abilities
Lex Fridman (1:35:48.640)
and it sets a really nice bar
Mark Zuckerberg (1:35:50.640)
because I think AI can really help in translation
Lex Fridman (1:35:53.600)
for people that don't speak the language at all,
Lex Fridman (1:35:55.720)
but to actually capture the magic of the chemistry,
Lex Fridman (1:36:00.120)
the translation, which would make the metaverse
Mark Zuckerberg (1:36:03.240)
super immersive.
Lex Fridman (1:36:04.800)
I mean, that's exciting.
Mark Zuckerberg (1:36:05.840)
You remove the barrier of language, period.
Lex Fridman (1:36:08.700)
Yeah, so when people think about translation,
Mark Zuckerberg (1:36:11.240)
I think a lot of that is they're thinking about text to text,
Lex Fridman (1:36:14.240)
but speech to speech, I think is a whole nother thing.
Lex Fridman (1:36:17.120)
And I mean, one of the big lessons on that,
Lex Fridman (1:36:19.080)
which I was referring to before is I think early models,
Mark Zuckerberg (1:36:22.120)
it's like, all right, they take speech,
Lex Fridman (1:36:23.800)
they translate it to text,
Mark Zuckerberg (1:36:25.080)
translate the text to another language
Lex Fridman (1:36:26.680)
and then kind of output that as speech in that language.
Lex Fridman (1:36:29.400)
And you don't wanna do that.
Lex Fridman (1:36:30.440)
You just wanna be able to go directly from speech
Mark Zuckerberg (1:36:32.260)
in one language to speech in another language
Lex Fridman (1:36:34.180)
and build up the models to do that.
Lex Fridman (1:36:36.400)
And I mean, I think one of the,
Lex Fridman (1:36:39.140)
there have been,
Mark Zuckerberg (1:36:40.760)
when you look at the progress in machine learning,
Lex Fridman (1:36:42.860)
there have been big advances in the techniques,
Mark Zuckerberg (1:36:47.240)
some of the advances in self supervised learning,
Lex Fridman (1:36:51.560)
which I know you talked to Jan about
Lex Fridman (1:36:52.920)
and he's like one of the leading thinkers in this area.
Lex Fridman (1:36:55.320)
I just think that that stuff is really exciting,
Lex Fridman (1:36:57.560)
but then you couple that with the ability
Lex Fridman (1:36:59.840)
to just throw larger and larger amounts of compute
Mark Zuckerberg (1:37:02.480)
at training these models.
Lex Fridman (1:37:04.000)
And you can just do a lot of things
Mark Zuckerberg (1:37:05.600)
that were harder to do before.
Lex Fridman (1:37:09.360)
But we're asking more of our systems too, right?
Lex Fridman (1:37:12.960)
So if you think about the applications
Lex Fridman (1:37:14.880)
that we're gonna need for the metaverse,
Mark Zuckerberg (1:37:18.400)
or think about it, okay,
Lex Fridman (1:37:19.400)
so let's talk about AR here for a second.
Mark Zuckerberg (1:37:21.480)
You're gonna have these glasses,
Lex Fridman (1:37:23.140)
they're gonna look hopefully
Mark Zuckerberg (1:37:24.760)
like a normal ish looking pair of glasses,
Lex Fridman (1:37:28.140)
but they're gonna be able to put holograms in the world
Lex Fridman (1:37:31.240)
and intermix virtual and physical objects in your scene.
Lex Fridman (1:37:35.980)
And one of the things that's gonna be unique about this
Mark Zuckerberg (1:37:39.080)
compared to every other computing device
Lex Fridman (1:37:41.240)
that you've had before,
Mark Zuckerberg (1:37:42.580)
is that this is gonna be the first computing device
Lex Fridman (1:37:45.160)
that has all the same signals
Mark Zuckerberg (1:37:47.560)
about what's going on around you that you have.
Lex Fridman (1:37:49.480)
Right, so your phone,
Mark Zuckerberg (1:37:50.480)
you can have it take a photo or a video,
Lex Fridman (1:37:54.140)
but I mean, these glasses are gonna,
Mark Zuckerberg (1:37:56.600)
whenever you activate them,
Lex Fridman (1:37:57.480)
they're gonna be able to see what you see
Mark Zuckerberg (1:37:59.000)
from your perspective,
Lex Fridman (1:38:00.240)
they're gonna be able to hear what you hear
Mark Zuckerberg (1:38:01.540)
because the microphones and all that
Lex Fridman (1:38:03.440)
are gonna be right around where your ears are.
Lex Fridman (1:38:05.800)
So you're gonna want an AI assistant,
Lex Fridman (1:38:08.160)
that's a new kind of AI assistant
Mark Zuckerberg (1:38:10.240)
that can basically help you process the world
Lex Fridman (1:38:13.880)
from this first person perspective
Mark Zuckerberg (1:38:17.040)
or from the perspective that you have.
Lex Fridman (1:38:18.580)
And the utility of that is gonna be huge,
Lex Fridman (1:38:21.800)
but the kinds of AI models that we're gonna need
Lex Fridman (1:38:25.440)
are going to be just,
Mark Zuckerberg (1:38:28.640)
I don't know, there's a lot that we're gonna need
Lex Fridman (1:38:30.040)
to basically make advances in.
Lex Fridman (1:38:31.840)
But I mean, but that's why I think these concepts
Lex Fridman (1:38:33.760)
of the metaverse and the advances in AI
Mark Zuckerberg (1:38:36.600)
are so fundamentally interlinked
Lex Fridman (1:38:40.200)
that I mean, they're kind of enabling each other.
Mark Zuckerberg (1:38:42.880)
Yeah, like the world builder is a really cool idea.
Lex Fridman (1:38:45.440)
Like you can be like a Bob Ross,
Mark Zuckerberg (1:38:47.240)
like I'm gonna put a little tree right here.
Lex Fridman (1:38:49.120)
Yeah.
Mark Zuckerberg (1:38:49.940)
I need a little tree, it's missing a little tree.
Lex Fridman (1:38:51.200)
And then, but at scale,
Mark Zuckerberg (1:38:52.960)
like enriching your experience in all kinds of ways.
Lex Fridman (1:38:55.680)
You mentioned the assistant too,
Mark Zuckerberg (1:38:56.960)
that's really interesting how you can have AI assistants
Lex Fridman (1:39:00.020)
helping you out on different levels
Mark Zuckerberg (1:39:01.640)
of sort of intimacy of communication.
Lex Fridman (1:39:04.000)
It could be just like scheduling
Mark Zuckerberg (1:39:05.480)
or it could be like almost like therapy.
Lex Fridman (1:39:08.120)
Clearly I need some.
Lex Fridman (1:39:09.880)
So let me ask you,
Lex Fridman (1:39:11.160)
you're one of the most successful people ever.
Mark Zuckerberg (1:39:14.000)
You've built an incredible company
Lex Fridman (1:39:16.240)
that has a lot of impact.
Lex Fridman (1:39:18.080)
What advice do you have for young people today?
Lex Fridman (1:39:23.120)
How to live a life they can be proud of?
Lex Fridman (1:39:25.480)
How to build something that can have a big positive impact
Lex Fridman (1:39:30.420)
on the world?
Mark Zuckerberg (1:39:31.260)
Well, let's break that down.
Lex Fridman (1:39:37.460)
Cause I think you proud of, have a big positive impact.
Mark Zuckerberg (1:39:41.300)
Well, you're actually listening.
Lex Fridman (1:39:42.320)
And how to live your life
Mark Zuckerberg (1:39:43.880)
are actually three different things that I think,
Lex Fridman (1:39:47.580)
I mean, they could line up,
Lex Fridman (1:39:48.900)
but, and also like what age of people are you talking to?
Lex Fridman (1:39:52.460)
Cause I mean, I can like.
Mark Zuckerberg (1:39:53.500)
High school and college.
Lex Fridman (1:39:54.700)
So you don't really know what you're doing,
Lex Fridman (1:39:56.500)
but your dream big.
Lex Fridman (1:39:58.220)
And you really have a chance to do something unprecedented.
Mark Zuckerberg (1:40:02.280)
Yeah.
Lex Fridman (1:40:04.180)
So I guess just to.
Mark Zuckerberg (1:40:05.020)
Also for people my age.
Lex Fridman (1:40:06.300)
Okay, so let's maybe start with the kind of most
Mark Zuckerberg (1:40:09.620)
philosophical and abstract version of this.
Lex Fridman (1:40:12.060)
Every night when I put my daughters to bed,
Mark Zuckerberg (1:40:16.220)
we go through this thing and like,
Lex Fridman (1:40:20.380)
they call it the good night things.
Mark Zuckerberg (1:40:21.740)
Cause we're basically what we talk about at night.
Lex Fridman (1:40:25.440)
And I just, I go through them.
Mark Zuckerberg (1:40:29.700)
Sounds like a good show.
Lex Fridman (1:40:31.460)
The good night things.
Mark Zuckerberg (1:40:32.700)
Yeah.
Lex Fridman (1:40:33.540)
Priscilla's always asking, she's like,
Lex Fridman (1:40:34.360)
can I get good night things?
Lex Fridman (1:40:35.200)
Like, I don't know.
Mark Zuckerberg (1:40:36.040)
You go to bed too early.
Lex Fridman (1:40:37.100)
But it's,
Lex Fridman (1:40:41.580)
but I basically go through with Max and Augie,
Lex Fridman (1:40:46.340)
what are the things that are most important in life?
Mark Zuckerberg (1:40:48.940)
Right.
Lex Fridman (1:40:49.780)
That I just, it's like, what do I want them to remember
Lex Fridman (1:40:51.560)
and just have like really ingrained in them as they grow up?
Lex Fridman (1:40:53.940)
And it's health, right?
Mark Zuckerberg (1:40:56.740)
Making sure that you take care of yourself
Lex Fridman (1:40:58.800)
and keep yourself in good shape,
Lex Fridman (1:41:00.700)
loving friends and family, right?
Lex Fridman (1:41:02.940)
Because having the relationships,
Mark Zuckerberg (1:41:05.380)
the family and making time for friends,
Lex Fridman (1:41:08.700)
I think is perhaps one of the most important things.
Lex Fridman (1:41:13.820)
And then the third is maybe a little more amorphous,
Lex Fridman (1:41:16.040)
but it is something that you're excited about for the future.
Lex Fridman (1:41:19.420)
And when I'm talking to a four year old,
Lex Fridman (1:41:21.300)
often I'll ask her what she's excited about
Mark Zuckerberg (1:41:23.500)
for tomorrow or the week ahead.
Lex Fridman (1:41:25.220)
But I think for most people, it's really hard.
Mark Zuckerberg (1:41:29.580)
I mean, the world is a heavy place.
Lex Fridman (1:41:31.440)
And I think like the way that we navigate it
Mark Zuckerberg (1:41:34.780)
is that we have things that we're looking forward to.
Lex Fridman (1:41:37.340)
So whether it is building AR glasses for the future
Mark Zuckerberg (1:41:41.620)
or being able to celebrate my 10 year wedding anniversary
Lex Fridman (1:41:45.520)
with my wife that's coming up,
Mark Zuckerberg (1:41:47.380)
it's like, I think people,
Lex Fridman (1:41:48.620)
you know, you have things that you're looking forward to.
Mark Zuckerberg (1:41:51.860)
Or for the girls, it's often I want to see mom
Lex Fridman (1:41:53.860)
in the morning, right?
Mark Zuckerberg (1:41:54.700)
It's just, but it's like that's a really critical thing.
Lex Fridman (1:41:57.020)
And then the last thing is I ask them every day,
Lex Fridman (1:42:00.340)
what did you do today to help someone?
Lex Fridman (1:42:04.340)
Because I just think that that's a really critical thing
Mark Zuckerberg (1:42:07.140)
is like, it's easy to kind of get caught up in yourself
Lex Fridman (1:42:10.740)
and kind of stuff that's really far down the road,
Lex Fridman (1:42:14.300)
but like, did you do something just concrete today
Lex Fridman (1:42:17.520)
to help someone?
Mark Zuckerberg (1:42:18.360)
And, you know, it can just be as simple as, okay, yeah,
Lex Fridman (1:42:21.060)
I helped set the table for lunch, right?
Mark Zuckerberg (1:42:23.420)
Or, you know, this other kid in our school
Lex Fridman (1:42:26.440)
was having a hard time with something
Lex Fridman (1:42:27.900)
and I like helped explain it to him.
Lex Fridman (1:42:29.260)
But in that those are, that's sort of like,
Mark Zuckerberg (1:42:32.940)
if you were to boil down my overall life philosophy
Lex Fridman (1:42:36.080)
into what I try to impart to my kids,
Mark Zuckerberg (1:42:40.140)
those are the things that I think are really important.
Lex Fridman (1:42:43.000)
So, okay, so let's say college.
Lex Fridman (1:42:44.320)
So if you're a graduate in college,
Lex Fridman (1:42:45.860)
probably more practical advice, I'm always very focused
Mark Zuckerberg (1:42:52.340)
on people.
Lex Fridman (1:42:53.180)
And I think the most important decision
Mark Zuckerberg (1:42:57.140)
you're probably gonna make if you're in college
Lex Fridman (1:42:59.320)
is who you surround yourself with,
Mark Zuckerberg (1:43:01.620)
because you become like the people
Lex Fridman (1:43:02.960)
you surround yourself with.
Lex Fridman (1:43:04.620)
And I sort of have this hiring heuristic at Metta,
Lex Fridman (1:43:09.620)
which is that I will only hire someone to work for me
Mark Zuckerberg (1:43:13.620)
if I could see myself working for them.
Lex Fridman (1:43:17.220)
Not necessarily that I want them to run the company
Mark Zuckerberg (1:43:19.020)
because I like my job, but in an alternate universe,
Lex Fridman (1:43:22.420)
if it was their company and I was looking
Lex Fridman (1:43:23.980)
to go work somewhere, would I be happy to work for them?
Lex Fridman (1:43:27.060)
And I think that that's a helpful heuristic
Mark Zuckerberg (1:43:31.220)
to help balance, you know,
Lex Fridman (1:43:33.060)
when you're building something like this,
Mark Zuckerberg (1:43:33.900)
there's a lot of pressure to, you know,
Lex Fridman (1:43:36.060)
you wanna build out your team,
Mark Zuckerberg (1:43:37.380)
because there's a lot of stuff that you need to get done.
Lex Fridman (1:43:39.620)
And everyone always says, don't compromise on quality,
Lex Fridman (1:43:41.860)
but there's this question of, okay,
Lex Fridman (1:43:42.940)
well, how do you know that someone is good enough?
Lex Fridman (1:43:44.020)
And I think my answer is, I would want someone
Lex Fridman (1:43:46.860)
to be on my team if I would work for them.
Lex Fridman (1:43:50.660)
But I think it's actually a pretty similar answer
Lex Fridman (1:43:53.300)
to like, if you were choosing friends or a partner
Mark Zuckerberg (1:43:58.300)
or something like that.
Lex Fridman (1:43:59.460)
So when you're kind of in college,
Mark Zuckerberg (1:44:01.900)
trying to figure out what your circle is gonna be,
Lex Fridman (1:44:03.580)
trying to figure out, you know,
Mark Zuckerberg (1:44:04.420)
you're evaluating data,
Lex Fridman (1:44:05.260)
your circle is gonna be trying to figure out, you know,
Mark Zuckerberg (1:44:07.620)
you're evaluating different job opportunities.
Lex Fridman (1:44:09.300)
Who are the people, even if they're gonna be peers
Mark Zuckerberg (1:44:12.980)
in what you're doing,
Lex Fridman (1:44:14.420)
who are the people who in an alternate university,
Mark Zuckerberg (1:44:17.020)
you would wanna work for them,
Lex Fridman (1:44:18.700)
because you think you're gonna learn a lot from them,
Mark Zuckerberg (1:44:20.420)
because they know, because they are kind of values aligned
Lex Fridman (1:44:24.180)
on the things that you care about,
Lex Fridman (1:44:25.500)
and they're gonna like, and they're gonna push you,
Lex Fridman (1:44:28.260)
but also they know different things
Lex Fridman (1:44:29.500)
and have different experiences
Lex Fridman (1:44:30.660)
that are kind of more of what you wanna become like
Mark Zuckerberg (1:44:32.940)
over time.
Lex Fridman (1:44:33.780)
But I don't know, I think probably people are too,
Mark Zuckerberg (1:44:37.020)
in general, objective focused,
Lex Fridman (1:44:39.100)
and maybe not focused enough on the connections
Lex Fridman (1:44:42.700)
and the people who they're basically building relationships
Lex Fridman (1:44:46.620)
with.
Mark Zuckerberg (1:44:47.460)
I don't know what it says about me,
Lex Fridman (1:44:48.300)
but my place in Austin now has seven legged robots.
Lex Fridman (1:44:53.540)
So I'm surrounded myself by robots,
Lex Fridman (1:44:55.420)
which is probably something I should look into.
Lex Fridman (1:44:59.140)
What kind of world would you like to see your daughters
Lex Fridman (1:45:02.500)
grow up in, even after you're gone?
Mark Zuckerberg (1:45:09.300)
Well, I think one of the promises of all the stuff
Lex Fridman (1:45:11.500)
that is getting built now is that it can be a world
Mark Zuckerberg (1:45:15.580)
where more people can just live out their imagination.
Lex Fridman (1:45:21.780)
One of my favorite quotes,
Mark Zuckerberg (1:45:23.420)
I think it was attributed to Picasso,
Lex Fridman (1:45:25.140)
it's that all children are artists,
Lex Fridman (1:45:26.780)
and the challenge is how do you remain one
Lex Fridman (1:45:28.340)
when you grow up?
Lex Fridman (1:45:29.580)
And if you have kids, this is pretty clear,
Lex Fridman (1:45:33.620)
I mean, they just have wonderful imaginations.
Lex Fridman (1:45:36.260)
And part of what I think is gonna be great
Lex Fridman (1:45:38.980)
about the creator economy and the metaverse
Lex Fridman (1:45:41.380)
and all this stuff is this notion around
Lex Fridman (1:45:44.740)
that a lot more people in the future
Mark Zuckerberg (1:45:46.300)
are gonna get to work doing creative stuff
Lex Fridman (1:45:49.020)
than what I think today we would just consider
Mark Zuckerberg (1:45:51.420)
traditional labor or service.
Lex Fridman (1:45:53.740)
And I think that that's awesome.
Lex Fridman (1:45:56.300)
And that's a lot of what people are here to do
Lex Fridman (1:46:00.140)
is collaborate together, work together,
Mark Zuckerberg (1:46:03.140)
think of things that you wanna build and go do it.
Lex Fridman (1:46:06.420)
And I don't know, one of the things
Mark Zuckerberg (1:46:08.540)
that I just think is striking,
Lex Fridman (1:46:09.380)
so I teach my daughters some basic coding with Scratch.
Mark Zuckerberg (1:46:13.660)
I mean, they're still obviously really young,
Lex Fridman (1:46:15.420)
but I think of coding as building,
Mark Zuckerberg (1:46:18.340)
where it's like when I'm coding,
Lex Fridman (1:46:19.780)
I'm building something that I want to exist.
Lex Fridman (1:46:22.300)
But my youngest daughter, she's very musical
Lex Fridman (1:46:27.980)
and pretty artistic and she thinks about coding as art.
Mark Zuckerberg (1:46:32.820)
She calls it code art, not the code,
Lex Fridman (1:46:35.540)
but the output of what she is making.
Mark Zuckerberg (1:46:37.540)
It's like, she's just very interesting visually
Lex Fridman (1:46:39.340)
in what she can kind of output and how it can move around.
Lex Fridman (1:46:42.580)
And do we need to fix that?
Lex Fridman (1:46:45.020)
Are we good?
Lex Fridman (1:46:45.860)
What happened?
Lex Fridman (1:46:47.460)
Do we have to clap, Alexa?
Mark Zuckerberg (1:46:49.460)
Yeah, so I was just talking about Augie and her code art,
Lex Fridman (1:46:53.020)
but I mean, to me, this is like a beautiful thing, right?
Mark Zuckerberg (1:46:56.540)
The notion that like for me,
Lex Fridman (1:46:58.700)
coding was this functional thing and I enjoyed it.
Lex Fridman (1:47:01.380)
And it like helped build something utilitarian,
Lex Fridman (1:47:04.620)
but that for the next generation of people,
Mark Zuckerberg (1:47:06.820)
it will be even more an expression
Lex Fridman (1:47:10.460)
of their kind of imagination and artistic sense
Mark Zuckerberg (1:47:14.900)
for what they want to exist.
Lex Fridman (1:47:15.940)
So I don't know if that happens,
Mark Zuckerberg (1:47:17.620)
if we can help bring about this world
Lex Fridman (1:47:20.420)
where a lot more people can,
Mark Zuckerberg (1:47:23.580)
that that's like their existence going forward
Lex Fridman (1:47:25.900)
is being able to basically create
Lex Fridman (1:47:28.620)
and live out all these different kinds of art.
Lex Fridman (1:47:32.940)
I just think that that's like a beautiful
Lex Fridman (1:47:34.420)
and wonderful thing and will be very freeing for humanity
Lex Fridman (1:47:37.940)
to spend more of our time on the things that matter to us.
Mark Zuckerberg (1:47:40.420)
Yeah, allow more and more people to express their art
Lex Fridman (1:47:43.140)
in the full meaning of that word.
Mark Zuckerberg (1:47:45.140)
That's a beautiful vision.
Lex Fridman (1:47:46.900)
We mentioned that you are mortal.
Lex Fridman (1:47:50.260)
Are you afraid of death?
Lex Fridman (1:47:51.900)
Do you think about your mortality?
Lex Fridman (1:47:56.300)
And are you afraid of it?
Lex Fridman (1:48:01.220)
You didn't sign up for this on a podcast, did you?
Mark Zuckerberg (1:48:03.100)
No, I mean, it's an interesting question.
Lex Fridman (1:48:07.060)
I mean, I'm definitely aware of it.
Mark Zuckerberg (1:48:08.700)
I do a fair amount of like extreme sport type stuff.
Lex Fridman (1:48:13.700)
So like, so I'm definitely aware of it.
Lex Fridman (1:48:19.700)
And you're flirting with it a bit.
Lex Fridman (1:48:22.220)
I train hard.
Mark Zuckerberg (1:48:23.700)
I mean, so it's like, if I'm gonna go out
Lex Fridman (1:48:25.100)
in like a 15 foot wave.
Mark Zuckerberg (1:48:27.700)
Go out big.
Lex Fridman (1:48:28.540)
Well, then it's like, all right,
Mark Zuckerberg (1:48:29.700)
I'll make sure we have the right safety gear
Lex Fridman (1:48:31.540)
and like make sure that I'm like used to that spot
Lex Fridman (1:48:34.540)
and all that stuff.
Lex Fridman (1:48:35.580)
But like, but you know, I mean, you.
Mark Zuckerberg (1:48:37.740)
The risk is still there.
Lex Fridman (1:48:38.940)
You take some head blows along the way.
Mark Zuckerberg (1:48:40.780)
Yes, but definitely aware of it.
Lex Fridman (1:48:45.340)
Definitely would like to stay safe.
Mark Zuckerberg (1:48:48.020)
I have a lot of stuff that I want to build and want to.
Lex Fridman (1:48:52.020)
Does it freak you out that it's finite though?
Mark Zuckerberg (1:48:55.460)
That there's a deadline when it's all over
Lex Fridman (1:48:59.180)
and that there'll be a time when your daughters are around
Lex Fridman (1:49:01.500)
and you're gone?
Lex Fridman (1:49:03.060)
I don't know.
Mark Zuckerberg (1:49:03.900)
That doesn't freak me out.
Lex Fridman (1:49:04.940)
I think, I don't know.
Mark Zuckerberg (1:49:09.780)
Constraints are helpful.
Lex Fridman (1:49:16.220)
Yeah.
Mark Zuckerberg (1:49:17.300)
Yeah, the finiteness makes ice cream
Lex Fridman (1:49:20.260)
taste more delicious somehow.
Mark Zuckerberg (1:49:21.740)
The fact that it's gonna be over.
Lex Fridman (1:49:23.140)
There's something about that with the metaverse too.
Mark Zuckerberg (1:49:25.700)
You want, we talked about this identity earlier,
Lex Fridman (1:49:28.500)
like having just one, like NFTs.
Mark Zuckerberg (1:49:30.260)
There's something powerful about the constraint
Lex Fridman (1:49:34.340)
of finiteness or uniqueness.
Mark Zuckerberg (1:49:36.900)
That this moment is singular in history.
Lex Fridman (1:49:39.740)
But I mean, a lot of,
Mark Zuckerberg (1:49:41.060)
as you go through different waves of technology,
Lex Fridman (1:49:42.700)
I think a lot of what is interesting is
Lex Fridman (1:49:44.340)
what becomes in practice infinite
Lex Fridman (1:49:48.020)
or kind of there can be many, many of a thing
Lex Fridman (1:49:51.500)
and then what ends up still being constrained.
Lex Fridman (1:49:53.660)
So the metaverse should hopefully allow
Mark Zuckerberg (1:50:00.340)
a very large number or maybe in practice,
Lex Fridman (1:50:04.220)
hopefully close to an infinite amount of expression
Lex Fridman (1:50:06.860)
and worlds, but we'll still only have
Lex Fridman (1:50:09.700)
a finite amount of time.
Mark Zuckerberg (1:50:11.220)
Yes.
Lex Fridman (1:50:12.060)
I think living longer I think is good.
Lex Fridman (1:50:18.020)
And obviously all of my, our philanthropic work is,
Lex Fridman (1:50:21.700)
it's not focused on longevity,
Lex Fridman (1:50:23.380)
but it is focused on trying to achieve
Lex Fridman (1:50:25.940)
what I think is a possible goal in this century,
Mark Zuckerberg (1:50:29.620)
which is to be able to cure, prevent
Lex Fridman (1:50:31.020)
or manage all diseases.
Lex Fridman (1:50:33.460)
So I certainly think people kind of getting sick
Lex Fridman (1:50:36.140)
and dying is a bad thing because,
Lex Fridman (1:50:37.740)
and I'm dedicating almost all of my capital
Lex Fridman (1:50:40.260)
towards advancing research in that area to push on that,
Mark Zuckerberg (1:50:44.500)
which I mean, we could do a whole,
Lex Fridman (1:50:45.460)
another one of these podcasts about that
Mark Zuckerberg (1:50:46.940)
because that's a fascinating topic.
Lex Fridman (1:50:49.660)
I mean, this is with your wife Priscilla Chan,
Mark Zuckerberg (1:50:51.740)
you formed the Chan Zuckerberg Initiative,
Lex Fridman (1:50:54.140)
gave away 99% or pledged to give away 99%
Mark Zuckerberg (1:50:57.020)
of Facebook non meta shares.
Lex Fridman (1:50:59.260)
I mean, like you said, we could talk forever
Mark Zuckerberg (1:51:01.980)
about all the exciting things you're working on there,
Lex Fridman (1:51:06.100)
including the sort of moonshot of eradicating disease
Mark Zuckerberg (1:51:11.300)
by the mid century marker.
Lex Fridman (1:51:13.260)
I don't actually know if you're gonna ever eradicate it,
Lex Fridman (1:51:15.420)
but I think you can get to a point where you
Lex Fridman (1:51:17.980)
can either cure things that happened, right?
Lex Fridman (1:51:20.900)
So people get diseases, but you can cure them.
Lex Fridman (1:51:22.940)
Prevent is probably closest to eradication
Mark Zuckerberg (1:51:25.620)
or just be able to manage as sort of like ongoing things
Lex Fridman (1:51:28.860)
that are not gonna ruin your life.
Lex Fridman (1:51:33.260)
And I think that that's possible.
Lex Fridman (1:51:34.300)
I think saying that there's gonna be no disease at all
Mark Zuckerberg (1:51:37.060)
probably is not possible within the next several decades.
Lex Fridman (1:51:41.540)
Basic thing is increase the quality of life
Lex Fridman (1:51:44.300)
and maybe keep the finiteness
Lex Fridman (1:51:46.820)
because it makes everything taste more delicious.
Mark Zuckerberg (1:51:50.180)
Maybe that's just being a romantic 20th century human.
Lex Fridman (1:51:54.740)
Maybe, but I mean, but it was an intentional decision
Mark Zuckerberg (1:51:57.140)
to not focus on our philanthropy on like explicitly
Lex Fridman (1:52:01.780)
on longevity or living forever.
Mark Zuckerberg (1:52:03.460)
Yes.
Lex Fridman (1:52:06.980)
If at the moment of your death, and by the way,
Mark Zuckerberg (1:52:09.060)
I like that the lights went out
Lex Fridman (1:52:11.540)
when we started talking about death.
Mark Zuckerberg (1:52:13.380)
You get to meet God.
Lex Fridman (1:52:14.220)
It does make it a lot more dramatic.
Mark Zuckerberg (1:52:15.660)
It does.
Lex Fridman (1:52:17.940)
I should get closer to the mic.
Mark Zuckerberg (1:52:19.740)
At the moment of your death, you get to meet God
Lex Fridman (1:52:23.020)
and you get to ask one question.
Lex Fridman (1:52:26.140)
What question would you like to ask?
Lex Fridman (1:52:29.820)
Or maybe a whole conversation.
Mark Zuckerberg (1:52:31.180)
I don't know.
Lex Fridman (1:52:32.020)
It's up to you.
Mark Zuckerberg (1:52:32.860)
It's more dramatic when it's just one question.
Lex Fridman (1:52:37.100)
Well, if it's only one question and I died,
Mark Zuckerberg (1:52:42.980)
I would just wanna know that Priscilla and my family,
Lex Fridman (1:52:48.020)
like if they were gonna be okay.
Mark Zuckerberg (1:52:50.900)
That might depend on the circumstances of my death.
Lex Fridman (1:52:54.620)
But I think that in most circumstances that I can think of,
Mark Zuckerberg (1:52:58.060)
that's probably the main thing that I would care about.
Lex Fridman (1:53:01.100)
Yeah, I think God will hear that question and be like,
Mark Zuckerberg (1:53:02.820)
all right, fine, you get in.
Lex Fridman (1:53:04.260)
That's the right question to ask.
Lex Fridman (1:53:06.700)
Is it?
Lex Fridman (1:53:07.540)
I don't know.
Mark Zuckerberg (1:53:08.380)
The humility and selfishness.
Lex Fridman (1:53:09.580)
All right, you're in.
Mark Zuckerberg (1:53:10.820)
I mean, but well, maybe.
Lex Fridman (1:53:14.540)
They're gonna be fine.
Mark Zuckerberg (1:53:15.380)
Don't worry, you're in.
Lex Fridman (1:53:16.220)
Okay, but I mean, one of the things that I think
Mark Zuckerberg (1:53:18.220)
I struggle with at least is on the one hand,
Lex Fridman (1:53:22.300)
that's probably the thing that's closest to me
Lex Fridman (1:53:25.620)
and maybe the most common human experience.
Lex Fridman (1:53:29.420)
But I don't know, one of the things that I just struggle with
Mark Zuckerberg (1:53:32.380)
in terms of running this large enterprise is like,
Lex Fridman (1:53:38.100)
should the thing that I care more about
Lex Fridman (1:53:41.020)
be that responsibility?
Lex Fridman (1:53:44.860)
And I think it's shifted over time.
Mark Zuckerberg (1:53:49.300)
I mean, like before I really had a family
Lex Fridman (1:53:52.060)
that was like the only thing I cared about.
Lex Fridman (1:53:53.860)
And at this point, I mean, I care deeply about it,
Lex Fridman (1:53:59.980)
but yeah, I think that that's not as obvious of a question.
Mark Zuckerberg (1:54:06.060)
Yeah, we humans are weird.
Lex Fridman (1:54:07.860)
You get this ability to impact millions of lives
Lex Fridman (1:54:12.780)
and it's definitely something, billions of lives,
Lex Fridman (1:54:15.660)
it's something you care about,
Lex Fridman (1:54:16.980)
but the weird humans that are closest to us,
Lex Fridman (1:54:21.100)
those are the ones that mean the most.
Lex Fridman (1:54:23.700)
And I suppose that's the dream of the metaverse
Lex Fridman (1:54:26.140)
is to connect, form small groups like that
Mark Zuckerberg (1:54:29.300)
where you can have those intimate relationships.
Lex Fridman (1:54:31.700)
Let me ask you the big, ridiculous.
Mark Zuckerberg (1:54:33.620)
Well, and to be able to be close,
Lex Fridman (1:54:36.940)
not just based on who you happen to be next to.
Mark Zuckerberg (1:54:39.900)
I think that's what the internet is already doing
Lex Fridman (1:54:41.980)
is allowing you to spend more of your time
Mark Zuckerberg (1:54:44.540)
not physically proximate.
Lex Fridman (1:54:46.540)
I mean, I always think when you think about the metaverse,
Mark Zuckerberg (1:54:49.940)
people ask this question about the real world.
Lex Fridman (1:54:52.140)
It's like the virtual world versus the real world.
Lex Fridman (1:54:54.860)
And it's like, no, the real world is a combination
Lex Fridman (1:54:58.180)
of the virtual world and the physical world.
Lex Fridman (1:55:00.100)
But I think over time, as we get more technology,
Lex Fridman (1:55:04.060)
the physical world is becoming less of a percent
Mark Zuckerberg (1:55:06.780)
of the real world.
Lex Fridman (1:55:08.180)
And I think that that opens up a lot of opportunities
Mark Zuckerberg (1:55:10.940)
for people, because you can work in different places.
Lex Fridman (1:55:13.460)
You can stay more close to, stay closer to people
Mark Zuckerberg (1:55:17.580)
who are in different places.
Lex Fridman (1:55:18.420)
So I think that's good.
Mark Zuckerberg (1:55:19.300)
Removing barriers of geography
Lex Fridman (1:55:21.380)
and then barriers of language.
Mark Zuckerberg (1:55:23.140)
That's a beautiful vision.
Lex Fridman (1:55:25.860)
Big, ridiculous question.
Lex Fridman (1:55:27.580)
What do you think is the meaning of life?
Lex Fridman (1:55:44.100)
I think that, well, there are probably a couple
Mark Zuckerberg (1:55:46.560)
of different ways that I would go at this.
Lex Fridman (1:55:52.020)
But I think it gets back to this last question
Mark Zuckerberg (1:55:53.900)
that we talked about, about the duality
Lex Fridman (1:55:55.380)
between you have the people around you
Mark Zuckerberg (1:55:58.660)
who you care the most about,
Lex Fridman (1:56:00.380)
and then there's like this bigger thing
Mark Zuckerberg (1:56:03.060)
that maybe you're building.
Lex Fridman (1:56:05.820)
And I think that in my own life, I mean,
Mark Zuckerberg (1:56:07.340)
I sort of think about this tension,
Lex Fridman (1:56:09.360)
but I mean, it's like, I started this whole company
Lex Fridman (1:56:11.860)
and my life's work is around human connection.
Lex Fridman (1:56:15.100)
So I think it's intellectually probably the thing
Mark Zuckerberg (1:56:22.100)
that I go to first is just that human connection
Lex Fridman (1:56:27.720)
is the meaning.
Lex Fridman (1:56:29.560)
And I mean, I think that it's a thing
Lex Fridman (1:56:31.160)
that our society probably systematically undervalues.
Mark Zuckerberg (1:56:36.860)
I mean, I just remember when I was growing up
Lex Fridman (1:56:39.280)
and in school, it's like, do your homework
Lex Fridman (1:56:43.140)
and then go play with your friends after.
Lex Fridman (1:56:45.200)
And it's like, no, well, what if playing
Lex Fridman (1:56:47.020)
with your friends is the point?
Lex Fridman (1:56:50.180)
That sounds like an argument your daughter would make.
Mark Zuckerberg (1:56:52.340)
Well, I mean, I don't know, I just think it's interesting.
Lex Fridman (1:56:54.620)
Homework doesn't even matter, man.
Mark Zuckerberg (1:56:56.340)
Well, I think it's interesting because it's,
Lex Fridman (1:56:58.540)
and people, I think people tend to think
Mark Zuckerberg (1:57:02.700)
about that stuff as wasting time,
Lex Fridman (1:57:05.180)
or that's like what you do in the free time that you have.
Lex Fridman (1:57:08.100)
But like, what if that's actually the point?
Lex Fridman (1:57:11.140)
So that's one.
Lex Fridman (1:57:12.580)
But here's maybe a different way of counting out this,
Lex Fridman (1:57:14.760)
which is maybe more like religious in nature.
Mark Zuckerberg (1:57:17.880)
I mean, I always like,
Lex Fridman (1:57:22.200)
there's a rabbi who I've studied with
Mark Zuckerberg (1:57:25.400)
who kind of gave me this,
Lex Fridman (1:57:27.980)
we were talking through Genesis and the Bible and the Torah
Lex Fridman (1:57:31.820)
and they're basically walking through,
Lex Fridman (1:57:36.100)
it's like, okay, you go through the seven days of creation
Lex Fridman (1:57:40.740)
and it's basically, it's like,
Lex Fridman (1:57:45.660)
why does the Bible start there?
Mark Zuckerberg (1:57:48.100)
Right, it's like it could have started anywhere,
Lex Fridman (1:57:49.460)
right, in terms of like how to live.
Lex Fridman (1:57:52.000)
But basically it starts with talking about
Lex Fridman (1:57:54.620)
how God created people in his, her image.
Lex Fridman (1:58:00.560)
But the Bible starts by talking about
Lex Fridman (1:58:02.720)
how God created everything.
Lex Fridman (1:58:04.740)
So I actually think that there's like a compelling argument
Lex Fridman (1:58:11.240)
that I think I've always just found meaningful
Lex Fridman (1:58:12.980)
and inspiring that a lot of the point
Lex Fridman (1:58:18.400)
of what sort of religion has been telling us
Mark Zuckerberg (1:58:22.980)
that we should do is to create and build things.
Lex Fridman (1:58:30.060)
So these things are not necessarily at odds.
Mark Zuckerberg (1:58:32.100)
I mean, I think like, I mean, that's,
Lex Fridman (1:58:34.780)
and I think probably to some degree
Mark Zuckerberg (1:58:36.100)
you'd expect me to say something like this
Lex Fridman (1:58:37.720)
because I've dedicated my life to creating things
Mark Zuckerberg (1:58:39.760)
that help people connect.
Lex Fridman (1:58:40.600)
So, I mean, that's sort of the fusion of,
Mark Zuckerberg (1:58:43.780)
I mean, getting back to what we talked about earlier,
Lex Fridman (1:58:45.260)
it's, I mean, what I studied in school
Lex Fridman (1:58:46.500)
or psychology and computer science, right?
Lex Fridman (1:58:48.260)
So it's, I mean, these are like the two themes
Mark Zuckerberg (1:58:50.580)
that I care about, but I don't know for me,
Lex Fridman (1:58:54.460)
that's kind of what I think about, that's what matters.
Mark Zuckerberg (1:58:57.140)
To create and to love, which is the ultimate form
Lex Fridman (1:59:02.200)
of connection.
Mark Zuckerberg (1:59:03.980)
I think this is one hell of an amazing replay experience
Lex Fridman (1:59:07.220)
in the metaverse.
Lex Fridman (1:59:08.060)
So whoever is using our avatars years from now,
Lex Fridman (1:59:11.980)
I hope you had fun and thank you for talking today.
Mark Zuckerberg (1:59:14.780)
Thank you.
Lex Fridman (1:59:16.460)
Thanks for listening to this conversation
Mark Zuckerberg (1:59:18.180)
with Mark Zuckerberg.
Lex Fridman (1:59:19.500)
To support this podcast, please check out our sponsors
Mark Zuckerberg (1:59:22.060)
in the description.
Lex Fridman (1:59:23.660)
And now, let me leave you with the end of the poem, If,
Mark Zuckerberg (1:59:27.820)
by Roger Kipling.
Lex Fridman (1:59:30.820)
If you can talk with crowds and keep your virtue,
Mark Zuckerberg (1:59:34.160)
or walk with kings, nor lose the common touch,
Lex Fridman (1:59:37.820)
if neither foes nor loving friends can hurt you,
Mark Zuckerberg (1:59:41.220)
if all men count with you, but none too much.
Lex Fridman (1:59:47.100)
If you can fill the unforgiving minute
Mark Zuckerberg (1:59:49.260)
with 60 seconds worth of distance run,
Lex Fridman (1:59:52.340)
yours is the earth and everything that's in it.
Lex Fridman (1:59:56.340)
And which is more, you'll be a man, my son.
Lex Fridman (1:59:59.580)
Thank you for listening and hope to see you next time.
Mark Zuckerberg (20:00.420)
if you look at the list of what are the top apps,
Lex Fridman (20:03.460)
social apps are already number one, two, three.
Lex Fridman (20:06.460)
So that's kind of becoming a critical thing, but I don't know.
Lex Fridman (20:10.860)
I would imagine for someone like you,
Mark Zuckerberg (20:12.580)
it'll be until we get a lot of the work things dialed in.
Lex Fridman (20:17.980)
When this is just like much more adopted
Lex Fridman (20:20.860)
and clearly better than Zoom for VC,
Lex Fridman (20:24.260)
when if you're doing your coding or your writing
Mark Zuckerberg (20:27.100)
or whatever it is in VR,
Lex Fridman (20:29.420)
which it's not that far off to imagine that
Mark Zuckerberg (20:31.260)
because pretty soon you're just gonna be able
Lex Fridman (20:32.660)
to have a screen that's bigger than,
Mark Zuckerberg (20:34.220)
it'll be your ideal setup and you can bring it with you
Lex Fridman (20:36.140)
and put it on anywhere
Lex Fridman (20:37.540)
and have your kind of ideal workstation.
Lex Fridman (20:39.780)
So I think that there are a few things to work out on that,
Lex Fridman (20:42.580)
but I don't think that that's more than five years off.
Lex Fridman (20:46.940)
And then you'll get a bunch of other things
Mark Zuckerberg (20:48.120)
that like aren't even possible
Lex Fridman (20:50.220)
or you don't even think about using a phone
Lex Fridman (20:52.040)
or PC for today, like fitness, right?
Lex Fridman (20:54.440)
So, I mean, I know you're, we were talking before
Mark Zuckerberg (20:57.460)
about how you're into running
Lex Fridman (20:58.980)
and like I'm really into a lot of things
Mark Zuckerberg (21:00.780)
around fitness as well,
Lex Fridman (21:02.740)
different things in different places.
Mark Zuckerberg (21:04.100)
I got really into hydrofoiling recently
Lex Fridman (21:06.060)
and surfing and I used to fence competitively.
Mark Zuckerberg (21:12.420)
I like run.
Lex Fridman (21:13.260)
So, and you were saying that you were thinking
Mark Zuckerberg (21:14.860)
about trying different martial arts
Lex Fridman (21:16.380)
and I tried to trick you and convince you
Mark Zuckerberg (21:18.180)
into doing Brazilian Jiu Jitsu.
Lex Fridman (21:19.940)
Or you actually mentioned that that was one
Mark Zuckerberg (21:21.500)
you're curious about and I don't know.
Lex Fridman (21:23.220)
Is that a trick?
Mark Zuckerberg (21:24.140)
Yeah, I don't know.
Lex Fridman (21:26.020)
We're in the metaverse now.
Mark Zuckerberg (21:27.420)
Yeah, no, I took that seriously.
Lex Fridman (21:29.580)
I thought that that was a real suggestion.
Mark Zuckerberg (21:34.260)
That would be an amazing chance
Lex Fridman (21:36.380)
if we ever step on the mat together
Lex Fridman (21:37.780)
and just like roll around.
Lex Fridman (21:39.020)
I'll show you some moves.
Mark Zuckerberg (21:40.100)
Well, give me a year to train and then we can do it.
Lex Fridman (21:43.700)
You know, you've seen Rocky IV
Mark Zuckerberg (21:44.780)
where the Russian faces off the American.
Lex Fridman (21:46.340)
I'm the Russian in this picture.
Lex Fridman (21:47.980)
And then you're the Rocky, the underdog
Lex Fridman (21:49.980)
that gets to win in the end.
Mark Zuckerberg (21:51.340)
The idea of me as Rocky and like fighting is...
Lex Fridman (21:56.180)
If he dies, he dies.
Mark Zuckerberg (21:58.080)
Sorry, I just had to.
Lex Fridman (22:00.740)
I mean.
Mark Zuckerberg (22:01.580)
Anyway, yeah.
Lex Fridman (22:02.640)
But I mean, a lot of aspects of fitness.
Mark Zuckerberg (22:05.900)
You know, I don't know if you've tried supernatural
Lex Fridman (22:08.780)
on Quest or...
Lex Fridman (22:10.180)
So first of all, can I just comment on the fact
Lex Fridman (22:12.060)
every time I played around with Quest 2,
Mark Zuckerberg (22:15.020)
I just, I get giddy every time I step into virtual reality.
Lex Fridman (22:18.780)
So you mentioned productivity and all those kinds of things.
Mark Zuckerberg (22:20.860)
That's definitely something I'm excited about,
Lex Fridman (22:23.820)
but really I just love the possibilities
Mark Zuckerberg (22:26.780)
of stepping into that world.
Lex Fridman (22:28.820)
Maybe it's the introvert in me,
Lex Fridman (22:30.460)
but it just feels like the most convenient way
Lex Fridman (22:34.060)
to travel into worlds,
Mark Zuckerberg (22:37.500)
into worlds that are similar to the real world
Lex Fridman (22:40.340)
or totally different.
Mark Zuckerberg (22:41.660)
It's like Alice in Wonderland.
Lex Fridman (22:42.860)
Just try out crazy stuff.
Mark Zuckerberg (22:44.660)
The possibilities are endless.
Lex Fridman (22:45.780)
And I just, I personally am just love,
Mark Zuckerberg (22:50.900)
get excited for stepping in those virtual worlds.
Lex Fridman (22:53.980)
So I'm a huge fan.
Mark Zuckerberg (22:55.020)
In terms of the productivity as a programmer,
Lex Fridman (22:58.300)
I spend most of my day programming.
Mark Zuckerberg (23:00.060)
That's really interesting also,
Lex Fridman (23:01.980)
but then you have to develop the right IDEs.
Mark Zuckerberg (23:04.340)
You have to develop, like there has to be a threshold
Lex Fridman (23:07.380)
where a large amount of the programming community
Mark Zuckerberg (23:09.340)
moves there, but the collaborative aspects
Lex Fridman (23:11.860)
that are possible in terms of meetings,
Mark Zuckerberg (23:14.260)
in terms of when two coders are working together,
Lex Fridman (23:18.380)
I mean, the possibilities there are super, super exciting.
Mark Zuckerberg (23:21.740)
I think that in building this, we sort of need to balance.
Lex Fridman (23:27.100)
There are gonna be some new things
Mark Zuckerberg (23:28.160)
that you just couldn't do before.
Lex Fridman (23:29.700)
And those are gonna be the amazing experiences.
Lex Fridman (23:31.540)
So teleporting to any place, right?
Lex Fridman (23:33.420)
Whether it's a real place or something that people made.
Lex Fridman (23:38.620)
And I mean, some of the experiences
Lex Fridman (23:40.340)
around how we can build stuff in new ways,
Mark Zuckerberg (23:42.100)
where a lot of the stuff that,
Lex Fridman (23:44.740)
when I'm coding stuff, it's like, all right,
Mark Zuckerberg (23:46.060)
you code it and then you build it
Lex Fridman (23:47.180)
and then you see it afterwards.
Lex Fridman (23:48.260)
But increasingly it's gonna be possible to,
Lex Fridman (23:50.500)
you're in a world and you're building the world
Mark Zuckerberg (23:52.780)
as you are in it and kind of manipulating it.
Lex Fridman (23:55.780)
One of the things that we showed at our Inside the Lab
Mark Zuckerberg (23:59.900)
for recent artificial intelligence progress
Lex Fridman (24:02.500)
is this Builder Bot program,
Mark Zuckerberg (24:03.940)
where now you can just talk to it and say,
Lex Fridman (24:07.180)
hey, okay, I'm in this world,
Mark Zuckerberg (24:08.460)
like put some trees over there and it'll do that.
Lex Fridman (24:10.740)
And like, all right, put some bottles of water
Mark Zuckerberg (24:13.220)
on our picnic blanket and it'll do that
Lex Fridman (24:17.060)
and you're in the world.
Lex Fridman (24:17.900)
And I think there are gonna be new paradigms for coding.
Lex Fridman (24:19.940)
So yeah, there are gonna be some things
Mark Zuckerberg (24:22.100)
that I think are just pretty amazing,
Lex Fridman (24:24.620)
especially the first few times that you do them,
Lex Fridman (24:26.580)
but that you're like, whoa,
Lex Fridman (24:28.300)
like I've never had an experience like this.
Lex Fridman (24:30.620)
But most of your life, I would imagine,
Lex Fridman (24:34.260)
is not doing things that are amazing for the first time.
Mark Zuckerberg (24:38.180)
A lot of this in terms of,
Lex Fridman (24:39.620)
I mean, just answering your question from before around,
Lex Fridman (24:42.020)
what is it gonna take
Lex Fridman (24:42.860)
before you're spending most of your time in this?
Mark Zuckerberg (24:45.060)
Well, first of all, let me just say it as an aside,
Lex Fridman (24:48.180)
the goal isn't to have people spend a lot more time
Mark Zuckerberg (24:50.340)
in computing.
Lex Fridman (24:51.180)
It's to make it so that. I'm asking for myself.
Lex Fridman (24:52.340)
Yeah, it's to make it. When will I spend all my time in?
Lex Fridman (24:54.460)
Yeah, it's to make computing more natural.
Lex Fridman (24:57.060)
But I think you will spend most of your computing time
Lex Fridman (25:02.740)
in this when it does the things
Mark Zuckerberg (25:04.900)
that you use computing for somewhat better.
Lex Fridman (25:07.300)
So maybe having your perfect workstation
Mark Zuckerberg (25:10.540)
is a 5% improvement on your coding productivity.
Lex Fridman (25:15.140)
Maybe it's not like a completely new thing.
Lex Fridman (25:19.300)
But I mean, look, if I could increase the productivity
Lex Fridman (25:21.620)
of every engineer at Meta by 5%,
Mark Zuckerberg (25:25.500)
we'd buy those devices for everyone.
Lex Fridman (25:27.620)
And I imagine a lot of other companies would too.
Lex Fridman (25:30.340)
And that's how you start getting to the scale
Lex Fridman (25:31.860)
that I think makes this rival
Mark Zuckerberg (25:34.500)
some of the bigger computing platforms that exist today.
Lex Fridman (25:37.020)
Let me ask you about identity.
Mark Zuckerberg (25:38.300)
We talked about the avatar.
Lex Fridman (25:40.460)
How do you see identity in the Metaverse?
Mark Zuckerberg (25:42.740)
Should the avatar be tied to your identity
Lex Fridman (25:46.420)
or can I be anything in the Metaverse?
Lex Fridman (25:49.300)
Like, can I be whatever the heck I want?
Lex Fridman (25:52.180)
Can I even be a troll?
Lex Fridman (25:53.660)
So there's exciting freeing possibilities
Lex Fridman (25:57.420)
and there's the darker possibilities too.
Lex Fridman (26:00.740)
Yeah, I mean, I think that there's gonna be a range, right?
Lex Fridman (26:03.180)
So we're working on, for expression and avatars,
Mark Zuckerberg (26:10.180)
on one end of the spectrum are kind of expressive
Lex Fridman (26:13.100)
and cartoonish avatars.
Lex Fridman (26:14.940)
And then on the other end of the spectrum
Lex Fridman (26:16.460)
are photorealistic avatars.
Lex Fridman (26:18.500)
And I just think the reality is
Lex Fridman (26:20.660)
that there are gonna be different use cases
Mark Zuckerberg (26:22.100)
for different things.
Lex Fridman (26:23.260)
And I guess there's another axis.
Lex Fridman (26:25.140)
So if you're going from photorealistic to expressive,
Lex Fridman (26:28.700)
there's also like representing you directly
Mark Zuckerberg (26:31.100)
versus like some fantasy identity.
Lex Fridman (26:33.660)
And I think that there are gonna be things
Lex Fridman (26:35.340)
on all ends of that spectrum too, right?
Lex Fridman (26:37.860)
So you'll want photo, like in some experience,
Lex Fridman (26:41.020)
you might wanna be like a photorealistic dragon, right?
Lex Fridman (26:44.300)
Or if I'm playing Onward,
Mark Zuckerberg (26:46.980)
or just this military simulator game,
Lex Fridman (26:50.940)
I think getting to be more photorealistic as a soldier
Mark Zuckerberg (26:53.620)
in that could enhance the experience.
Lex Fridman (26:57.780)
There are times when I'm hanging out with friends
Mark Zuckerberg (26:59.540)
where I want them to know it's me.
Lex Fridman (27:02.060)
So a kind of cartoonish or expressive version of me is good.
Lex Fridman (27:06.180)
But there are also experiences like,
Lex Fridman (27:09.580)
VRChat does this well today,
Mark Zuckerberg (27:11.580)
where a lot of the experience is kind of dressing up
Lex Fridman (27:14.900)
and wearing a fantastical avatar
Mark Zuckerberg (27:17.780)
that's almost like a meme or is humorous.
Lex Fridman (27:19.580)
So you come into an experience
Lex Fridman (27:21.300)
and it's almost like you have like a built in icebreaker
Lex Fridman (27:24.540)
because like you see people and you're just like,
Mark Zuckerberg (27:27.300)
all right, I'm cracking up at what you're wearing
Lex Fridman (27:29.940)
because that's funny.
Lex Fridman (27:30.780)
And it's just like, where'd you get that?
Lex Fridman (27:31.900)
Or, oh, you made that?
Mark Zuckerberg (27:32.740)
That's, it's awesome.
Lex Fridman (27:35.500)
Whereas, okay, if you're going into a work meeting,
Mark Zuckerberg (27:38.900)
maybe a photorealistic version of your real self
Lex Fridman (27:41.740)
is gonna be the most appropriate thing for that.
Lex Fridman (27:43.540)
So I think the reality is there aren't going to be,
Lex Fridman (27:47.340)
it's not just gonna be one thing.
Mark Zuckerberg (27:50.500)
You know, my own sense of kind of how you wanna
Lex Fridman (27:54.380)
express identity online has sort of evolved over time.
Lex Fridman (27:56.860)
And that, you know, early days in Facebook,
Lex Fridman (27:58.620)
I thought, okay, people are gonna have one identity.
Lex Fridman (28:00.260)
And now I think that's clearly not gonna be the case.
Lex Fridman (28:02.100)
I think you're gonna have all these different things
Lex Fridman (28:04.420)
and there's utility in being able to do different things.
Lex Fridman (28:07.300)
So some of the technical challenges
Mark Zuckerberg (28:10.100)
that I'm really interested in around it
Lex Fridman (28:12.140)
are how do you build the software
Lex Fridman (28:14.180)
to allow people to seamlessly go between them?
Lex Fridman (28:17.100)
So say, so you could view them
Mark Zuckerberg (28:19.300)
as just completely discrete points on a spectrum,
Lex Fridman (28:25.140)
but let's talk about the metaverse economy for a second.
Mark Zuckerberg (28:28.500)
Let's say I buy a digital shirt
Lex Fridman (28:31.220)
for my photorealistic avatar, which by the way,
Mark Zuckerberg (28:34.420)
I think at the time where we're spending a lot of time
Lex Fridman (28:36.460)
in the metaverse doing a lot of our work meetings
Mark Zuckerberg (28:38.740)
in the metaverse and et cetera,
Lex Fridman (28:40.260)
I would imagine that the economy around virtual clothing
Mark Zuckerberg (28:42.460)
as an example is going to be quite as big.
Lex Fridman (28:44.660)
Why wouldn't I spend almost as much money
Mark Zuckerberg (28:47.100)
in investing in my appearance or expression
Lex Fridman (28:49.780)
for my photorealistic avatar for meetings
Mark Zuckerberg (28:52.420)
as I would for whatever I'm gonna wear in my video chat.
Lex Fridman (28:55.540)
But the question is, okay, so you,
Mark Zuckerberg (28:56.700)
let's say you buy some shirt
Lex Fridman (28:57.620)
for your photorealistic avatar.
Mark Zuckerberg (28:59.780)
Wouldn't it be cool if there was a way
Lex Fridman (29:02.620)
to basically translate that into a more expressive thing
Lex Fridman (29:07.620)
for your kind of cartoonish or expressive avatar?
Lex Fridman (29:11.220)
And there are multiple ways to do that.
Mark Zuckerberg (29:12.580)
You can view them as two discrete points and okay,
Lex Fridman (29:14.940)
maybe if a designer sells one thing,
Mark Zuckerberg (29:18.220)
then it actually comes in a pack and there's two
Lex Fridman (29:19.940)
and you can use either one on that,
Lex Fridman (29:22.340)
but I actually think this stuff might exist more
Lex Fridman (29:24.420)
as a spectrum in the future.
Lex Fridman (29:26.100)
And that's what I do think the direction
Lex Fridman (29:29.460)
on some of the AI advances that is happening
Mark Zuckerberg (29:33.380)
to be able to, especially stuff around like style transfer,
Lex Fridman (29:35.980)
being able to take a piece of art or express something
Lex Fridman (29:39.860)
and say, okay, paint me this photo in the style of Gauguin
Lex Fridman (29:44.900)
or whoever it is that you're interested in.
Mark Zuckerberg (29:49.060)
Take this shirt and put it in the style
Lex Fridman (29:51.300)
of what I've designed for my expressive avatar.
Mark Zuckerberg (29:55.220)
I think that's gonna be pretty compelling.
Lex Fridman (29:56.940)
And so the fashion, you might be buying like a generator,
Mark Zuckerberg (30:00.060)
like a closet that generates a style.
Lex Fridman (30:03.260)
And then like with the GANs,
Mark Zuckerberg (30:05.540)
you'll be able to infinitely generate outfits
Lex Fridman (30:08.180)
thereby making it, so the reason I wear the same thing
Mark Zuckerberg (30:10.780)
all the time is I don't like choice.
Lex Fridman (30:12.380)
You've talked about the same thing,
Lex Fridman (30:15.140)
but now you don't even have to choose.
Lex Fridman (30:16.700)
Your closet generates your outfit for you every time.
Lex Fridman (30:19.580)
So you have to live with the outfit it generates.
Lex Fridman (30:23.460)
I mean, you could do that, although,
Mark Zuckerberg (30:25.500)
no, I think that that's, I think some people will,
Lex Fridman (30:27.500)
but I think like, I think there's going to be a huge aspect
Mark Zuckerberg (30:31.300)
of just people doing creative commerce here.
Lex Fridman (30:35.900)
So I think that there is going to be a big market
Mark Zuckerberg (30:37.860)
around people designing digital clothing.
Lex Fridman (30:41.060)
But the question is, if you're designing digital clothing,
Lex Fridman (30:43.020)
do you need to design, if you're the designer,
Lex Fridman (30:44.860)
do you need to make it for each kind of specific discrete
Mark Zuckerberg (30:48.140)
point along a spectrum, or are you just designing it
Lex Fridman (30:51.500)
for kind of a photo realistic case or an expressive case,
Mark Zuckerberg (30:54.140)
or can you design one
Lex Fridman (30:55.140)
and have it translate across these things?
Mark Zuckerberg (30:57.460)
If I buy a style from a designer who I care about,
Lex Fridman (31:01.780)
and now I'm a dragon, is there a way to morph that
Lex Fridman (31:04.220)
so it goes on the dragon in a way that makes sense?
Lex Fridman (31:07.660)
And that I think is an interesting AI problem
Mark Zuckerberg (31:09.460)
because you're probably not going to make it
Lex Fridman (31:10.820)
so that designers have to go design for all those things.
Lex Fridman (31:14.700)
But the more useful the digital content is that you buy
Lex Fridman (31:17.900)
in a lot of uses, in a lot of use cases,
Mark Zuckerberg (31:21.220)
the more that economy will just explode.
Lex Fridman (31:23.420)
And that's a lot of what all of the,
Mark Zuckerberg (31:28.100)
we were joking about NFTs before,
Lex Fridman (31:29.700)
but I think a lot of the promise here is that
Mark Zuckerberg (31:32.580)
if the digital goods that you buy are not just tied
Lex Fridman (31:35.020)
to one platform or one use case,
Mark Zuckerberg (31:37.060)
they end up being more valuable,
Lex Fridman (31:38.260)
which means that people are more willing
Lex Fridman (31:39.820)
and more likely to invest in them,
Lex Fridman (31:41.300)
and that just spurs the whole economy.
Lex Fridman (31:44.220)
But the question is, that's a fascinating positive aspect,
Lex Fridman (31:47.260)
but the potential negative aspect is that
Mark Zuckerberg (31:50.780)
you can have people concealing their identity
Lex Fridman (31:52.660)
in order to troll or even not people, bots.
Lex Fridman (31:57.060)
So how do you know in the metaverse
Lex Fridman (31:58.780)
that you're talking to a real human or an AI
Lex Fridman (32:02.060)
or a well intentioned human?
Lex Fridman (32:03.940)
Is that something you think about,
Lex Fridman (32:04.980)
something you're concerned about?
Lex Fridman (32:06.940)
Well, let's break that down into a few different cases.
Mark Zuckerberg (32:10.260)
I mean, because knowing that you're talking to someone
Lex Fridman (32:11.980)
who has good intentions is something that I think
Mark Zuckerberg (32:13.880)
is not even solved in pretty much anywhere.
Lex Fridman (32:17.860)
But I mean, if you're talking to someone who's a dragon,
Mark Zuckerberg (32:20.380)
I think it's pretty clear that they're not representing
Lex Fridman (32:22.000)
themselves as a person.
Mark Zuckerberg (32:23.300)
I think probably the most pernicious thing
Lex Fridman (32:25.300)
that you want to solve for is,
Mark Zuckerberg (32:30.140)
I think probably one of the scariest ones is
Lex Fridman (32:32.300)
how do you make sure that someone isn't impersonating you?
Mark Zuckerberg (32:35.020)
So, okay, you're in a future version of this conversation,
Lex Fridman (32:39.340)
and we have photorealistic avatars,
Lex Fridman (32:41.700)
and we're doing this in work rooms
Lex Fridman (32:43.320)
or whatever the future version of that is,
Lex Fridman (32:44.980)
and someone walks in who looks like me.
Lex Fridman (32:48.860)
How do you know that that's me?
Lex Fridman (32:50.300)
And one of the things that we're thinking about
Lex Fridman (32:54.200)
is it's still a pretty big AI project
Mark Zuckerberg (32:57.500)
to be able to generate photorealistic avatars
Lex Fridman (32:59.500)
that basically can like,
Lex Fridman (33:00.880)
they work like these codecs of you, right?
Lex Fridman (33:03.380)
So you kind of have a map from your headset
Lex Fridman (33:06.220)
and whatever sensors of what your body's actually doing,
Lex Fridman (33:08.020)
and it takes the model and it kind of displays it in VR.
Lex Fridman (33:11.180)
But there's a question, which is,
Lex Fridman (33:12.660)
should there be some sort of biometric security
Lex Fridman (33:15.420)
so that when I put on my VR headset
Lex Fridman (33:18.220)
or I'm going to go use that avatar,
Lex Fridman (33:20.940)
I need to first prove that I am that?
Lex Fridman (33:24.340)
And I think you probably are gonna want something like that.
Lex Fridman (33:26.780)
So as we're developing these technologies,
Lex Fridman (33:31.120)
we're also thinking about the security for things like that
Mark Zuckerberg (33:34.500)
because people aren't gonna wanna be impersonated.
Lex Fridman (33:37.060)
That's a huge security issue.
Mark Zuckerberg (33:41.700)
Then you just get the question
Lex Fridman (33:42.900)
of people hiding behind fake accounts
Mark Zuckerberg (33:46.580)
to do malicious things,
Lex Fridman (33:48.300)
which is not gonna be unique to the metaverse,
Mark Zuckerberg (33:51.020)
although certainly in a environment
Lex Fridman (33:56.140)
where it's more immersive
Lex Fridman (33:57.300)
and you have more of a sense of presence,
Lex Fridman (33:58.640)
it could be more painful.
Lex Fridman (34:01.740)
But this is obviously something
Lex Fridman (34:03.140)
that we've just dealt with for years
Mark Zuckerberg (34:06.480)
in social media and the internet more broadly.
Lex Fridman (34:08.740)
And there, I think there have been a bunch of tactics
Mark Zuckerberg (34:13.140)
that I think we've just evolved to,
Lex Fridman (34:17.900)
we've built up these different AI systems
Mark Zuckerberg (34:20.480)
to basically get a sense of,
Lex Fridman (34:21.880)
is this account behaving in the way that a person would?
Lex Fridman (34:26.340)
And it turns out,
Lex Fridman (34:28.340)
so in all of the work that we've done around,
Mark Zuckerberg (34:31.940)
we call it community integrity
Lex Fridman (34:33.340)
and it's basically like policing harmful content
Lex Fridman (34:36.920)
and trying to figure out where to draw the line.
Lex Fridman (34:38.340)
And there are all these like really hard
Lex Fridman (34:39.800)
and philosophical questions around like,
Lex Fridman (34:41.300)
where do you draw the line on some of this stuff?
Lex Fridman (34:42.880)
And the thing that I've kind of found the most effective
Lex Fridman (34:47.820)
is as much as possible trying to figure out
Mark Zuckerberg (34:51.260)
who are the inauthentic accounts
Lex Fridman (34:53.380)
or where are the accounts that are behaving
Mark Zuckerberg (34:55.540)
in an overall harmful way at the account level,
Lex Fridman (34:58.420)
rather than trying to get into like policing
Lex Fridman (35:00.460)
what they're saying, right?
Lex Fridman (35:01.420)
Which I think the metaverse is gonna be even harder
Mark Zuckerberg (35:03.700)
because the metaverse I think will have more properties of,
Lex Fridman (35:07.320)
it's almost more like a phone call, right?
Mark Zuckerberg (35:09.220)
Or it's not like I post a piece of content
Lex Fridman (35:12.380)
and is that piece of content good or bad?
Lex Fridman (35:14.700)
So I think more of this stuff will have to be done
Lex Fridman (35:16.260)
at the level of the account.
Lex Fridman (35:19.420)
But this is the area where,
Lex Fridman (35:21.740)
between the kind of counter intelligence teams
Mark Zuckerberg (35:27.140)
that we built up inside the company
Lex Fridman (35:28.480)
and like years of building just different AI systems
Mark Zuckerberg (35:33.220)
to basically detect what is a real account and what isn't.
Lex Fridman (35:36.900)
I'm not saying we're perfect,
Lex Fridman (35:37.940)
but like this is an area where I just think
Lex Fridman (35:39.920)
we are like years ahead of basically anyone else
Mark Zuckerberg (35:43.580)
in the industry in terms of having built those capabilities.
Lex Fridman (35:48.100)
And I think that that just is gonna be incredibly important
Mark Zuckerberg (35:50.180)
for this next wave of things.
Lex Fridman (35:51.540)
And like you said, on a technical level,
Mark Zuckerberg (35:53.460)
on a philosophical level,
Lex Fridman (35:54.980)
it's an incredibly difficult problem to solve.
Mark Zuckerberg (35:59.260)
By the way, I would probably like to open source my avatar
Lex Fridman (36:03.220)
so there could be like millions of Lexis walking around
Mark Zuckerberg (36:05.940)
just like an army.
Lex Fridman (36:07.040)
Like Agent Smith?
Mark Zuckerberg (36:08.500)
Agent Smith, yeah, exactly.
Lex Fridman (36:10.700)
So the Unity ML folks built a copy of me
Lex Fridman (36:16.420)
and they sent it to me.
Lex Fridman (36:18.420)
So there's a person running around
Lex Fridman (36:20.220)
and I've just been doing reinforcement learning on it.
Lex Fridman (36:22.500)
I was gonna release it
Mark Zuckerberg (36:25.420)
because just to have sort of like thousands of Lexis
Lex Fridman (36:29.860)
doing reinforcement.
Lex Fridman (36:31.180)
So they fall over naturally,
Lex Fridman (36:32.460)
they have to learn how to like walk around and stuff.
Lex Fridman (36:34.900)
So I love that idea,
Lex Fridman (36:36.660)
this tension between biometric security,
Mark Zuckerberg (36:39.140)
you want to have one identity,
Lex Fridman (36:40.340)
but then certain avatars, you might have to have many.
Mark Zuckerberg (36:43.620)
I don't know which is better security,
Lex Fridman (36:45.400)
sort of flooding the world with Lexis
Lex Fridman (36:48.140)
and thereby achieving security
Lex Fridman (36:49.440)
or really being protective of your identity.
Mark Zuckerberg (36:51.780)
I have to ask you a security question actually.
Lex Fridman (36:53.860)
Well, how does flooding the world with Lexis help me know
Lex Fridman (36:56.860)
in our conversation that I'm talking to the real Lex?
Lex Fridman (36:59.640)
I completely destroy the trust
Lex Fridman (37:01.580)
in all my relationships then, right?
Lex Fridman (37:03.060)
If I flood,
Mark Zuckerberg (37:04.180)
cause then it's, yeah, that.
Lex Fridman (37:07.820)
I think that one's not gonna work that well for you.
Mark Zuckerberg (37:09.500)
It's not gonna work that well for the original copy.
Lex Fridman (37:11.860)
It probably fits some things.
Mark Zuckerberg (37:13.380)
Like if you're a public figure
Lex Fridman (37:14.820)
and you're trying to have a bunch of,
Mark Zuckerberg (37:18.500)
if you're trying to show up
Lex Fridman (37:19.480)
in a bunch of different places in the future,
Mark Zuckerberg (37:21.060)
you'll be able to do that in the metaverse.
Lex Fridman (37:23.500)
So that kind of replication I think will be useful.
Lex Fridman (37:26.260)
But I do think that you're gonna want a notion of like,
Lex Fridman (37:29.260)
I am talking to the real one.
Mark Zuckerberg (37:31.500)
Yeah.
Lex Fridman (37:32.700)
Yeah, especially if the fake ones start outperforming you
Lex Fridman (37:35.660)
and all your private relationships
Lex Fridman (37:37.460)
and then you're left behind.
Mark Zuckerberg (37:38.740)
I mean, that's a serious concern I have with clones.
Lex Fridman (37:41.060)
Again, the things I think about.
Mark Zuckerberg (37:43.340)
Okay, so I recently got, I use QNAP NAS storage.
Lex Fridman (37:48.380)
So just storage for video and stuff.
Lex Fridman (37:50.220)
And I recently got hacked.
Lex Fridman (37:51.460)
This is the first time for me with ransomware.
Mark Zuckerberg (37:53.540)
It's not me personally, it's all QNAP devices.
Lex Fridman (37:58.700)
So the question that people have
Mark Zuckerberg (38:00.780)
is about security in general.
Lex Fridman (38:03.300)
Because I was doing a lot of the right things
Mark Zuckerberg (38:05.060)
in terms of security and nevertheless,
Lex Fridman (38:06.820)
ransomware basically disabled my device.
Lex Fridman (38:10.940)
Is that something you think about?
Lex Fridman (38:12.060)
What are the different steps you could take
Lex Fridman (38:13.780)
to protect people's data on the security front?
Lex Fridman (38:16.940)
I think that there's different solutions for,
Lex Fridman (38:21.380)
and strategies where it makes sense to have stuff
Lex Fridman (38:23.660)
kind of put behind a fortress, right?
Lex Fridman (38:25.460)
So the centralized model versus the decentralizing.
Lex Fridman (38:30.220)
Then I think both have strengths and weaknesses.
Lex Fridman (38:32.140)
So I think anyone who says, okay,
Lex Fridman (38:33.260)
just decentralize everything, that'll make it more secure.
Mark Zuckerberg (38:36.660)
I think that that's tough because,
Lex Fridman (38:38.980)
I mean, the advantage of something like encryption
Mark Zuckerberg (38:42.740)
is that we run the largest encrypted service
Lex Fridman (38:46.420)
in the world with WhatsApp.
Lex Fridman (38:47.700)
And we're one of the first to roll out
Lex Fridman (38:49.580)
a multi platform encryption service.
Lex Fridman (38:52.660)
And that's something that I think was a big advance
Lex Fridman (38:55.980)
for the industry.
Lex Fridman (38:57.180)
And one of the promises that we can basically make
Lex Fridman (38:59.300)
because of that, our company doesn't see
Mark Zuckerberg (39:02.260)
when you're sending an encrypted message
Lex Fridman (39:04.500)
and to an encrypted message,
Lex Fridman (39:05.860)
what the content is of what you're sharing.
Lex Fridman (39:07.860)
So that way, if someone hacks Meta servers,
Mark Zuckerberg (39:11.540)
they're not gonna be able to access the WhatsApp message
Lex Fridman (39:14.700)
that you're sending to your friend.
Lex Fridman (39:16.940)
And that I think matters a lot to people
Lex Fridman (39:19.100)
because obviously if someone is able to compromise
Mark Zuckerberg (39:21.900)
a company's servers and that company has hundreds
Lex Fridman (39:23.900)
of millions or billions of people,
Mark Zuckerberg (39:25.220)
then that ends up being a very big deal.
Lex Fridman (39:27.900)
The flip side of that is, okay,
Mark Zuckerberg (39:29.380)
all the content is on your phone.
Lex Fridman (39:32.900)
Are you following security best practices on your phone?
Mark Zuckerberg (39:35.860)
If you lose your phone, all your content is gone.
Lex Fridman (39:38.060)
So that's an issue.
Mark Zuckerberg (39:39.620)
Maybe you go back up your content from WhatsApp
Lex Fridman (39:42.300)
or some other service in an iCloud or something,
Lex Fridman (39:45.740)
but then you're just at Apple's whims about,
Lex Fridman (39:47.940)
are they gonna go turn over the data to some government
Lex Fridman (39:51.920)
or are they gonna get hacked?
Lex Fridman (39:53.380)
So a lot of the time it is useful to have data
Mark Zuckerberg (39:57.340)
in a centralized place too because then you can train
Lex Fridman (40:00.580)
systems that can just do much better personalization.
Mark Zuckerberg (40:04.740)
I think that in a lot of cases, centralized systems
Lex Fridman (40:08.580)
can offer, especially if you're a serious company,
Mark Zuckerberg (40:13.460)
you're running the state of the art stuff
Lex Fridman (40:16.020)
and you have red teams attacking your own stuff
Lex Fridman (40:19.540)
and you're putting out bounty programs
Lex Fridman (40:24.260)
and trying to attract some of the best hackers in the world
Mark Zuckerberg (40:26.260)
to go break into your stuff all the time.
Lex Fridman (40:27.820)
So any system is gonna have security issues,
Lex Fridman (40:30.500)
but I think the best way forward is to basically try
Lex Fridman (40:34.340)
to be as aggressive and open about hardening
Mark Zuckerberg (40:36.460)
the systems as possible, not trying to kind of hide
Lex Fridman (40:39.140)
and pretend that there aren't gonna be issues,
Mark Zuckerberg (40:40.740)
which I think is over time why a lot of open source systems
Lex Fridman (40:43.820)
have gotten relatively more secure is because they're open
Lex Fridman (40:46.540)
and it's not, rather than pretending that there aren't
Lex Fridman (40:48.740)
gonna be issues, just people surface them quicker.
Lex Fridman (40:50.900)
So I think you want to adopt that approach as a company
Lex Fridman (40:53.760)
and just constantly be hardening yourself.
Mark Zuckerberg (40:56.640)
Trying to stay one step ahead of the attackers.
Lex Fridman (41:01.060)
It's an inherently adversarial space.
Mark Zuckerberg (41:03.940)
I think it's an interesting security is interesting
Lex Fridman (41:07.260)
because of the different kind of threats
Mark Zuckerberg (41:09.140)
that we've managed over the last five years,
Lex Fridman (41:11.820)
there are ones where basically the adversaries
Mark Zuckerberg (41:15.540)
keep on getting better and better.
Lex Fridman (41:16.820)
So trying to kind of interfere with security
Mark Zuckerberg (41:21.580)
is certainly one area of this.
Lex Fridman (41:23.100)
If you have nation states that are trying
Mark Zuckerberg (41:24.860)
to interfere in elections or something,
Lex Fridman (41:27.300)
they're kind of evolving their tactics.
Mark Zuckerberg (41:29.460)
Whereas on the other hand, I don't want to be too simplistic
Lex Fridman (41:32.780)
about it, but if someone is saying something hateful,
Mark Zuckerberg (41:36.660)
people usually aren't getting smarter and smarter
Lex Fridman (41:38.700)
about how they say hateful things.
Lex Fridman (41:40.460)
So maybe there's some element of that,
Lex Fridman (41:42.600)
but it's a very small dynamic compared
Mark Zuckerberg (41:44.820)
to how advanced attackers and some of these other places
Lex Fridman (41:48.520)
get over time.
Mark Zuckerberg (41:49.980)
I believe most people are good,
Lex Fridman (41:51.360)
so they actually get better over time
Lex Fridman (41:53.660)
and not being less hateful
Lex Fridman (41:55.420)
because they realize it's not fun being hateful.
Mark Zuckerberg (42:00.060)
That's at least the belief I have.
Lex Fridman (42:01.980)
But first, bathroom break.
Mark Zuckerberg (42:04.940)
Sure, okay.
Lex Fridman (42:06.840)
So we'll come back to AI,
Lex Fridman (42:08.180)
but let me ask some difficult questions now.
Lex Fridman (42:11.020)
Social Dilemma is a popular documentary
Mark Zuckerberg (42:13.820)
that raised concerns about the effects
Lex Fridman (42:15.500)
of social media on society.
Mark Zuckerberg (42:17.540)
You responded with a point by point rebuttal titled,
Lex Fridman (42:20.820)
What the Social Dilemma Gets Wrong.
Mark Zuckerberg (42:23.120)
People should read that.
Lex Fridman (42:24.980)
I would say the key point they make
Mark Zuckerberg (42:26.700)
is because social media is funded by ads,
Lex Fridman (42:29.580)
algorithms want to maximize attention and engagement
Lex Fridman (42:33.280)
and an effective way to do so is to get people angry
Lex Fridman (42:38.020)
at each other, increase division and so on.
Lex Fridman (42:40.940)
Can you steel man their criticisms and arguments
Lex Fridman (42:44.260)
that they make in the documentary
Mark Zuckerberg (42:46.260)
as a way to understand the concern
Lex Fridman (42:48.580)
and as a way to respond to it?
Mark Zuckerberg (42:53.060)
Well, yeah, I think that's a good conversation to have.
Lex Fridman (42:56.860)
I don't happen to agree with the conclusions
Lex Fridman (43:00.420)
and I think that they make a few assumptions
Lex Fridman (43:02.100)
that are just very big jumps
Mark Zuckerberg (43:06.300)
that I don't think are reasonable to make.
Lex Fridman (43:08.880)
But I understand overall why people would be concerned
Mark Zuckerberg (43:13.880)
that our business model and ads in general,
Lex Fridman (43:19.380)
we do make more money
Lex Fridman (43:20.640)
as people use the service more in general, right?
Lex Fridman (43:23.280)
So as a kind of basic assumption, okay,
Mark Zuckerberg (43:26.680)
do we have an incentive for people to build a service
Lex Fridman (43:29.400)
that people use more?
Mark Zuckerberg (43:31.200)
Yes, on a lot of levels.
Lex Fridman (43:32.840)
I mean, we think what we're doing is good.
Lex Fridman (43:34.480)
So we think that if people are finding it useful,
Lex Fridman (43:37.140)
they'll use it more.
Mark Zuckerberg (43:38.560)
Or if you just look at it as this sort of,
Lex Fridman (43:41.320)
if the only thing we cared about is money,
Mark Zuckerberg (43:43.280)
which is not for anyone who knows me,
Lex Fridman (43:46.200)
but okay, we're a company.
Lex Fridman (43:47.820)
So let's say you just kind of simplified it down to that,
Lex Fridman (43:51.360)
then would we want people to use the services more?
Mark Zuckerberg (43:53.840)
Yes, and then you get to the second question,
Lex Fridman (43:57.240)
which is does kind of getting people agitated
Lex Fridman (44:02.920)
make them more likely to use the services more?
Lex Fridman (44:07.480)
And I think from looking at other media in the world,
Mark Zuckerberg (44:12.480)
especially TV, and there's the old news adage,
Lex Fridman (44:17.280)
if it bleeds, it leads.
Mark Zuckerberg (44:18.680)
Like I think that this is,
Lex Fridman (44:20.020)
there are a bunch of reasons why someone might think
Mark Zuckerberg (44:25.520)
that that kind of provocative content
Lex Fridman (44:30.520)
would be the most engaging.
Mark Zuckerberg (44:32.640)
Now, what I've always found is two things.
Lex Fridman (44:35.640)
One is that what grabs someone's attention in the near term
Mark Zuckerberg (44:39.180)
is not necessarily something
Lex Fridman (44:40.840)
that they're going to appreciate having seen
Mark Zuckerberg (44:43.640)
or going to be the best over the long term.
Lex Fridman (44:45.320)
So I think what a lot of people get wrong
Mark Zuckerberg (44:47.460)
is that I'm not building this company
Lex Fridman (44:50.400)
to make the most money or get people to spend the most time
Mark Zuckerberg (44:53.300)
on this in the next quarter or the next year.
Lex Fridman (44:55.760)
I've been doing this for 17 years at this point,
Lex Fridman (44:58.980)
and I'm still relatively young,
Lex Fridman (45:00.380)
and I have a lot more that I wanna do
Mark Zuckerberg (45:02.040)
over the coming decades.
Lex Fridman (45:03.380)
So I think that it's too simplistic to say,
Mark Zuckerberg (45:08.380)
hey, this might increase time in the near term,
Lex Fridman (45:11.820)
therefore, it's what you're gonna do.
Mark Zuckerberg (45:13.440)
Because I actually think a deeper look
Lex Fridman (45:15.340)
at kind of what my incentives are,
Mark Zuckerberg (45:17.240)
the incentives of a company
Lex Fridman (45:18.300)
that are focused on the long term,
Mark Zuckerberg (45:20.540)
is to basically do what people
Lex Fridman (45:22.720)
are gonna find valuable over time,
Mark Zuckerberg (45:24.160)
not what is gonna draw people's attention today.
Lex Fridman (45:26.780)
The other thing that I'd say is that,
Mark Zuckerberg (45:29.980)
I think a lot of times people look at this
Lex Fridman (45:31.500)
from the perspective of media
Mark Zuckerberg (45:34.580)
or kind of information or civic discourse,
Lex Fridman (45:37.780)
but one other way of looking at this is just that,
Lex Fridman (45:40.860)
okay, I'm a product designer, right?
Lex Fridman (45:42.540)
Our company, we build products,
Lex Fridman (45:45.180)
and a big part of building a product
Lex Fridman (45:47.340)
is not just the function and utility
Mark Zuckerberg (45:49.020)
of what you're delivering,
Lex Fridman (45:50.180)
but the feeling of how it feels, right?
Lex Fridman (45:52.020)
And we spend a lot of time talking about virtual reality
Lex Fridman (45:55.660)
and how the kind of key aspect of that experience
Mark Zuckerberg (45:58.820)
is the feeling of presence, which it's a visceral thing.
Lex Fridman (46:01.980)
It's not just about the utility that you're delivering,
Mark Zuckerberg (46:03.940)
it's about like the sensation.
Lex Fridman (46:05.940)
And similarly, I care a lot about how people feel
Mark Zuckerberg (46:10.420)
when they use our products,
Lex Fridman (46:11.420)
and I don't want to build products that make people angry.
Mark Zuckerberg (46:15.300)
I mean, that's like not, I think,
Lex Fridman (46:17.020)
what we're here on this earth to do,
Mark Zuckerberg (46:18.460)
is to build something that people spend a bunch of time doing
Lex Fridman (46:22.180)
and it just kind of makes them angrier at other people.
Mark Zuckerberg (46:24.020)
I mean, I think that that's not good.
Lex Fridman (46:26.300)
That's not what I think would be
Mark Zuckerberg (46:30.120)
sort of a good use of our time
Lex Fridman (46:32.040)
or a good contribution to the world.
Mark Zuckerberg (46:33.660)
So, okay, it's like people, they tell us
Lex Fridman (46:36.500)
on a per content basis, does this thing,
Lex Fridman (46:39.220)
do I like it?
Lex Fridman (46:40.060)
Do I love it?
Lex Fridman (46:40.880)
Does it make me angry?
Lex Fridman (46:41.720)
Does it make me sad?
Lex Fridman (46:42.980)
And based on that, we choose to basically show content
Lex Fridman (46:47.220)
that makes people angry less,
Mark Zuckerberg (46:49.180)
because of course, if you're designing a product
Lex Fridman (46:52.720)
and you want people to be able to connect
Lex Fridman (46:56.320)
and feel good over a long period of time,
Lex Fridman (46:59.180)
then that's naturally what you're gonna do.
Mark Zuckerberg (47:02.080)
So, I don't know, I think overall,
Lex Fridman (47:06.580)
I understand at a high level,
Mark Zuckerberg (47:10.540)
if you're not thinking too deeply about it,
Lex Fridman (47:13.680)
why that argument might be appealing.
Lex Fridman (47:16.100)
But I just think if you actually look
Lex Fridman (47:19.220)
at what our real incentives are,
Mark Zuckerberg (47:20.940)
not just like if we were trying to optimize
Lex Fridman (47:25.100)
for the next week,
Lex Fridman (47:26.820)
but like as people working on this,
Lex Fridman (47:28.980)
like why are we here?
Lex Fridman (47:30.480)
And I think it's pretty clear
Lex Fridman (47:32.900)
that that's not actually how you would wanna
Mark Zuckerberg (47:34.300)
design the system.
Lex Fridman (47:35.740)
I guess one other thing that I'd say is that,
Mark Zuckerberg (47:37.780)
while we're focused on the ads business model,
Lex Fridman (47:40.820)
I do think it's important to note that a lot
Mark Zuckerberg (47:43.400)
of these issues are not unique to ads.
Lex Fridman (47:45.380)
I mean, so take like a subscription news business model,
Mark Zuckerberg (47:47.900)
for example, I think that has just as many
Lex Fridman (47:50.920)
potential pitfalls.
Mark Zuckerberg (47:53.180)
Maybe if someone's paying for a subscription,
Lex Fridman (47:55.240)
you don't get paid per piece of content that they look at,
Lex Fridman (47:57.940)
but say for example, I think like a bunch
Lex Fridman (48:02.640)
of the partisanship that we see could potentially
Mark Zuckerberg (48:06.180)
be made worse by you have these kind of partisan
Lex Fridman (48:12.620)
news organizations that basically sell subscriptions
Lex Fridman (48:15.740)
and they're only gonna get people on one side
Lex Fridman (48:17.580)
to basically subscribe to them.
Lex Fridman (48:19.900)
So their incentive is not to print content
Lex Fridman (48:22.780)
or produce content that's kind of centrist
Mark Zuckerberg (48:26.100)
or down the line either.
Lex Fridman (48:27.820)
I bet that what a lot of them find is that
Mark Zuckerberg (48:30.060)
if they produce stuff that's kind of more polarizing
Lex Fridman (48:32.480)
or more partisan, then that is what gets
Mark Zuckerberg (48:35.460)
the more subscribers.
Lex Fridman (48:36.860)
So I think that this stuff is all,
Mark Zuckerberg (48:40.260)
there's no perfect business model.
Lex Fridman (48:41.940)
Everything has pitfalls.
Mark Zuckerberg (48:44.340)
The thing that I think is great about advertising
Lex Fridman (48:46.500)
is it makes it so the consumer service is free,
Mark Zuckerberg (48:48.740)
which if you believe that everyone should have a voice
Lex Fridman (48:50.860)
and everyone should be able to connect,
Mark Zuckerberg (48:52.020)
then that's a great thing, as opposed to building
Lex Fridman (48:55.000)
a luxury service that not everyone can afford.
Lex Fridman (48:57.240)
But look, every business model, you have to be careful
Lex Fridman (49:00.020)
about how you're implementing what you're doing.
Mark Zuckerberg (49:02.500)
You responded to a few things there.
Lex Fridman (49:04.620)
You spoke to the fact that there is a narrative
Mark Zuckerberg (49:07.260)
of malevolence, like you're leaning into them,
Lex Fridman (49:12.420)
making people angry just because it makes more money
Mark Zuckerberg (49:15.020)
in the short term, that kind of thing.
Lex Fridman (49:16.300)
So you responded to that.
Lex Fridman (49:17.880)
But there's also kind of reality of human nature.
Lex Fridman (49:22.060)
Just like you spoke about, there's fights,
Mark Zuckerberg (49:25.380)
arguments we get in and we don't like ourselves afterwards,
Lex Fridman (49:28.700)
but we got into them anyway.
Lex Fridman (49:30.340)
So our longterm growth is, I believe for most of us,
Lex Fridman (49:34.540)
has to do with learning, challenging yourself,
Mark Zuckerberg (49:38.220)
improving, being kind to each other,
Lex Fridman (49:40.960)
finding a community of people that you connect with
Mark Zuckerberg (49:47.620)
on a real human level, all that kind of stuff.
Lex Fridman (49:50.520)
But it does seem when you look at social media
Mark Zuckerberg (49:54.660)
that a lot of fights break out,
Lex Fridman (49:56.540)
a lot of arguments break out,
Mark Zuckerberg (49:58.180)
a lot of viral content ends up being sort of outrage
Lex Fridman (50:03.020)
in one direction or the other.
Lex Fridman (50:04.820)
And so it's easy from that to infer the narrative
Lex Fridman (50:08.020)
that social media companies are letting
Mark Zuckerberg (50:11.220)
this outrage become viral.
Lex Fridman (50:13.940)
And so they're increasing the division in the world.
Mark Zuckerberg (50:16.820)
I mean, perhaps you can comment on that
Lex Fridman (50:18.820)
or further, how can you be,
Lex Fridman (50:21.140)
how can you push back on this narrative?
Lex Fridman (50:25.780)
How can you be transparent about this battle?
Mark Zuckerberg (50:28.420)
Because I think it's not just motivation or financials,
Lex Fridman (50:33.540)
it's a technical problem too,
Mark Zuckerberg (50:36.000)
which is how do you improve longterm wellbeing
Lex Fridman (50:41.020)
of human beings?
Mark Zuckerberg (50:43.020)
I think that going through some of the design decisions
Lex Fridman (50:47.940)
would be a good conversation.
Lex Fridman (50:49.660)
But first, I actually think,
Lex Fridman (50:51.780)
I think you acknowledged that,
Mark Zuckerberg (50:54.260)
that narrative is somewhat anecdotal.
Lex Fridman (50:56.900)
And I think it's worth grounding this conversation
Mark Zuckerberg (50:59.500)
in the actual research that has been done on this,
Lex Fridman (51:02.600)
which by and large finds that social media
Lex Fridman (51:07.980)
is not a large driver of polarization, right?
Lex Fridman (51:10.780)
And, I mean, there's been a number of economists
Lex Fridman (51:14.820)
and social scientists and folks who have studied this.
Lex Fridman (51:18.380)
In a lot of polarization, it varies around the world.
Mark Zuckerberg (51:21.220)
If social media is basically in every country,
Lex Fridman (51:23.100)
Facebook's in pretty much every country
Mark Zuckerberg (51:24.580)
except for China and maybe North Korea.
Lex Fridman (51:27.180)
And you see different trends in different places
Mark Zuckerberg (51:32.460)
where in a lot of countries polarization is declining,
Lex Fridman (51:37.000)
in some it's flat, in the US it's risen sharply.
Lex Fridman (51:41.660)
So the question is, what are the unique phenomenon
Lex Fridman (51:44.660)
in the different places?
Lex Fridman (51:45.980)
And I think for the people who are trying to say,
Lex Fridman (51:47.580)
hey, social media is the thing that's doing this.
Mark Zuckerberg (51:50.220)
I think that that clearly doesn't hold up
Lex Fridman (51:52.940)
because social media is a phenomenon
Mark Zuckerberg (51:54.480)
that is pretty much equivalent
Lex Fridman (51:56.020)
in all of these different countries.
Lex Fridman (51:57.740)
And you have researchers like this economist at Stanford,
Lex Fridman (52:00.600)
Matthew Genskow, who has just written at length about this.
Lex Fridman (52:05.300)
And it's a bunch of books by political scientists,
Lex Fridman (52:10.420)
Ezra Klein and folks, why we're polarized,
Mark Zuckerberg (52:13.060)
basically goes through this decades long analysis in the US.
Lex Fridman (52:17.100)
Before I was born, basically talking about
Mark Zuckerberg (52:19.500)
some of the forces in kind of partisan politics
Lex Fridman (52:22.780)
and Fox News and different things
Mark Zuckerberg (52:25.160)
that predate the internet in a lot of ways
Lex Fridman (52:27.640)
that I think are likely larger contributors.
Lex Fridman (52:30.100)
So to the contrary on this,
Lex Fridman (52:32.220)
not only is it pretty clear that social media
Mark Zuckerberg (52:35.340)
is not a major contributor,
Lex Fridman (52:37.600)
but most of the academic studies that I've seen
Mark Zuckerberg (52:40.060)
actually show that social media use
Lex Fridman (52:42.640)
is correlated with lower polarization.
Lex Fridman (52:45.400)
And Genskow, the same person who just did the study
Lex Fridman (52:48.640)
that I cited about longitudinal polarization
Mark Zuckerberg (52:51.640)
across different countries,
Lex Fridman (52:54.180)
also did a study that basically showed
Mark Zuckerberg (52:57.480)
that if you looked after the 2016 election in the US,
Lex Fridman (53:02.120)
the voters who were the most polarized
Mark Zuckerberg (53:05.280)
were actually the ones who were not on the internet.
Lex Fridman (53:07.560)
So, and there have been recent other studies,
Mark Zuckerberg (53:10.280)
I think in Europe and around the world,
Lex Fridman (53:12.840)
basically showing that as people stop using social media,
Mark Zuckerberg (53:16.720)
they tend to get more polarized.
Lex Fridman (53:19.200)
Then there's a deeper analysis around,
Mark Zuckerberg (53:21.360)
okay, well, polarization actually isn't even one thing.
Lex Fridman (53:24.740)
Cause you know, having different opinions on something
Mark Zuckerberg (53:26.520)
isn't, I don't think that that's by itself bad.
Lex Fridman (53:28.900)
What people who study this say is most problematic
Mark Zuckerberg (53:33.920)
is what they call affective polarization,
Lex Fridman (53:35.920)
which is basically are you,
Lex Fridman (53:37.920)
do you have negative feelings towards people
Lex Fridman (53:40.040)
of another group?
Lex Fridman (53:41.040)
And the way that a lot of scholars study this
Lex Fridman (53:43.760)
is they basically ask a group,
Lex Fridman (53:46.780)
would you let your kids marry someone of group X?
Lex Fridman (53:50.600)
Whatever the groups are that you're worried
Mark Zuckerberg (53:53.320)
that someone might have negative feelings towards.
Lex Fridman (53:55.520)
And in general, use of social media
Mark Zuckerberg (53:58.160)
has corresponded to decreases
Lex Fridman (53:59.880)
in that kind of affective polarization.
Lex Fridman (54:01.960)
So I just wanna, I think we should talk
Lex Fridman (54:04.760)
through the design decisions and how we handle
Mark Zuckerberg (54:07.720)
the kind of specific pieces of content,
Lex Fridman (54:10.720)
but overall, I think it's just worth grounding
Mark Zuckerberg (54:13.280)
that discussion in the research that's existed
Lex Fridman (54:15.600)
that I think overwhelmingly shows
Mark Zuckerberg (54:17.440)
that the mainstream narrative around this
Lex Fridman (54:19.560)
is just not right.
Lex Fridman (54:21.040)
But the narrative does take hold
Lex Fridman (54:24.060)
and it's compelling to a lot of people.
Mark Zuckerberg (54:27.920)
There's another question I'd like to ask you on this.
Lex Fridman (54:31.300)
I was looking at various polls and saw that you're
Mark Zuckerberg (54:34.980)
one of the most disliked tech leaders today,
Lex Fridman (54:38.080)
54% unfavorable rating.
Mark Zuckerberg (54:41.400)
Elon Musk is 23%.
Lex Fridman (54:43.240)
It's basically everybody has a very high unfavorable rating
Mark Zuckerberg (54:46.260)
that are tech leaders.
Lex Fridman (54:48.000)
Maybe you can help me understand that.
Lex Fridman (54:50.640)
Why do you think so many people dislike you?
Lex Fridman (54:54.720)
Some even hate you.
Lex Fridman (54:56.880)
And how do you regain their trust and support?
Lex Fridman (54:59.160)
Given everything you just said,
Lex Fridman (55:02.360)
why are you losing the battle
Lex Fridman (55:05.360)
in explaining to people what actual impact
Lex Fridman (55:09.440)
social media has on society?
Lex Fridman (55:12.440)
Well, I'm curious if that's a US survey or world.
Mark Zuckerberg (55:16.760)
It is US, yeah.
Lex Fridman (55:17.960)
So I think that there's a few dynamics.
Mark Zuckerberg (55:19.360)
One is that our brand
Lex Fridman (55:24.360)
has been somewhat uniquely challenged in the US
Mark Zuckerberg (55:27.880)
compared to other places.
Lex Fridman (55:29.040)
It's not that there are.
Mark Zuckerberg (55:29.880)
I mean, other countries, we have issues too,
Lex Fridman (55:32.660)
but I think in the US, there was this dynamic where
Mark Zuckerberg (55:36.920)
if you look at like the next sentiment
Lex Fridman (55:38.880)
of kind of coverage or attitude towards us,
Mark Zuckerberg (55:42.880)
before 2016, I think that there were probably
Lex Fridman (55:44.880)
very few months, if any, where it was negative.
Lex Fridman (55:47.480)
And since 2016, I think that there probably
Lex Fridman (55:49.440)
been very few months, if any, then it's been positive.
Mark Zuckerberg (55:51.960)
Politics.
Lex Fridman (55:53.600)
But I think it's a specific thing.
Lex Fridman (55:55.360)
And this is very different from other places.
Lex Fridman (55:56.960)
So I think in a lot of other countries in the world,
Mark Zuckerberg (55:59.840)
the sentiment towards meta and our services
Lex Fridman (56:02.440)
is extremely positive.
Mark Zuckerberg (56:04.840)
In the US, we have more challenges.
Lex Fridman (56:06.600)
And I think compared to other companies,
Mark Zuckerberg (56:09.800)
you can look at certain industries,
Lex Fridman (56:12.520)
I think if you look at it from like a partisan perspective,
Mark Zuckerberg (56:16.320)
not from like a political perspective,
Lex Fridman (56:18.060)
but just kind of culturally,
Mark Zuckerberg (56:19.080)
it's like there are people who are probably
Lex Fridman (56:20.240)
more left of center and there are people
Mark Zuckerberg (56:21.360)
who are more right of center,
Lex Fridman (56:22.520)
and there's kind of blue America and red America.
Mark Zuckerberg (56:25.880)
There are certain industries that I think
Lex Fridman (56:27.660)
maybe one half of the country has a more positive view
Mark Zuckerberg (56:30.880)
towards than another.
Lex Fridman (56:32.200)
And I think we're in a,
Mark Zuckerberg (56:36.400)
one of the positions that we're in that I think
Lex Fridman (56:38.320)
is really challenging is that because of a lot
Mark Zuckerberg (56:41.720)
of the content decisions that we've basically
Lex Fridman (56:44.760)
had to arbitrate, and because we're not a partisan company,
Mark Zuckerberg (56:49.560)
we're not a Democrat company or a Republican company,
Lex Fridman (56:52.640)
we're trying to make the best decisions we can
Mark Zuckerberg (56:55.080)
to help people connect and help people have as much voice
Lex Fridman (56:58.480)
as they can while having some rules
Mark Zuckerberg (57:01.120)
because we're running a community.
Lex Fridman (57:04.920)
The net effect of that is that we're kind of constantly
Mark Zuckerberg (57:07.420)
making decisions that piss off people in both camps.
Lex Fridman (57:12.760)
And the effect that I've sort of seen is that
Mark Zuckerberg (57:17.160)
when we make a decision that is,
Lex Fridman (57:21.480)
that's a controversial one that's gonna upset,
Mark Zuckerberg (57:24.440)
say about half the country,
Lex Fridman (57:27.880)
those decisions are all negative sum,
Mark Zuckerberg (57:30.340)
from a brand perspective, because it's not like,
Lex Fridman (57:33.640)
if we make that decision in one way
Lex Fridman (57:35.640)
and say half the country is happy
Lex Fridman (57:37.840)
about that particular decision that we make,
Mark Zuckerberg (57:40.000)
they tend to not say, oh, sweet, meta got that one right.
Lex Fridman (57:43.720)
They're just like, ah, you didn't mess that one up.
Lex Fridman (57:46.160)
But their opinion doesn't tend to go up by that much.
Lex Fridman (57:48.960)
Whereas the people who kind of are on the other side of it
Lex Fridman (57:52.840)
are like, God, how could you mess that up?
Lex Fridman (57:55.080)
How could you possibly think that that piece of content
Lex Fridman (57:57.760)
is okay and should be up and should not be censored?
Lex Fridman (58:00.120)
Or, and so I think the, whereas if you leave it up
Mark Zuckerberg (58:04.940)
and, you know, it's, or if you take it down,
Lex Fridman (58:09.220)
the people who thought it should be taken down or,
Mark Zuckerberg (58:11.360)
you know, it's like, all right, fine, great.
Lex Fridman (58:12.740)
You didn't mess that one up.
Lex Fridman (58:14.080)
So our internal assessment of,
Lex Fridman (58:16.120)
and the kind of analytics on our brand
Mark Zuckerberg (58:17.920)
are basically anytime one of these big controversial things
Lex Fridman (58:20.540)
comes up in society,
Mark Zuckerberg (58:23.740)
our brand goes down with half of the country.
Lex Fridman (58:26.080)
And then like, if you,
Lex Fridman (58:27.600)
and then if you just kind of extrapolate that out,
Lex Fridman (58:29.600)
it's just been very challenging for us to try to navigate
Lex Fridman (58:33.200)
what is a polarizing country in a principled way,
Lex Fridman (58:36.640)
where we're not trying to kind of hew to one side
Mark Zuckerberg (58:38.600)
or the other, we're trying to do
Lex Fridman (58:39.440)
what we think is the right thing.
Lex Fridman (58:41.040)
But that's what I think is the right thing
Lex Fridman (58:43.220)
for us to do though.
Mark Zuckerberg (58:44.060)
So, I mean, that's what we'll try to keep doing.
Lex Fridman (58:47.360)
Just as a human being, how does it feel though,
Mark Zuckerberg (58:50.160)
when you're giving so much of your day to day life
Lex Fridman (58:53.380)
to try to heal division, to try to do good in the world,
Mark Zuckerberg (58:58.060)
as we've talked about, that so many people in the US,
Lex Fridman (59:02.260)
the place you call home have a negative view
Mark Zuckerberg (59:06.480)
of you as a leader, as a human being
Lex Fridman (59:09.120)
and the company you love?
Mark Zuckerberg (59:14.040)
Well, I mean, it's not great,
Lex Fridman (59:15.880)
but I mean, look, if I wanted people to think positively
Mark Zuckerberg (59:21.040)
about me as a person,
Lex Fridman (59:25.760)
I don't know, I'm not sure if you go build a company.
Mark Zuckerberg (59:27.960)
I mean, it's like.
Lex Fridman (59:28.800)
Or a social media company.
Mark Zuckerberg (59:30.280)
It seems exceptionally difficult to do
Lex Fridman (59:32.040)
with a social media company.
Mark Zuckerberg (59:32.880)
Yeah, so, I mean, I don't know,
Lex Fridman (59:34.800)
there is a dynamic where a lot of the other people
Mark Zuckerberg (59:39.760)
running these companies, internet companies,
Lex Fridman (59:42.160)
have sort of stepped back and they just do things
Mark Zuckerberg (59:45.320)
that are sort of, I don't know, less controversial.
Lex Fridman (59:49.480)
And some of it may be that they just get tired over time.
Mark Zuckerberg (59:52.720)
But, you know, it's, so I don't know.
Lex Fridman (59:55.400)
I think that, you know, running a company is hard,
Mark Zuckerberg (59:58.120)
building something at scale is hard.
Lex Fridman (59:59.880)
You only really do it for a long period of time
🔗 相关节目