监视资本主义：智能陷阱（2020）

They were different when newspapers came in and the printing press came in,
电视问世的时候 与现在很不同
and they were different when television came in,
当时有三个主要的…
and you had three major networks and...
-网络 -当时 我也想说
At the time. -At the time. That's what I'm saying.
但我在说 这是一个全新的层次
But I'm saying the idea that there's a new level
这个新的层次 以前也发生过很多次
and that new level has happened so many times before.
这只是我们看见的 最新的一个层次
I mean, this is just the latest new level that we've seen.
有这样一种说法：“我们去适应它
There's this narrative that, you know, "We'll just adapt to it.
我们要学着与这些设备共存
We'll learn how to live with these devices,
就像我们学着 和其他所有事物共存一样”
just like we've learned how to live with everything else."
但这个说法没有注意到的是 有些东西明显是全新的
And what this misses is there's something distinctly new here.
或许这其中最危险的是
Perhaps the most dangerous piece of all this is the fact
这是由技术驱动的 在成指数地向前发展
that it's driven by technology that's advancing exponentially.
（计算机处理能力）
大体上 如果你看从20世纪60年代至今
Roughly, if you say from, like, the 1960s to today,
计算机处理能力增长了万亿倍
processing power has gone up about a trillion times.
我们身边没有任何其他东西 以这个速率增长
Nothing else that we have has improved at anything near that rate.
比如 汽车速度基本上才实现翻倍
Like, cars are, you know, roughly twice as fast.
几乎所有其他的东西都显得微不足道
And almost everything else is negligible.
或许最重要的是
And perhaps most importantly,
我们人类… 我们的生理 我们的大脑 根本没有丝毫进化
our humanour physiology, our brains have evolved not at all.
（没有手♥机♥的时间）
人类的思想、身体和体质
Human beings, at a mind and body and sort of physical level,
基本上不会改变了
are not gonna fundamentally change.
我们可以搞基因工程 在未来开♥发♥新的人类物种
We can do genetic engineering and develop new kinds of human beings,
但是现实来讲 你生活在大脑这个硬件下
but realistically speaking, you're living inside of hardware, a brain,
已经存在几百万年了
that was, like, millions of years old,
然后出现了这样一个屏幕 在屏幕的另一端
and then there's this screen, and then on the opposite side of the screen,
有上千名工程师和超级计算机
there's these thousands of engineers and supercomputers
有着与你不同的目标
that have goals that are different than your goals,
那么 这个游戏谁能赢呢？谁会赢？
and so, who's gonna win in that game? Who's gonna win?
我们怎么会输？
How are we losing?
-我不知道 -他在哪里？这太不正常了
I don't know. -Where is he? This is not normal.
我给他推送朋友和家人的内容太多 他烦了吗？
Did I overwhelm him with friends and family content?
-或许吧 -或许是广♥告♥太多了
Probably. -Well, maybe it was all the ads.
不 一定出了严重的问题 我们切换到复苏模式吧
No. Something's very wrong. Let's switch to resurrection mode.
当你想到人工智能 人工智能会毁掉世界
When you think of AI, you know, an AI's gonna ruin the world,
你会看到《终结者》 看到施瓦辛格…
and you see, like, a Terminator, and you see Arnold Schwarzenegger.
我会回来的
I'll be back.
…你会看到无人机 你觉得
You see drones, and you think, like,
“人工智能会杀人的”
" we're gonna kill people with AI."
人们忽略的是 人工智能 现在已经在运营着当今世界了
And what people miss is that AI already runs today's world right now.
甚至谈论“人工智能”都只是暗喻
Even talking about "an AI" is just a metaphor.
在谷歌♥这种公♥司♥ 有超级大的房♥间
At these companies like... like Google, there's just massive, massive rooms,
有些在地下 有些在水下
some of them underground, some of them underwater,
房♥间里全是电脑
of just computers.
无数的电脑 连绵不绝
Tons and tons of computers, as far as the eye can see.
它们互相之间在内部深度连接
They're deeply interconnected with each other
在运行着极其复杂的程序
and running extremely complicated programs,
始终不停地在彼此之间交换信息
sending information back and forth between each other all the time.
他们会运行很多不同的程序
And they'll be running many different programs,
在同样的机器上 有不同的产品
many different products on those same machines.
有些东西可以被描述为简单算法
Some of those things could be described as simple algorithms,
有些算法太过复杂
some could be described as algorithms
就可以被称为“智能”
that are so complicated, you would call them intelligence.
我想说 算法是内嵌在代码中的观点…
I like to say that algorithms are opinions embedded in code...
算法并不是客观的
and that algorithms are not objective.
算法被某种成功的定义优化
Algorithms are optimized to some definition of success.
所以 如果你能想象
So, if you can imagine,
一个商业公♥司♥成功的定义
if a... if a coercial enterprise builds an algorithm
需要依靠算法
to their definition of success,
那就是商业利益
it's a coercial interest.
通常都有利润
It's usually profit.
你给电脑一个目标 说“我想要这个结果”
You are giving the computer the goal state, "I want this outcome,"
电脑自己去学习 怎样实现
and then the computer itself is learning how to do it.
这是“机器学习”概念的由来
That's where the term "machine learning" comes from.
所以 每一天 都会更好一点
And so, every day, it gets slightly better
在正确命令获取正确的推送数据
at picking the right posts in the right order
让你在这个产品上花的时间越来越多
so that you spend longer and longer in that product.
没人能真正明白 为了实现这个目标
And no one really understands what they're doing
他们在做什么
in order to achieve that goal.
算法有着自己的思想 虽然是人写的
The algorithm has a mind of its own, so even though a person writes it,
它写出来的目的
it's written in a way
是你建立一个机器 这个机器会自己改变
that you kind of build the machine, and then the machine changes itself.
这种公♥司♥ 员工非常少
There's only a handful of people at these companies,
在脸书、推特和其他公♥司♥
at Facebook and Twitter and other companies...
只有几个人能明白 这些系统的工作原理
There's only a few people who understand how those systems work,
虽然他们不需要完全理解
and even they don't necessarily fully understand
某一条特定的内容 会发生什么
what's gonna happen with a particular piece of content.
作为人类 我们几乎 已经失去了对这些系统的控制
So, as humans, we've almost lost control over these systems.
因为是它们在控制我们看到的信息
Because they're controlling, you know, the information that we see,
更多的是它们在控制我们 而不是我们控制它们
they're controlling us more than we're controlling them.
在他的地理区域
Cross-referencing him
对他和可以对比的人 进行交叉参照
against comparables in his geographic zone.
他的心理测定相似者
His psychometric doppelgangers.
在他的地区 有13694个和他行为相似的人
There are 13,694 people behaving just like him in his region.
-他们中盛行什么？ -我们需要真实的好东西
What's trending with them? -We need something actually good
才能进行有效的复苏
for a proper resurrection,
因为平常的那些东西已经不起作用了
given that the typical stuff isn't working.
学校那个可爱的姑娘都没用了
Not even that cute girl from school.
我的分♥析♥显示 用极端中心内容搞政♥治♥
My analysis shows that going political with Extreme Center content
有62.3%的几率能够获得长♥期♥参与
has a 62.3 percent chance of long-term engagement.
还不错
That's not bad.
想用它引导 还不太够
It's not good enough to lead with.
好 所以我们已经试过 通知给他圈人照片
Okay, okay, so we've tried notifying him about tagged photos,
邀请、实事、甚至是瑞贝卡的私信
invitations, current events, even a direct message from Rebecca.
但是用户01265923010呢？
But what about User 01265923010?
是 本点赞了她所有的发帖
Yeah, Ben loved all of her posts.
几个月的所有发帖 真的点了所有 然后就没有然后了
For months and, like, literally all of them, and then nothing.
我算出了通知安娜的内容
I calculate a 92.3 percent chance of resurrection
会有92.3%的复苏概率
with a notification about Ana.
（新感情）
还有她的新朋友
And her new friend.
（没有手♥机♥的时间）
（你的前女友有了新感情！）
不是吧
you gotta be kiddin' me.
好吧
Okay.
（安娜与路易斯正在热恋）
什么？
What?
当！我们回来了！
Bam! We're back!
我们继续挣钱 兄弟们
Let's get back to making money, boys.
好 让他们和整个世界联♥系♥起来
Yes, and connecting Ben with the entire world.
我给他看所有他可能喜欢的信息
I'm giving him access to all the information he might like.
你们是否想过 这些推送对本是好的吗？
Hey, do you guys ever wonder if, you know, like, the feed is good for Ben?
-没想过 -没有
No. -No.
我在你身上下了咒语
♪ I put a spell on you ♪
因为你是我的
♪ 'Cause you're mine ♪
你最好停止你做的事情
♪ You better stop the things you do ♪
我不骗你
♪ I ain't lyin' ♪
（A/B测试 极端中心）
不 我不骗你
♪ No, I ain't lyin' ♪
你知道我无法忍♥受
♪ You know I can't stand it ♪
你在四处奔跑
♪ You're runnin' around ♪
爸爸 你更清楚
♪ You know better, Daddy ♪
我无法忍♥受 因为你将我放下
♪ I can't stand it 'Cause you put me down ♪
耶
♪ Yeah, yeah ♪
我在你身上下了咒语
♪ I put a spell on you ♪
因为你是我的
♪ Because you're mine ♪
你是我的
♪ You're mine ♪
想象一下 你在用脸书…
So, imagine you're on Facebook...
你的对手是人工智能
and you're effectively playing against this artificial intelligence
它知道你的一切
that knows everything about you,
能够预测你未来的举动 你对它却一无所知
can anticipate your next move, and you know literally nothing about it,
除了上面有猫的视频和出生日期
except that there are cat videos and birthdays on it.