The incredible inventions of intuitive AI | Maurice Conti

Translator: Leslie Gauthier Reviewer: Camille Martínez

譯者: 易帆 余 審譯者: SF Huang

How many of you are creatives,

你們有多少人是創意人、

designers, engineers, entrepreneurs, artists,

設計師、工程師、企業家、藝術家？

or maybe you just have a really big imagination?

或者你只是有無遠弗屆的想像力？

Show of hands? (Cheers)

請舉一下手？（歡呼聲）

That's most of you.

現場大部分人都是。

I have some news for us creatives.

我有一些消息要給我們的創意人。

Over the course of the next 20 years,

接下來的 20 年，

more will change around the way we do our work

很多我們工作的方式，

than has happened in the last 2,000.

將會遠遠不同於過去的 2000 年。

In fact, I think we're at the dawn of a new age in human history.

實際上，我認為我們正處在 人類歷史新世代的黎明。

Now, there have been four major historical eras defined by the way we work.

人類工作的方式， 有四個主要的歷史階段。

The Hunter-Gatherer Age lasted several million years.

人類歷經了幾百萬年的 狩獵採集時代。

And then the Agricultural Age lasted several thousand years.

然後經歷了幾千年的農業時代。

The Industrial Age lasted a couple of centuries.

工業時代則延續了幾世紀。

And now the Information Age has lasted just a few decades.

而目前的資訊時代才走了幾十年。

And now today, we're on the cusp of our next great era as a species.

如今，身為人類的我們， 即將邁入下一個偉大的時代。

Welcome to the Augmented Age.

歡迎來到「擴增時代」。

In this new era, your natural human capabilities are going to be augmented

在這個新時代， 人類天生的能力將會被強化擴增，

by computational systems that help you think,

電腦計算系統將幫助你思考、

robotic systems that help you make,

機械人系統協助你製造、

and a digital nervous system

遠超過你自然感官強度的 數位神經系統，

that connects you to the world far beyond your natural senses.

能夠讓你與全世界接軌。

Let's start with cognitive augmentation.

我們先從「認知擴增」談起。

How many of you are augmented cyborgs?

現場有多少人是「強化的半機械人」？

(Laughter)

（笑聲）

I would actually argue that we're already augmented.

其實我想說的是， 我們都已經被強化、擴增了。

Imagine you're at a party,

想像你正在參加一場派對，

and somebody asks you a question that you don't know the answer to.

有人問了你一個 你不知道如何回答的問題。

If you have one of these, in a few seconds, you can know the answer.

如果你有這個，只要幾秒鐘， 你就會得到答案。

But this is just a primitive beginning.

但這也只是剛開始而已。

Even Siri is just a passive tool.

甚至 Siri 也只是個被動工具。

In fact, for the last three-and-a-half million years,

實際上，在過去的 350 萬年，

the tools that we've had have been completely passive.

我們所有的工具都是被動的。

They do exactly what we tell them and nothing more.

它們只會照我們的指令去做， 僅此而已。

Our very first tool only cut where we struck it.

我們最早使用的工具， 遵循一個口令一個動作的指示。

The chisel only carves where the artist points it.

藝術家指哪裡，雕刻刀就雕刻哪裡。

And even our most advanced tools do nothing without our explicit direction.

即使最先進的工具，如果沒有 我們明確的指令也不會工作。

In fact, to date, and this is something that frustrates me,

說真的，時到今日， 有件事仍讓我感覺很挫敗，

we've always been limited

我們一直以來都被限制在

by this need to manually push our wills into our tools --

「需要動手將我們的意念 傳達給工具」的這種迷思框框中——

like, manual, literally using our hands,

就是得動手去做，即使有了電腦 還是得靠雙手。

even with computers.

But I'm more like Scotty in "Star Trek."

但我還是比較喜歡當 <<星際迷航>>裡的史考迪。

(Laughter)

（笑聲）

I want to have a conversation with a computer.

我也想跟電腦對話。

I want to say, "Computer, let's design a car,"

當我說，「電腦， 我們來設計一輛車吧！」

and the computer shows me a car.

然後電腦就會顯示一輛車給我看。

And I say, "No, more fast-looking, and less German,"

然後我說：「不，要拉風一點， 德國味兒少一點。」

and bang, the computer shows me an option.

接著「蹦」， 電腦給了我一個新選擇。

(Laughter)

（笑聲）

That conversation might be a little ways off,

這樣的對話可能有點不切實際，

probably less than many of us think,

也許沒有我們認為的那麼不切實際，

but right now,

但現在，

we're working on it.

我們正在做這件事。

這些工具將帶領我們大躍進， 從被動轉為衍生。

Tools are making this leap from being passive to being generative.

Generative design tools use a computer and algorithms

「衍生設計工具 」 是利用電腦及演算法，

to synthesize geometry

合成出幾何結構，

to come up with new designs all by themselves.

產製出新的設計圖， 全部都是它們自己構思出來的。

All it needs are your goals and your constraints.

你只需要設定目標及限制條件。

I'll give you an example.

我給各位舉個例子。

In the case of this aerial drone chassis,

就拿這個無人機底盤為例，

all you would need to do is tell it something like,

你唯一要做的， 就是告訴它你的需求，

it has four propellers,

像是，你要四個螺旋槳的，

you want it to be as lightweight as possible,

它越輕越好，

and you need it to be aerodynamically efficient.

空氣動力學表現效率佳的。

Then what the computer does is it explores the entire solution space:

電腦做的，就是探索 所有可能的解決方案：

every single possibility that solves and meets your criteria --

每一個能解決且符合你標準的 可能方案——

millions of them.

有上百萬個。

It takes big computers to do this.

這需要大型電腦才能做到。

But it comes back to us with designs

但它回饋給我們的設計方案，

that we, by ourselves, never could've imagined.

是我們單憑自己無法想像出來的 設計方案。

And the computer's coming up with this stuff all by itself --

電腦憑藉著自己的能力 做出這些東西——

no one ever drew anything,

我們人類沒有動筆畫任何東西，

and it started completely from scratch.

完全是它自己從頭、從零畫起的。

And by the way, it's no accident

順便一提，這可不是偶然......

that the drone body looks just like the pelvis of a flying squirrel.

無人機的機體長的像飛鼠的骨盆，

(Laughter)

（笑聲）

It's because the algorithms are designed to work

那是因為演算法的計算模式，

the same way evolution does.

是遵循生物演化模式而設計的。

What's exciting is we're starting to see this technology

令人興奮的是， 我們開始見證這樣的科技

out in the real world.

在現實世界中實現。

We've been working with Airbus for a couple of years

我們與空中巴士 （歐洲最大飛機製造商）

on this concept plane for the future.

合作開發未來的概念機 已經好幾年了，

It's a ways out still.

這計畫目前還在進行。

But just recently we used a generative-design AI

但最近，我們用了 衍生設計的人工智慧

to come up with this.

做出了這一個。

This is a 3D-printed cabin partition that's been designed by a computer.

這是一個 3D 列印的客艙隔間板， 由一台電腦所設計。

It's stronger than the original yet half the weight,

它比原款式還要堅固， 但重量只有原本的一半，

and it will be flying in the Airbus A320 later this year.

今年稍晚，它將跟 A320 空中巴士一起飛上天。

So computers can now generate;

所以現在電腦會主動生成、衍生了；

they can come up with their own solutions to our well-defined problems.

它們可以對界定明確的問題， 給出自己的答案。

But they're not intuitive.

但它們並不是靠直覺做事。

They still have to start from scratch every single time,

它們還是每次都得從頭開始，

and that's because they never learn.

因為它們不會學習。

Unlike Maggie.

不像瑪姬。

(Laughter)

（笑聲）

Maggie's actually smarter than our most advanced design tools.

瑪姬其實比我們最先進的 設計工具都還要聰明。

What do I mean by that?

這是什麼意思？

If her owner picks up that leash,

如果狗主人拿起狗鍊，

Maggie knows with a fair degree of certainty

瑪姬就知道有相當的確定性，

it's time to go for a walk.

主人要帶她去散步了。

And how did she learn?

她是怎麼知道的？

Well, every time the owner picked up the leash, they went for a walk.

因為，每當主人拿起狗鍊， 他們就會一起去散步。

And Maggie did three things:

瑪姬會做三件事：

she had to pay attention,

她必須專注、

she had to remember what happened

必須記得發生過什麼事、

and she had to retain and create a pattern in her mind.

必須在腦中記憶並產生一個模式。

Interestingly, that's exactly what

有趣的是，這正是電腦科學家

computer scientists have been trying to get AIs to do

過去 60 年來，一直嘗試 要讓人工智慧做的事。

for the last 60 or so years.

Back in 1952,

回想一下 1952 年，

they built this computer that could play Tic-Tac-Toe.

科學家建立了這一台電腦， 它會玩井字遊戲。

Big deal.

真了不起。

Then 45 years later, in 1997,

45 年後，1997 年，

Deep Blue beats Kasparov at chess.

深藍擊敗了當時的西洋棋世界冠軍 卡司帕洛夫，

2011, Watson beats these two humans at Jeopardy,

2011年，華生 (IBM電腦) 在<<危機邊緣>>擊敗這兩個人，

which is much harder for a computer to play than chess is.

對電腦來說，這比下棋難多了。

In fact, rather than working from predefined recipes,

事實上，華生並不是從 預先定義的題庫中來找答案，

Watson had to use reasoning to overcome his human opponents.

它必須使用推理 來擊敗它的人類對手。

And then a couple of weeks ago,

就在幾個禮拜前，

DeepMind's AlphaGo beats the world's best human at Go,

DeepMind 的阿爾法圍棋 擊敗了世界圍棋冠軍，

which is the most difficult game that we have.

而圍棋是我們人類最複雜的遊戲。

In fact, in Go, there are more possible moves

事實上，圍棋走法的可能性

than there are atoms in the universe.

超過全宇宙的原子數量。

So in order to win,

所以為了取得勝利，

what AlphaGo had to do was develop intuition.

阿爾法圍棋必須學會使用直覺。

And in fact, at some points, AlphaGo's programmers didn't understand

實際上，有些下法， 阿爾法圍棋的程式人員也不懂

why AlphaGo was doing what it was doing.

為什麼阿爾法圍棋要那樣下。

And things are moving really fast.

世界變化真快。

I mean, consider -- in the space of a human lifetime,

我的意思是，想像一下—— 在人類壽命這麼長的時間裡，

computers have gone from a child's game

電腦已經從小孩子的遊戲

to what's recognized as the pinnacle of strategic thought.

發展到策略思考的頂尖水平。

What's basically happening

電腦基本上的發展，

is computers are going from being like Spock

已經從史巴克大副進化到

to being a lot more like Kirk.

寇克艦長。

(Laughter)

（笑聲）

Right? From pure logic to intuition.

對吧？從純粹的邏輯運算 到直覺判斷。

Would you cross this bridge?

你們會跨過這座橋嗎？

Most of you are saying, "Oh, hell no!"

大部分人應該都說， 「喔，打死我也不要！」

(Laughter)

（笑聲）

And you arrived at that decision in a split second.

你瞬間就可以做出這個決定。

You just sort of knew that bridge was unsafe.

你就是隱約知道那座橋並不安全。

And that's exactly the kind of intuition

這種直覺判斷，

that our deep-learning systems are starting to develop right now.

就是目前我們深度學習系統 正在發展的能力。

Very soon, you'll literally be able

很快的，各位就可以

to show something you've made, you've designed,

把你製作、設計出來的東西

to a computer,

拿給電腦評判，

and it will look at it and say,

然後它看完後會說，

"Sorry, homie, that'll never work. You have to try again."

「抱歉，兄弟，這東西行不通， 你再試試別的吧！」

Or you could ask it if people are going to like your next song,

或者你可以問它， 人們會不會喜歡你的新歌？

or your next flavor of ice cream.

或者你冰淇淋的新口味？

Or, much more importantly,

再或者，更重要的，

you could work with a computer to solve a problem

你可以跟電腦一起解決

我們從未面臨過的問題。

that we've never faced before.

For instance, climate change.

例如，氣候變遷問題。

We're not doing a very good job on our own,

我們自己沒有做得很好的事，

we could certainly use all the help we can get.

我們當然可以利用身邊 各種資源來幫忙解決。

That's what I'm talking about,

這就是我接下來要談的，

technology amplifying our cognitive abilities

科技強化了我們的認知能力，

so we can imagine and design things that were simply out of our reach

所以我們可以想像並設計出， 當我們還未具有強化擴增能力時

as plain old un-augmented humans.

所未能創造出來的東西。

So what about making all of this crazy new stuff

那麼，製造這些我們即將發明設計的

that we're going to invent and design?

瘋狂新產品會如何呢？

I think the era of human augmentation is as much about the physical world

我認為在人類擴增的時代， 現實世界

及虛擬智慧領域 與其皆有不分軒輊的重要相關性。

as it is about the virtual, intellectual realm.

How will technology augment us?

科技將會如何強化我們？

In the physical world, robotic systems.

在現實世界，就是機械人系統。

OK, there's certainly a fear

沒錯，很多人擔心，

that robots are going to take jobs away from humans,

機械人會搶走人類的工作，

and that is true in certain sectors.

在某些領域，確實是如此。

But I'm much more interested in this idea

但，我對以下的想法比較有興趣，

that humans and robots working together are going to augment each other,

就是，人類與機械人 將會一起工作並互相強化，

and start to inhabit a new space.

並開創出一種新的共生空間。

This is our applied research lab in San Francisco,

這是我們在舊金山的 應用研究實驗室，

where one of our areas of focus is advanced robotics,

我們專研的領域之一就是 高階機械人，

specifically, human-robot collaboration.

特別是人機合作的領域。

And this is Bishop, one of our robots.

這是畢夏普， 我們其中的一個機器人。

As an experiment, we set it up

在實驗裡，我將它設定為

to help a person working in construction doing repetitive tasks --

在建築領域中， 幫助人類做重複性的工作——

tasks like cutting out holes for outlets or light switches in drywall.

比如說，在石牆上打出一個 插座孔或電燈開關孔。

(Laughter)

（笑聲）

So, Bishop's human partner can tell what to do in plain English

所以，畢夏普的人類夥伴 就可以用簡單的英語和手勢

and with simple gestures,

告訴它該做什麼，

kind of like talking to a dog,

有點像是在跟狗狗說話。

and then Bishop executes on those instructions

然後畢夏普會以完美的準確度

with perfect precision.

執行人類所下達的指令。

We're using the human for what the human is good at:

我們讓人類做人類擅長的事，像是：

awareness, perception and decision making.

需要意識力、洞察力、做決策的工作。

And we're using the robot for what it's good at:

我們讓機械人做機械人擅長的事， 像是：

precision and repetitiveness.

準確度及重複性的工作。

Here's another cool project that Bishop worked on.

畢夏普還有另一個很酷的專案。

The goal of this project, which we called the HIVE,

這個專案的目標， 我們稱它為<<蜂巢>>，

was to prototype the experience of humans, computers and robots

主要目標是把人類、電腦、 機械人的經驗結合起來，

all working together to solve a highly complex design problem.

一起工作解決極複雜的設計問題。

The humans acted as labor.

人類的工作是

They cruised around the construction site, they manipulated the bamboo --

在建築基地巡邏監工 並熟練地操作竹子——

which, by the way, because it's a non-isomorphic material,

順便一提，因為每一根竹子的 材料性質都不一樣，

is super hard for robots to deal with.

所以機械人操作起來非常困難。

But then the robots did this fiber winding,

但機械人做的是彎曲竹子的纖維，

which was almost impossible for a human to do.

這種事人類幾乎做不來。

And then we had an AI that was controlling everything.

然後我們讓一台人工智慧 來控制所有的東西。

It was telling the humans what to do, telling the robots what to do

它會告訴人類要做什麼， 告訴機械人要做什麼，

and keeping track of thousands of individual components.

並且對成千上萬個部件 進行持續的追蹤。

What's interesting is,

有趣的是，

building this pavilion was simply not possible

要建造出這樣的亭狀建築物，

without human, robot and AI augmenting each other.

如果沒有人類、機械、人工智慧的 互補強化，根本不可能做得出來。

OK, I'll share one more project. This one's a little bit crazy.

好，我再分享一個專案， 這個有點瘋狂。

We're working with Amsterdam-based artist Joris Laarman and his team at MX3D

我們與阿姆斯特丹的藝術家 尤爾斯‧拉曼和他的 MX3D 團隊，

to generatively design and robotically print

正使用衍生性設計 與機械列印的方式，

the world's first autonomously manufactured bridge.

打造世界第一座機械人自造的橋梁。

So, Joris and an AI are designing this thing right now, as we speak,

所以，就在我們談話的這一刻，

尤爾斯正和人工智慧一起 在阿姆斯特丹設計這座橋梁。

in Amsterdam.

And when they're done, we're going to hit "Go,"

等他們設計完成後， 我們就會按下「啟動」開關，

and robots will start 3D printing in stainless steel,

讓機械人開始 用不鏽鋼 3D 列印出橋梁，

and then they're going to keep printing, without human intervention,

在沒有人類的介入幫忙下， 它們會持續地列印

until the bridge is finished.

直到橋樑完工為止。

So, as computers are going to augment our ability

所以，電腦將強化

to imagine and design new stuff,

我們的想像及設計新事物的能力，

robotic systems are going to help us build and make things

機械人系統將協助我們製造

that we've never been able to make before.

我們以前無法製造的東西。

But what about our ability to sense and control these things?

但是我們感知和控制 這些東西的能力呢？

What about a nervous system for the things that we make?

我們製成東西的神經系統 又如何呢？

Our nervous system, the human nervous system,

我們的神經系統，人類的神經系統，

tells us everything that's going on around us.

可以告訴我們周遭發生的每一件事。

But the nervous system of the things we make is rudimentary at best.

但這些東西的神經系統， 最多只能算「尚未成熟」。

For instance, a car doesn't tell the city's public works department

比如說，車輛本身 不會主動通告市政府的工部門，

that it just hit a pothole at the corner of Broadway and Morrison.

說它在經過百老匯和 莫里森轉角口時撞到水坑。

A building doesn't tell its designers

建築物本身不會告知它的設計師，

whether or not the people inside like being there,

裡面的居民是否喜歡住在那裏，

and the toy manufacturer doesn't know

玩具製造商也不知道

他們的玩具 現在是跟誰在玩、在哪玩、

if a toy is actually being played with --

how and where and whether or not it's any fun.

是不是玩的很開心。

Look, I'm sure that the designers imagined this lifestyle for Barbie

我確定設計師在設計芭比時，

一定想像過芭比的生活方式。

when they designed her.

(Laughter)

（笑聲）

But what if it turns out that Barbie's actually really lonely?

但要是芭比變的很孤單怎麼辦？

(Laughter)

（笑聲）

If the designers had known

如果設計師知道

what was really happening in the real world

他們設計的東西， 在真實世界裡發生了什麼事，

with their designs -- the road, the building, Barbie --

像是道路、建築物、芭比——

they could've used that knowledge to create an experience

那他們就可以運用所獲得的訊息，

that was better for the user.

為使用者創造出更好的使用體驗。

What's missing is a nervous system

我們欠缺的就是一個

connecting us to all of the things that we design, make and use.

可以連結所有我們設計、製造、 使用事物的神經系統。

What if all of you had that kind of information flowing to you

如果大家在真實世界，能收到 自己創造的東西所回饋的資訊，

from the things you create in the real world?

那會如何呢？

With all of the stuff we make,

所有我們製造的東西，

we spend a tremendous amount of money and energy --

我們花了很多錢跟精力 ──

實際上光是去年， 大約就有兩兆美金 ──

in fact, last year, about two trillion dollars --

convincing people to buy the things we've made.

去說服人們購買我們製造的東西。

But if you had this connection to the things that you design and create

但如果你所設計製造出來的東西 能連結傳送給你回饋的訊息，

after they're out in the real world,

不管是在它們上市以後，

after they've been sold or launched or whatever,

或是在賣出或發表以後，

we could actually change that,

我們就可以改變既有的銷售模式，

and go from making people want our stuff,

從說服人們來購買我們的產品，

to just making stuff that people want in the first place.

轉變成我們第一時間就做出 人們真正需要的東西。

The good news is, we're working on digital nervous systems

好消息是，我們正在研發的 這套數位神經系統

that connect us to the things we design.

能連結我們與我們所設計的產品。

We're working on one project

我們與在洛杉磯的

with a couple of guys down in Los Angeles called the Bandito Brothers

邦帝圖兄弟公司和他們的團隊

and their team.

正合作進行一個專案。

And one of the things these guys do is build insane cars

這幾個人做的其中一件事 就是製造「瘋狂賽車」，

that do absolutely insane things.

他們做的東西真的很瘋狂。

These guys are crazy --

這些人真的是瘋了──

(Laughter)

（笑聲）

in the best way.

不過是用最厲害的方式。

And what we're doing with them

我們跟他們一起合作的模式，

is taking a traditional race-car chassis

就是將傳統的賽車底盤

and giving it a nervous system.

安裝神經系統。

So we instrumented it with dozens of sensors,

所以，我們在底盤 安裝了好幾組感應器，

put a world-class driver behind the wheel,

然後請一位世界級車手來駕駛，

took it out to the desert and drove the hell out of it for a week.

把車送到沙漠連續開它個一禮拜。

And the car's nervous system captured everything

之後車子的神經系統

就能紀錄到車子發生的所有反應。

that was happening to the car.

We captured four billion data points;

我們抓到了 40 億個資料點；

all of the forces that it was subjected to.

所有底盤所承受的壓力數據。

And then we did something crazy.

然後我們做了一些瘋狂的事。

We took all of that data,

我們把所有的資料

and plugged it into a generative-design AI we call "Dreamcatcher."

連接到一個叫做「捕夢者」的 衍生設計人工智慧上 。

So what do get when you give a design tool a nervous system,

所以當你把神經系統 安裝到設計工具上，

and you ask it to build you the ultimate car chassis?

並請它幫你建造一個 終極汽車底盤時，你會得到什麼？

You get this.

你會得到這個。

This is something that a human could never have designed.

這是人類永遠無法設計出的東西。

Except a human did design this,

如果真有人這樣設計過，

but it was a human that was augmented by a generative-design AI,

那個人一定也是透過 衍生設計的人工智慧強化、

a digital nervous system

數位神經系統的強化、

and robots that can actually fabricate something like this.

和機械人一起合作， 才做得出來的東西。

So if this is the future, the Augmented Age,

所以，如果擴增時代 就是我們的未來，

and we're going to be augmented cognitively, physically and perceptually,

而我們的認知、體格、知覺 都將被強化、擴增，

what will that look like?

那會是怎樣的世界？

What is this wonderland going to be like?

那會是個什麼樣的美麗新世界？

I think we're going to see a world

我認為我們即將見證這麼一個世界，

where we're moving from things that are fabricated

一個東西從製造出來的變成

to things that are farmed.

「種 」出來的世界。

Where we're moving from things that are constructed

一個東西從建造出來的

to that which is grown.

變成自己「長 」出來的世界。

We're going to move from being isolated

我們將從自我隔離

to being connected.

轉變成相互交流。

And we'll move away from extraction

我們也將從奪取者

to embrace aggregation.

變成相互擁抱的給予者。

I also think we'll shift from craving obedience from our things

我也認為，我們將會從冀望產品 順從我們的指令，

to valuing autonomy.

轉變成重視其自主性。

Thanks to our augmented capabilities,

由於我們的擴增強化能力，

our world is going to change dramatically.

我們的世界將會有劇烈的變化。

We're going to have a world with more variety, more connectedness,

我們的世界會變得更多元、 更加連通、

more dynamism, more complexity,

更有活力、更多複雜的變化、

more adaptability and, of course,

更有適應力、當然

more beauty.

也會更美麗。

The shape of things to come

未來世界的雛型

will be unlike anything we've ever seen before.

是我們前所未見的。

Why?

為什麼？

Because what will be shaping those things is this new partnership

因為形塑這個世界的

between technology, nature and humanity.

將會是科技、自然與人類的 新結盟關係。

That, to me, is a future well worth looking forward to.

對我而言，那樣的未來 是值得我們期待的。

Thank you all so much.

非常感謝各位。

(Applause)

（掌聲）