With AI, Anyone Can Be a Coder Now | Thomas Dohmke | TED

譯者: 麗玲 辛

You know, I'm one of these adults

你知道嗎？我是那種到現在 還很喜歡玩樂高的大人。

that actually still loves playing with LEGO.

I loved them way back in the '80s in Berlin when I grew up,

早在 80 年代，我在柏林 成長時，我就愛上了樂高，

and I still love them.

而現在我依然很愛。

And these days, I build LEGO with my kids on Saturday afternoons.

最近，我每週六下午會和 我的孩子一起玩樂高。

And the reason that my love for LEGO

讓我對樂高的愛一直不減的原因，

has remained evergreen is, quite simply,

其實很簡單，

that LEGO is a system for realizing creativity

樂高是一種可以實現創意的系統，

with almost no barrier to entry.

幾乎沒有入門門檻。

And I’m not only a LEGO dad,

我不僅是個樂高爸爸，

I'm also the CEO of GitHub.

我還是 GitHub 的執行長。

And if you don't know GitHub,

如果你不知道 GitHub，

you can think of it as the home of coding.

你可以把它想成是程式設計的家園。

It's where all the software developers,

這裡是所有軟體開發者，

the chief nerds of our society,

也就是我們社會的 頂尖書呆子們，一起合作的地方。

collaborate together.

And it's part of our mission to make it as easy as possible

我們的使命之一

就是讓每一位開發者能夠 輕鬆地用程式碼構建大小創意。

for every developer to build small and big ideas with code.

But in contrast to LEGO,

但與樂高不同的是，

the process of building software feels daunting to most people.

對大多數人來說， 編寫軟體的過程令人望之生畏。

This all started to change

這一切在 2022 年末 ChatGPT 出現時開始改變。

when ChatGPT came along in late 2022.

Now we live in a world where intelligent machines understand us

現在我們生活在一個智能機器 能像我們理解它們一樣

as much as we understand them.

理解我們的世界。

All because of language.

這一切都是因為語言，

And this will forever change the way we create software.

這將永遠改變我們創建軟體的方式。

Up until now, in order to create software,

到目前為止，要創建軟體，

you had to be a professional software developer.

你得是專業的軟體開發者，

You had to understand, speak and interpret the highly complex,

你必須理解、使用和詮釋 我們稱之為程式碼的這種高度複雜，

sometimes nonsensical language of a machine that we call code.

有時候甚至是荒謬的機器語言。

Modern code still looks like hieroglyphics to most people.

現代的程式碼對大多數人來說 仍然像象形文字一樣難懂。

Here's an example.

舉個例子。

This, from the early 1940s,

這是來自 1940 年代早期的

is the world's first computer programming language,

世界上第一種電腦程式語言，

called Plankalkül.

叫做 Plankalkül。

It set the foundation for the modern code that we use today.

它為我們現今使用的 現代程式碼奠定了基礎。

And as you can see, it's a few numbers,

正如你所見，這裡有一些數字，

some bubbles and some big-ass brackets.

一些氣泡和一些大括號。

Not much humanity here, right?

沒什麼人性的氣息，對吧？

Flash forward about 20 years

再快轉約 20 年，

to the programming language called COBOL.

到了叫做 COBOL 的 程式語言時期。

COBOL was invented during the Eisenhower years,

COBOL 是 在艾森豪總統時期發明的，

but it remains an important language

但它仍然是許多 大型金融機構的重要語言。

for many of our largest financial institutions.

Wall Street, your savings account, your credit cards,

華爾街、你的存款帳戶、信用卡，

all run on this today.

現在都還在使用它。

And we see some familiar words here.

這裡我們看到了些熟悉的詞語。

But structurally, I think this doesn't make much sense to most of you.

但結構上，我想這對大多數人來說 並沒有什麼意義。

Flash forward another 30 years to 1991,

再快轉 30 年，到 1991 年，

and we saw the birth of Python,

我們看到了 Python 誕生，

one of the most popular programming languages in this era of AI.

這是 AI 時代 最受歡迎的程式語言之一。

In 80 years, we went from bubbles to brackets,

在這 80 年的時間裡， 我們從氣泡到括號，

to blips of English,

再到一些英語詞彙，

and yet, we got nowhere near as close as the intuitiveness of human language.

但我們離人類語言的 直觀性依然還很遠。

But then came June 2020,

然後來到 2020 年 6 月，

and we got early access to OpenAI's large language model,

我們獲得了 OpenAI 的大型語言模型 （當時叫做GPT-3）的早期使用權。

then called GPT-3.

It was COVID, we were all on lockdown,

那時正值新冠疫情封鎖期間，

I remember we were on a video call together.

我記得我們都在視訊通話。

We fed random programming exercises into this raw model,

我們把隨機的程式設計練習題 餵給這個原始模型，

and like magic,

像變魔術一樣，

it solved 93 percent of them during the first few takes.

它在前幾次嘗試 就解決了 93％ 的問題。

We at GitHub recognized we had something remarkable in our hands,

在 GitHub，我們意識到 我們手中掌握著非凡的東西。

and we quickly turned around a novel developer tool

並迅速推出了一款創新開發工具：

called GitHub Copilot:

名為 GitHub Copilot，

an AI assistant that predicts and completes code

一個為軟體開發者 預測和完成程式碼的 AI 助手。

for software developers.

Copilot is now the most adopted AI developer tool on the planet.

Copilot 現在是全球 採用最多的 AI 開發工具。

The age of programming has been reborn.

程式設計的時代得以重生。

But the possibilities of the breakthrough

但這一突破的可能性

went further than just these business results.

遠遠超出了這些商業成果。

Because the large language models that power ChatGPT and Copilot

因為支持 ChatGPT 和 Copilot 的大型語言模型

are trained on a vast library of human information,

訓練於龐大的人類資料庫，

they understand and interpret nearly every human language,

它們幾乎能理解和解釋

every major human language.

所有主要人類語言。

They seem to get us.

它們似乎理解我們。

We have struck a new fusion

我們在人類語言和機器語言之間

between the language of a human and a machine.

達成了新的融合。

With Copilot, any person can now build software in any human language

有了 Copilot，每個人現在 都可以用任何一種人類語言

with a single written prompt.

通過一些簡單的文字提示來構建軟體。

Goodbye to the bubbles and the big-ass bracket.

告別氣泡和大括號。

This is the most profound breakthrough to technology

這是自軟體開發誕生以來

since the genesis of software development itself.

技術上的最深遠的突破。

Today, there are over 100 million developers on GitHub.

今天，GitHub 上已有 超過一億名開發者。

That's about one percent of the world's population,

那大約是全球人口的百分之一左右。

you know, plus-minus.

I think that number is about to explode.

我認為這個數字即將急劇增加。

And I want to show you why, here on my MacBook.

我想在我的 MacBook 上 向你們展示為什麼。

We started it all with the original Copilot or how we say the OG Copilot,

我們最初使用原版 Copilot， 或稱之為 OG Copilot，

and it literally just predicted and completed code in the editor.

它實際上只是在編輯器中 預測和完成程式碼。

You can think of the editor as, you know, the Google Docs for developers.

你可以把編輯器想像成 開發者的 Google Docs。

And when you have a doc open, you know how it is, empty page,

當你打開一個文檔時， 你知道那種感覺，空白頁面，

what do I actually want to do?

我到底想做什麼？

And I mentioned LEGO.

我剛提到樂高。

So let’s build a 3D LEGO brick on a web page.

那麼，我們來在網頁上 構建一個 3D 樂高磚吧。

So what developers do, you know, they start typing.

開發者們會怎麼做呢？他們開始打字。

And so I typed in the JavaScript file,

我在 JavaScript 文件中輸入：

create a function to create a LEGO brick.

創建一個函數來創建一個樂高磚。

And you can see here this gray text, we call this ghost text.

你可以看到這裡的灰色文字， 我們稱之為幽靈文本。

This is coming from the large language model.

這來自大型語言模型。

So now I can just press the tab key and press enter.

現在我只需按下 Tab 鍵， 再按 Enter 鍵。

And I get another suggestion, you know, to create a LEGO tower.

我得到另一個建議，創建一個樂高塔。

也許我們稍後再做這個。

Maybe we do that later.

或者我可以這樣做：函數 draw LEGO brick。

Or I can just do: function draw LEGO brick.

And here again you see ghost text from Copilot right away available for me.

這裡你馬上又看到 Copilot 提供可用的幽靈文本。

And if I like what I'm seeing here,

如果我喜歡這裡看到的內容，

so I get into a mode of writing and understanding,

我就進入編寫和理解的模式，

I can just accept this.

可以直接接受這個建議。

Developers love that, right?

開發者們喜歡這一點，對吧？

Because instead of writing ten lines of code themselves

因為他們不用自己寫十行程式碼

or copy and pasting them from the internet,

或從網路上複製、貼上，

they get them right in their editor.

而是在編輯器中直接獲得這些內容。

They can stay in the flow.

他們可以保持在工作流程中。

Now what the OG Copilot didn’t offer me is a way to interact with this.

但是 OG Copilot 沒有 提供我與它互動的方式。

I cannot ask questions,

我不能問問題，

I cannot, you know, instruct it to do different things.

我不能指示它做不同的事情。

Last year we launched a new feature, Copilot chat,

去年我們推出了一個新功能， Copilot Chat，

and you can think about it as ChatGPT in your editor.

你可以把它想像成 在你的編輯器中的 ChatGPT。

So I can open this up here in the sidebar.

我可以在側邊欄打開這個功能。

And now I can tell it to create a whole web page

現在我可以叫它為我創建 一整個 3D 樂高磚的網頁。

with a 3D LEGO brick for me.

Now you know, similar to ChatGPT, it streams the response,

像 ChatGPT 一樣， 它會實時顯示回應，

and it gives me not only some code

不僅給我一些程式碼，

but it actually gives me an explanation.

居然還給我解釋。

You know, it starts writing code,

它開始編寫程式碼，

you can see the comments that explain what that code does.

你可以看到註釋 解釋這些程式碼的作用。

It uses an open-source library called Three.js.

它使用了一個叫做 Three.js 的開源庫。

And so you can kind of see here the idea of this empowering developers

你可以理解，這個概念是給予開發者

and people that want to learn development.

和想要學習程式的人自主能力。

And it ends, you know, with another explanation.

它以另一個解釋結束。

Now I can go here, inspect that code,

現在我可以檢查這些程式碼，

and I can actually push that button to copy it into my file.

並實際將它複製到我的文件中。

But I want to show you something else here.

但我想向你們展示另一件事。

你可能已經看到這個小話筒圖標。

And you might have already seen this little mic icon.

我可以用它與 Copilot 對話。

I can use that to speak to Copilot.

And I want to ask it, in German, what that code does

我想用德語問它，

這段編輯器左側的程式碼是什麼作用。

that is on the left side in the editor.

(Speaking German) Can you explain to me what that code does?

（講德語）這段程式碼是做什麼的？

And now Copilot responds again,

現在 Copilot 再次回應，

but it responds in German to me, right?

但它用德語回應我。

So it says, if I loosely translate,

它說，容我大致翻譯，

"Yes, of course, this JavaScript code defines a function

「當然，這段 JavaScript 程式碼定義了一個名為

named ‘drawLEGOBrick.’”

drawLEGOBrick 的函數。」

So you get the idea here.

你明白了吧？

A six-year-old in Berlin, in Mumbai and Rio,

一個六歲孩子，不論 在柏林、孟買或里約，

can now explore coding without their parents being around

即便父母沒有在身邊，

或沒有技術背景的情況下， 現在他們也可以探索程式設計。

or even having a technical background.

(Laughter)

（笑聲）

I mean, you know.

(Applause)

（掌聲）

Now what you also see is you still need to kind of figure out

現在你還會看到，

你仍然需要弄清楚如何 將這些都整合起來，對吧？

how you put that all together, right?

There’s a lot of technical stuff here.

這裡有很多技術細節。

I have code. I have to iterate on my machine.

我有程式碼，必須 在我的機器上反覆迭代。

I have to figure out how to deploy this to the cloud

我必須弄清楚如何將這些部署到雲端，

so I can share with my friends.

以便與我的朋友分享。

But here is my LEGO brick now.

現在，你看到的是我的樂高磚。

This is what it looks like

如果我做了開發者的所有步驟後， 它就會長這個樣子，

if I've done all these steps as a developer,

you can see now it’s a nicely rotating brick.

你現在可以看到 這個磚很漂亮，會旋轉。

我可以用滑鼠讓它旋轉。

I can actually use my mouse to turn it around.

These are the anti-studs here, the studs,

你會看到凹孔，還有凸點。

There's nice lighting effects.

光影效果很好。

I can even zoom into this and zoom out of this.

我還可以放大、縮小。

Now I don't want to do all this developer stuff anymore.

現在我不想再做這些 程式設計師的事情了。

I just want to channel my creativity straight into reality.

我只想直接將我的創意轉化為現實。

And so for the first time ever on stage,

我將首次在舞台上展示

I'm going to show you a new product that we call Copilot Workspace

我們的新產品，稱之為 Copilot Workspace,

that does exactly that.

它的功能正是如此。

So here is my workspace.

這是我的工作區。

And you can already see there's not an editor anymore.

你可以看到這裡已經不再是編輯器了。

I can just see a task, and I can enter a task.

我只看到一個任務欄， 然後我輸入一個任務。

And so now I have my LEGO brick,

我現在有樂高磚，

I want to now expand the LEGO brick into a LEGO house.

我想把這個樂高磚 擴展成一個樂高房子。

堆疊磚塊，形成一個樂高房子的形狀。

Stack the bricks in the shape of a LEGO house.

And I can do that also in German and in other languages.

我也可以用德語和其他語言來進行。

But for now, let's stick with English.

但現在，我們還是用英語吧。

I can save my task.

我可以將任務儲存。

And now what happens is that Copilot Workspace analyzes what I already have

Copilot Workspace 會分析我已有的內容，

and then describes what it proposes to me.

然後向我描述它建議的計劃。

Basically, it reframes my ask into a plan or a specification.

基本上，它會將我的需求 重新構建成一個計劃或規範。

And so you can see here, you know, it's all in natural language in our user.

你可以看到這些都是 使用者的自然語言。

Some file names, of course, but there is no code here.

當然有些文件名，但沒有程式碼。

It's all describing it in English.

它全部用英語描述。

I can actually go into this and edit it

我可以進入並編輯它，

and can make changes to this line,

並改動某一行，

or I can go down here and add another item

或者如果我覺得 計劃不完全符合我的要求，

if I feel like the plan is not exactly what I want.

我可以在這裡再添加一個項目。

I can go a step further and generate a plan,

我可以進一步生成一個計劃，

and now an agent runs through all my files I already have

接著一個程式工具會遍歷 我已有的所有文件，

and figures out how do I need to modify those files,

並找出我要如何修改這些文件，

or, you know, do I need to add files to my repository

或者說，我需要在程式庫 添加哪些文件，

so you know it wants to add a “createLEGOHouse” function

才能讓它添加一個函數 “createLEGOHouse” ，

and call the “createLEGOHouse” afterwards.

並在之後必要時調用該功能。

Looks good to me, so let's implement this.

這看起來不錯， 我們來跑一下這個程式。

And now Copilot uses my task, my specification,

現在 Copilot 使用 我的任務、規範、計劃，

my plan to write code for me.

為我編寫程式碼。

You can see here two files are queued,

你可以看到兩個文件 正在排隊等待處理，

the public/legoBrick.js file

public/legoBrick.js 文件，

and boom, there's already my code written for me, right?

一下子，程式碼已經寫好了，對吧？

I didn't have to touch code,

我不需要碰程式碼，

I didn't have to even know what code is.

我甚至不需要知道程式碼是什麼。

現在我看到它導入了 一些新內容到我的文件中，

Now I see here now it imports some new line into my file

and has written, you know, lots of code here that does those changes.

並寫了很多程式碼來進行這些更改。

So you want to see what that looks like, did we get a LEGO house?

你想看看它的樣子， 我們做好樂高房子了嗎？

So here's a button that lets me open a live preview,

這裡有一個按鈕讓我打開實時預覽，

so I can do this.

我可以這樣做。

And now the bricks fall from the sky and I have a LEGO house.

現在磚塊從天而降， 樂高房子蓋好了。

And you know, this is not a picture, right --

這不是圖片而已，對吧

(Applause)

（掌聲）

Yes, thank you.

是的，謝謝。

This is all live, this is the power of code,

以上都是實時的，這是程式碼的力量，

this is the power of streaming my creativity into reality

這是用自然語言將我的創意 轉化為現實的力量。

with natural language.

Now one last thing.

最後一件事。

Thank you, Copilot,

謝謝你，Copilot，

you have always to be nice to the AI.

永遠要對 AI 友善。

(Laughter)

（笑聲）

(Applause)

（掌聲）

Now, what you just saw were three leaps in three years.

你剛剛看到的是 這三年內的三次大躍進。

Three leaps that are more progress

這三次躍進

to the accessibility of computer programming

比我們在過去 100 年中 達成的程式設計可及性的進步更大。

than we have made in the last 100.

Remember how I said that one percent of the world's population is a developer?

記得我說過，世界人口的 百分之一是開發者嗎？

Now you can see how this will change.

現在你可以看到這將如何改變。

Copilot Workspace may still be a developer tool right now,

Copilot Workspace 現在可能還是一個開發者工具，

but soon enough these kind of developer tools will become mainstream.

但很快，類似的開發工具將成為主流。

Because, going forward, every person, no matter what language they speak,

因為，展望未來，每個人， 不管他們講什麼語言，

will also have the power to speak machine.

都將有能力跟機器說話。

Any human language is now the only skill

現在你只需要會任何一種人類語言，

that you need to start computer programming.

你就可以開始編寫程式碼。

This will lead to a globalized groundswell of software developers,

這將導致全球 軟體開發者大規模增長，

and it will reshape the geography of our global economy.

並將重塑全球經濟的地理格局。

And because of this,

正因如此，

I think by 2030, maybe even sooner,

我認為到 2030 年，甚至更早，

we will have more than one billion software developers on GitHub.

GitHub 上的 軟體開發者將超過十億人。

Think about that:

想想看：

10 percent of the world’s population will not only control a computer

世界人口的 10％ 將不僅能夠控制電腦，

but will also be able to create software

還能夠像騎自行車一樣創建軟體。

just [as] if they were riding a bicycle.

This will generate a new renaissance of human creativity with software.

這將帶來人類用軟體發揮創意的 新文藝復興時期。

Now anyone here in this room could have a brilliant idea right now:

在座的各位都可能有個絕妙的創意：

a website, an application,

一個網站，一個應用程式，

a cool computer game, an amazing song,

一個酷炫的電腦遊戲， 一首驚人的歌曲，

maybe even a cure for something.

甚至可能是某種疾病的治療方法。

For example, last year, over a couple of weeks,

例如，去年，我花了幾周的時間

I built an app that tracks all the flights I've ever taken in my life.

建立了一個應用程序， 追蹤我一生中乘坐過的所有航班。

Now I know what you're thinking.

我知道你在想什麼。

What a freaking nerd, right?

真是個怪咖，對吧？

And yeah, it's true, I love building stuff like this.

是的，這是真的， 我喜歡建造這樣的東西。

And with the help of AI,

有了 AI 的幫助，

now I can do this in English or in German

現在我可以用英語或德語

before I even finish a glass of wine.

在喝完一杯酒之前完成它。

And soon enough, this will be true for everyone here.

很快地，在座的各位也能做到。

The floodgates of nerditude have swung wide open.

書呆子的天地已經大開！

(Laughter)

（笑聲）

(Applause)

（掌聲）

Now this doesn’t mean

這並不意味著

that everyone will become a professional software developer

每個人都會成為專業的軟體開發者，

or even that they should.

甚至不應該這樣。

The profession of a professional software developer

專業軟體開發者的工作

is not going anywhere.

不會消失。

There will always be demand for those that design and maintain

我們仍然需要

設計和維護世界上大型軟體系統的人。

the largest software systems in the world.

We are adding millions of lines of code every single day

我們每天都在越來越複雜的系統裏 添加數百萬行程式碼，

to ever more complex systems,

and we are barely keeping up with maintaining the existing ones.

而我們幾乎跟不上 維護現有系統的步伐。

Like any infrastructure in this world out there,

就像這個世界上的任何基礎設施一樣，

we need real experts to preserve and renew it.

我們需要真正的專家來維護和更新它。

The point here is not a "will" or a "should."

這裡的重點不是「會」或「應該」。

It's that anyone can.

是任何人都可以。

All because the most powerful system that we have,

這一切都是因為 我們擁有最強大的系統—

any human language,

任何一種人類語言，

is now fused to the language of a machine.

現在可以與機器語言融合。

And very soon,

很快，

building software will be just as simple and joyful

構建軟體將像搭建樂高一樣 簡單、愉快。

as stacking a LEGO.

(Speaking German) Thank you very much.

（講德語）非常感謝大家。

(Applause)

（掌聲）

Bilawal Sidhu: Gosh, I've got to say, one billion developers

Bilawal Sidhu： 天哪，我得說，

一億開發者讓 GitHub 聽起來 更像 YouTube 和 TikTok，

makes GitHub sound more like YouTube and TikTok than it is today.

Just super exciting.

這太令人興奮了。

Got to ask you one question,

得問你一個問題，

perhaps the elephant in the room.

也許是眾所周知的問題。

Amazing talk.

精彩的演講。

So you said the developer is still in charge.

你說開發者仍然掌控局面。

You also said, "We've had three leaps in three years."

你也說過，「我們這三年內 有三次大躍進。」

Fast forwarding a little bit,

展望不遠的未來，

do you think humans will still need to be in the loop,

你認為人類還需要參與其中嗎？

or will these AI systems be able to autonomously build

還是這些 AI 系統將能夠自主構建

and maintain software?

和維護軟體？

TD: You know, the way I always think and talk about it

TD：我總是這樣想和談論它，

is that we called it Copilot for a reason.

我們稱它為 Copilot （副駕駛員）是有原因的。

We need a pilot.

我們需要一位主駕駛員。

We need a pilot that is creative, that can decide what to do.

我們需要有創意的駕駛員， 能夠決定做什麼事。

It’s kind of like a LEGO set.

這有點像樂高套件。

You need to take this big problem and break it down into smaller problems,

你需要把大問題分解成更小的問題，

into small building blocks.

變成小的組件。

And for that, you need a systems thinker.

為此，你需要一個系統思考者。

You need a human that can figure out, am I building a point of sale system?

你需要某個人，能夠弄清楚 我是在構建一個銷售點系統嗎？

Am I building an iPhone app?

我是在構建一個 iPhone 應用程式嗎？

Am I building a cool computer game?

我是在構建一個酷炫的電腦遊戲嗎？

Am I building the next Facebook?

我是在構建下一個 Facebook 嗎？

Those are very different systems.

這些是非常不同的系統。

這些組件將會越來越大。

Now these building blocks, they will grow in size.

Today it's, you know, a couple of lines of code,

今天可能是幾行程式碼，

maybe a whole file,

也許是一整個文件，

in the future, it might be a whole subsystem.

將來可能是一個完整的子系統。

So I get more work taken off my shoulders.

這樣我可以減輕更多的工作負擔。

But I'm still there, you know, covering the large system.

但我還掌控著整個系統。

And as I mentioned, you know, we're still running COBOL systems from the '60s.

正如我所提到的，我們仍在運用來自 60 年代的 COBOL 系統，

So we have lots of work to do.

所以我們有很多工作要做。

BS: Absolutely.

BS：絕對如此。

So we will be in charge orchestrating these systems

我們將在更高層次的 抽象層面上操控這些系統。

at a higher level of abstraction.

Thomas Dohmke, everybody, thank you.

請大家掌聲謝謝湯瑪斯·杜麥克。

TD: Thank you so much.

TD：非常感謝。