Name: 不斷佔據台新聞版面的 NVIDIA CEO 黃仁勳說不要學 code，而要學「這個」！("Don't Learn to Code, But Study This Instead..." says NVIDIA CEO Jensen Huang)
Uploaded: 2024-06-10T21:01:30.000Z
Duration: 11 min 35 s

情態副詞

Jensen Huang, co-founder and CEO of NVIDIA, the company that is at the heart of AI revolution, recently made a comment about what kids should learn in the future.

處於 AI 革命核心地位的輝達公司聯合創始人兼 CEO 黃仁勳最近就孩子們未來應該學習什麼發表了評論。

And let me spoil it to you, it's not computer science or programming.

讓我來告訴你，不是電腦科學或寫程式。

程度副詞

I'm going to say something and it's going to sound completely opposite of what people feel.

我要說的話，聽起來會與人們的感受完全相反。

You probably recall over [that] the course of the last 10 years, 15 years, almost everybody who sits on a stage like this would tell you it is vital that your children learn computer science.

你可能還記得，在過去的 10 年、15 年裡，幾乎所有坐在臺上的人都會告訴你，讓孩子們學習電腦科學至關重要。

And in fact, it's almost exactly the opposite.

而事實上，情況幾乎正好相反。

It is our job to create computing technology such that nobody has to program.

我們的工作就是創造電腦技術，讓所有人都不需要寫程式。

And that the programming language, it's human.

At Zymph Minds, a full service agency, where I'm a co-founder, we're in a close relationship with NVIDIA.

Zymph Minds 是一家提供全方位服務的公司，我是該公司的聯合創始人之一，我們與輝達公司保持著密切的合作關係。

And later in this video, I will include a clip of what Jensen said people should study,

在這部影片的稍後部分，我還會附上黃仁勳說人們應該學習的內容的片段，

which is also in alignment why we are expanding our team with people like Amy Zolotov, who comes from healthcare.

這也是我們擴大團隊的原因所在，比如 Amy Zolotov 就來自醫療保健行業。

Collectively, we trained and taught close to 20,000 people at this point, from people at all ages and walks of life who subscribe to my channel, to the amazing teams at the companies like General Electronics and Kraft Heinz.

目前，我們總共培訓和教導了近兩萬人，其中既有訂閱我頻道的各年齡段、各行各業的人，也有通用電子和卡夫亨氏等公司的優秀團隊。

Plus, we've been developing AI educational material, both for universities like Stanford or Yeshiva, all the way to developing courses on Uplimit.

此外，我們還一直在開發 AI 教材，包括為史丹佛大學或耶希瓦大學開發的教材，以及在 Uplimit 上開發的課程。

I think this shows that there is raising awareness about prompt engineering and the impact that you can have in your organization or your team if you just know better how to work with AI tools.

我認為，這表明人們對指令工程的認識在不斷提高，也表明如果你能更好地瞭解如何使用 AI 工具，就能對你的組織或團隊產生影響。

And I understand that when a girl on the internet says that prompting is the new programming language, that's one thing.

我明白，當一個女孩在網上說指令是新的程式語言時，那是一回事。

But I'm really happy to see that more people kind of receive a message when it is NVIDIA CEO on a stage at the World Government Summit, saying it out loud.

但我真的很高興看到更多人了解 NVIDIA 執行長在世界政府高峰會的舞台上大聲說出的訊息。

Everybody in the world is now a programmer.

現在世界上每個人都會寫程式。

This is the miracle of artificial intelligence.

For the very first time, we have closed the gap.

The technology divide has been completely closed.

And this is the reason why so many people can engage artificial intelligence.

這就是為什麼這麼多人能夠參與 AI 的原因。

It's the reason why every single government, every single industrial conference, every single company is talking about artificial intelligence today.

這也是今天每個政府、每個工業會議、每個公司都在談論 AI 的原因。

Because for the very first time, you can imagine everybody in your company being a technologist.

因為這是第一次，你可以想象公司裡的每個人都是科技專家。

This is a tremendous time for all of you to realize that the technology divide has been closed.

這是一個偉大的時刻，你們都應該意識到，科技鴻溝已經消除。

I hope more people and businesses will take it as a wake-up call and start looking at this quite seriously.

我希望更多的人和企業能以此為警鐘，開始認真對待這個問題。

Also to realize that to engage AI is a lot easier now than at any time in the history of computing.

也要意識到，現在使用 AI 比電腦歷史上的任何時候都容易得多。

As for 90% of people, prompting is going to just become easier and easier.

因為對於 90% 的人來說，下指令會變得越來越容易。

It's because of smart AI engineers and prompt engineers working in the background, optimizing that chat interface.

這是因為聰明的 AI 工程師和指令工程師在後臺工作，優化了聊天界面。

So that a few vague words can be taken, processed by a way bigger and more extensive prompt or the whole separate AI model, by the way.

這樣，幾個含糊不清的詞就能被一個更大、更廣泛的指令或整個獨立的 AI 模型所接收、處理。

All that optimization gets combined into the final prompt and sent to the AI model.

所有優化結果都會合併成最終指令，並發送給 AI 模型。

Therefore, you get these amazing outcomes like Delhi Tree or Copilot.

所以你會得到像 Delhi Tree 或 Copilot 這樣的驚人成果。

It starts with a prompt from you in an app.

它始於你在應用程式中的指令。

Copilot pre-processes the prompt through an approach called grounding.

Copilot 通過一種稱為「接地」的方法對指令進行預先處理。

Put simply, grounding improves the quality of the prompt.

簡單地說，接地可以提高指令的品質。

So you get answers that are relevant and actionable.

這樣，你就能得到相關且可行的答案。

One of the most important parts of grounding is making a call to the Microsoft Graph to retrieve your business content and context.

接地最重要的部分之一就是調用 Microsoft Graph 來檢索你的業務內容和上下文。

Copilot combines this user data from the graph with other inputs to improve the prompt.

Copilot 將圖表中的用戶數據與其他輸入數據相結合，以改進指令。

It then sends that modified prompt to the LLM.

然後，它會將修改後的指令發送給 LLM。

Copilot takes the response from the LLM and post-processes it.

Copilot 接收來自 LLM 的回應，並對其進行後處理。

This post-processing includes additional grounding calls to the graph, responsible AI checks, security, compliance and privacy reviews, and command generation.

這種後處理包括對圖形的額外基礎調用、負責任的 AI 檢查、安全性、合規性和隱私審查以及命令生成。

Finally, Copilot sends a response to the user and commands back to the apps.

最後，Copilot 會向用戶發送回覆，並向應用程式發送指令。

It's awesome if you're not specific and you are just really exploring.

如果你沒有特別的要求，只是想真正探索一下，那就太棒了。

Because sometimes these chatbots get so optimized that it gets horribly hilarious.

因為有時這些聊天機器人優化得太厲害了，會變得非常搞笑。

Gemini from Google went viral for absolutely wrong reasons, refusing to generate white people and over-optimizing for inclusivity and I could say stereotypes, but kind of phenomenal is that it became stereotypical.

谷歌的 Gemini 因絕對錯誤的原因而走紅，它拒絕生成白人，過度優化包容性，我可以說是刻板印象，但某種現象是，它變成了刻板印象。

I'm genuinely curious what impact it will have that NVIDIA CEO had this hot take.

我真的很好奇，輝達 CEO 的這一熱門話題會產生什麼影響。

But if you listen between the lines, even Sam Altman hinted at this future.

但如果你仔細聆聽，甚至 Sam Altman 也暗示了這一未來。

I think coding will still be important in the future.

我認為，寫程式在未來仍然很重要。

It's just going to change a little bit or a lot.

只是會有一點點或很大的變化。

One of the predictions I'm most confident about the future is it's just going to happen much faster than the past did.

我對未來最有信心的預測之一就是，未來的發展速度會比過去快得多。

And that, you know, young people are always the best at adapting to that, but now it's going to be extra valuable.

年輕人總是最善於適應這種變化，但現在這種變化將變得格外寶貴。

I think I agree with that framework incredibly much.

And I think anytime you can choose a skill or a class or an activity to do that you feel aligns with the learning how to adapt to new technology, pick it for sure.

我認為，無論何時，只要你能選擇一種技能、一門課程或一項活動，讓你覺得它與學習如何適應新技術相一致，就一定要選擇它。

But on specifics, I still think that software is actually a really important skill to build.

但具體來說，我仍然認為軟體實際上是一項非常重要的技能。

And some of it is it's not going to be about the mechanics, it's not going to be about the specifics of the programming language, it's not going to be about those details, but fundamentally getting computers to do what you intend.

其中的一些原因是，它與機械原理無關，與寫程式語言的具體細節無關，而是從根本上讓電腦按照你的意圖工作。

And so understanding the low-level details of how it works, that's a starting point.

瞭解其工作原理的底層細節是一個起點。

And you know, we do see people who have never coded before who build some of the best AI applications.

我們確實看到一些從未寫過程式的人開發出了一些最優秀的 AI 應用程式。

And so that's actually new, that you can actually get computers to do things without having to do it in this like contort yourself to the machine.

這實際上是一項新技術，你可以讓電腦做一些事情，而不用像這樣把自己扭曲成機器的樣子。

As it is right now, anyone can work with AI, anybody can instruct AI.

現在，任何人都可以使用 AI ，任何人都可以指導 AI。

Can you imagine that a couple years ago, this ability to interact with AI models was available only to researchers and actual developers?

你能想象幾年前，這種與 AI 模型互動的能力只有研究人員和實際開發人員才能使用嗎？

And just because it's available to everybody and it's getting easier, doesn't mean that people use it in the best way possible or the most effectively.

雖然人人都能使用，而且越來越容易，但這並不意味著人們會以最好的方式或最有效的方式使用它。

The way I like to think about this stuff is like a spectrum.

我喜歡用光譜的方式來思考這個問題。

On one side, you have like, oh my God, prompt engineering is the only thing you need to know.

一邊是「哦，天哪，指令工程是你唯一需要知道的東西」，

On the other spectrum is like, oh, it doesn't matter.

另一種觀點則認為「哦，沒關係。

Programming is going to be still there, and it's a solid skill.

程式設計將繼續存在。這是一項紮實的技能。」

And in our team, the way we like to look at things is a bit more leveled.

而在我們的團隊中，我們看待問題的方式更有層次感。

If you think about in a term of 10 years, if we all walk with the neural link chips in our head.

如果你思考一下 10 年後，如果我們都帶著大腦中的神經連接晶片行走。

This morning, Elon Musk announcing that for the first time, his company Neuralink has implanted a brain chip in a human.

今天上午，伊隆馬斯克宣佈，他的公司 Neuralink 首次為人類植入了大腦晶片。

And after the statement from Jensen, on the whole other spectrum, there's going to be a group of people who are like, "Okay, programmers, they are not valued, they are not important.

在黃仁勳的聲明之後，在另一個層面上，會有一群人認為：「好喔，寫程式的他們不被重視，他們不重要。

Don't study computer science and stuff like that."

不要學習電腦科學之類的東西。」

Because programming and computer science as hands on skill is one thing, but what it teaches people is problem solving.

因為寫程式和電腦科學作為一門動手技能是一回事，但它教給人們的是解決問題的方法。

And that is a hard skill, which won't get absolute anytime soon.

而這是一項艱苦的技能，不會在短期內得到絕對化。

And I want to say never, but we also said that creativity is probably going to be the last affected by AI, which turned out completely opposite.

我想說絕對不會，但我們曾說過創造力可能是最後一個受 AI 影響的，結果卻完全相反。

So anyway, taking a part being in the middle, educate yourself about prompt engineering; how to use AI, how it works, what are ethical uses of that.

所以無論如何，我們都要在中間扮演一個角色，對自己進行指令工程方面的教育，瞭解如何使用 AI 、AI 的運作原理以及 AI 的道德用途。

But if you're able to also learn about computer science and how computers work in general, develop deeper problem solving and critical thinking, really skills with programming, that is not going to hurt you.

但是，如果你能夠同時學習電腦科學和電腦的一般運作原理，培養更深層次的問題解決能力和批判性思維，以及真正的寫程式技能，這對你來說並沒有壞處。

Like I myself, I talk about prompt engineering, but I also decided to take CS50 course to kind of bridge the gap I have when it comes down to programming.

就像我自己一樣，我談論的是指令工程，但我也決定學習 CS50 課程，以彌補我在寫程式方面的不足。

Funny thing that a month ago, before Jensen's video came out, I made a video where I said that domain expertise is going to be extremely important when using AI tools.

有趣的是，一個月前，在黃仁勳的影片發佈之前，我製作了一個影片，我在影片中說，在使用 AI 工具時，領域專業知識將變得極其重要。

And here's one universal truth when it comes to prompting.

說到下指令，有一個普遍的真理。

You as a user have to have domain expertise and use specific language to condition the AI model and steer its responses to the right direction.

作為用戶，你必須具備相關領域的專業知識，並使用特定的語言來調節 AI 模型，並引導其反應向正確的方向發展。

Because if you know your stuff very well, not only you will be able to explain it well and make clear instructions, but you will also be able to detect hallucinations.

因為如果你對自己的東西非常瞭解，不僅能夠很好地解釋，做出清晰的指令，還能發現幻覺。

The countries, the people that understand how to solve a domain problem in digital biology or in education of young people or in manufacturing or in farming,

那些了解如何解決數位生物學、年輕人教育、製造業或農業領域問題的國家和人民，

those people who understand domain expertise now can utilize technology that is readily available to you.

那些了解領域專業知識的人現在可以利用你隨時可用的技術。

You now have a computer that will do what you tell it to do, to help automate your work, to amplify your productivity, to make you more efficient.

現在，你有了一臺電腦，它可以按照你的要求去做，幫助你實現工作自動化，提高你的工作效率，讓你更有效率。

And so I think that this is just a tremendous time.

所以我認為現在正是一個大好時機。

And going back to that question, what kids should study in the future?

回到這個問題，孩子們將來應該學習什麼？

If I were starting all over again, I would realize one thing, that one of the most complex fields of science is the understanding of biology, human biology.

如果重新來過，我會意識到一件事，科學中最複雜的領域之一就是對生物學、人類生物學的理解。

Not only is it complicated because it's so diverse, so complicated, so hard to understand, living and breathing, it is also incredibly impactful.

它的複雜性不僅在於它的多樣性、複雜性、活生生的難以理解性，還在於它的影響力令人難以置信。

Complicated technology, complicated science, incredibly impactful.

複雜的技術，複雜的科學，令人難以置信的影響。

And remember, we call this field life sciences and we call drug discovery, discovery.

請記住，我們稱這一領域為生命科學，稱藥物發現為「發現」。

As if you wander around the universe and all of a sudden, "Hey, look what I discovered."

就好像你在宇宙中漫步，突然說：「嘿，看我發現了什麼。」

Nobody in computer science, nobody in computers and nobody in the traditional industries that are very large today, nobody says car discovery.

在電腦科學領域，在電腦行業，在當今非常龐大的傳統行業，沒有人說過「發現」汽車。