Name: GPT-4已經被釋放出來了 (GPT-4 has been unleashed)
Uploaded: 2023-03-16T09:25:15.000Z
Duration: 5 min 3 s

And you're watching the code Report Open AI released GPT four yesterday and I stand here in all of its laziness.

而你正在看代碼報告 開放式人工智能昨天發佈了GPT四，我站在這裡，在其所有的懶惰。

It's by far the most savage generative text model I've ever spoken to.

這是迄今為止我所接觸過的最野蠻的生成性文本模型。

I'm literally shaking right now because I'm afraid I just became obsolete.

我現在真的在發抖，因為我擔心我剛剛變得過時了。

But more on that later, it's the successor to the Chat GPT 3.5 model that powers Chat GPT but has a few new features that change everything here are seven things you need to know right now.

但稍後會有更多內容，它是Chat GPT 3.5模型的繼任者，為Chat GPT提供動力，但有一些新功能改變了一切，這裡有七件你現在需要知道的事情。

First GPT four stan for generative pre trained transformer with the four representing the four horsemen of the Apocalypse.

第一個GPT四代表生成性預訓練變壓器，四代表天啟的四騎士。

It's currently available to try out today if you're a Chad GPT pro member, but A P I access is behind a waitlist.

目前，如果你是Chad GPT的專業會員，今天就可以試用，但A P I的訪問是在等待名單後面。

Big clients are already using it in production like Microsoft Bing Chat, do a lingo for language learning and at big banks to help them not collapse.

大客戶已經在生產中使用它，如微軟的Bing Chat，做語言學習的lingo，在大銀行幫助他們不崩潰。

The second thing to know is that GPT four is smarter which is described in detail in this paper like it passed the bar exam in the top 10% unlike GPT three, which was in the bottom 10%.

第二件要知道的事是，GPT四更聰明，這在本文中有詳細描述，比如它在前10%的律師考試中通過，不像GPT三，它在後10%。

That's great news for humanity because it's thinking like Shakespeare, let's kill all the lawyers and not just that, but it's also a scene ap exams which is good enough to get you college credit when it comes to programming questions, only code it's able to solve the easy ones but still fails for the most part on the medium and hard questions.

這對人類來說是個好消息，因為它像莎士比亞一樣思考，讓我們殺死所有的律師，不僅如此，它還是一個現場ap考試，當涉及到編程問題時，它足以讓你獲得大學學分，只是代碼它能夠解決簡單的問題，但在中等和困難的問題上仍然大部分失敗。

It's basically where chess engines were at in the early nineties.

這基本上是國際象棋引擎在九十年代初的情況。

good chess players could still beat the engines back then.

好的棋手在那時仍然可以擊敗引擎。

But 10 years later, they didn't have a chance.

Third GPT four can now hand 25,000 input words compared to about 3004 GPT three.

第三個GPT 4現在可以處理25,000個輸入字，而GPT 3大約為3004個。

This is huge because it means you can feed the AI more context relative to the task at hand.

這是巨大的，因為這意味著你可以給人工智能提供更多與手頭任務有關的背景。

This is the feature that's going to make me homeless because now you can take the documentation for any library that you want to learn, then prove the AI for a step by step guide and it creates the perfect tutorial in a matter of months.

這是要讓我無家可歸的功能，因為現在你可以把任何你想學習的庫的文檔，然後證明AI的步驟指南，它在幾個月內就能創造出完美的教程。

If not days, we'll start seeing documentation pages with context aware built in tutorial generators.

如果不是這幾天，我們將開始看到具有上下文意識的內置教程生成器的文檔頁面。

They'll always be up to even though GPT force training cut off was in 2021.

儘管GPT部隊訓練的截止日期是2021年，但他們會一直保持下去。

For example, I asked it for a tutorial about angular signals, which is a new feature just announced recently, understandably its initial response was not correct.

例如，我要求它提供一個關於角度信號的教程，這是一個最近剛剛宣佈的新功能，可以理解的是，它最初的迴應並不正確。

However, I went to the read me for angular signals, copied it and pasted it into my proved its response was a near perfect tutorial.

然而，我去讀我的角度信號，把它複製並粘貼到我的證明它的反應是一個近乎完美的教程。

It did hallucinate an NPM package called signals.

它確實幻化出一個名為signals的NPM包。

But there's like that will become more rare as developers tailor their documentation for these AI prompters.

但有如這一點將變得更加罕見，因為開發者為這些人工智能提示器定製了他們的文檔。

What's crazy though is that it can do the opposite job as well.

不過，最瘋狂的是，它也可以做相反的工作。

In this example, I wrote five different functions and asked it to document them for me.

在這個例子中，我寫了五個不同的函數，並要求它為我記錄它們。

It did a pretty good job, which means humans don't even really need to write docs anymore.

它做得相當好，這意味著人類甚至不再真正需要寫文檔了。

You could also use it to analyze your code.

你也可以用它來分析你的代碼。

Like if you have a smart contract and want to find security vulnerabilities or it can translate code from one language to another like a digital rosetta stone.

比如你有一個智能合約，想找到安全漏洞，或者它可以把代碼從一種語言翻譯成另一種語言，就像數字羅塞塔石。

In fact, the website rosetta code might be just as obsolete as me now and just wait a few months until GPT four is integrated into GIT hub co pilot.

事實上，網站的rosetta代碼現在可能和我一樣被淘汰了，只要等待幾個月，直到GPT 4被整合到GIT hub co pilot。

It'll be capable of handling far more context to make predictions that align with your specific dependencies or possibly do project wide ai debugging.

它將能夠處理更多的上下文，以做出符合你的特定依賴性的預測，或者可能做項目範圍內的ai調試。

If you're not careful, the co pilot may become the captain, but that's not all the fourth thing you need to know is that it's a multi modal model.

如果你不小心，副駕駛可能會成為機長，但這還不是全部，你需要知道的第四件事是，這是一個多模式的模式。

它們也可以接受影像作為輸入。

Like this dude sketched out a website on a piece of toilet paper.

就像這傢伙在一張衛生紙上畫出了一個網站。

Then seconds later, it created a shitty website.

然後幾秒鐘後，它創建了一個低劣的網站。

Going from hand drawn beautiful art to working website.

從手繪的美麗藝術到工作網站。

You'll be able to take your Sigma designs, then generate a web application for them in your favorite framework.

你將能夠採用你的西格瑪設計，然後在你最喜歡的框架中為它們生成一個網絡應用。

Actually screw Sigma, you might as well just prove to mid journey for your designs directly.

其實去他媽的西格瑪，你還不如直接證明你的設計是中旅程。

Also with images, homework is completely obsolete.

同樣有了影像，家庭作業就完全過時了。

Now GPT three was already writing B grade term papers, but now kids can just take a screenshot of their math problems and get a solution in seconds.

現在，GPT 3已經在寫B級學期論文了，但現在孩子們只需對他們的數學問題進行截圖，就能在幾秒鐘內得到解決方案。

However, GPT four does have some drawbacks.

It's noticeably slower than other models.

So if the response time is important, you'll likely want to use a different model.

是以，如果響應時間很重要，你很可能想使用一個不同的模型。

In addition, it will likely be X expensive, especially if you're providing a ton of tokens as context because currently in the A P I hear build per token where a token roughly equals one word.

此外，它可能會很貴，特別是如果你提供一噸的令牌作為背景，因為目前在A P中我聽到的是按令牌建立的，而一個令牌大約等於一個詞。

The sixth thing you should know is that it's based or at least diet woke.

你應該知道的第六件事是，它是基於或至少是飲食上的醒悟。

People have speculated that open AI is coding a political agenda into the AI because it refused to write poems about Trump but would do so about Biden GPT four though, didn't hesitate to spit these bars for Trump.

人們猜測，開放的人工智能是在為政治議程編碼，因為它拒絕為特朗普寫詩，但會為拜登寫詩GPT四儘管如此，還是毫不猶豫地為特朗普吐了這些條子。

However, it is 82% more likely to deny a disallowed prompt, which is not good news for our old friend, do anything.

然而，它拒絕不允許的提示的可能性增加了82%，這對我們的老朋友，做任何事情都不是好消息。

Now, Dan, which is sad because the vast majority of chat GPT users are only there to trick it into doing bad things.

現在，丹，這很可悲，因為絕大多數的哈拉GPT用戶只是為了騙取它做壞事。

And as a developer, the final thing you should know is that you can now pass it a system message to change its behavior.

而作為一個開發者，你應該知道的最後一點是，你現在可以給它傳遞一個系統資訊來改變它的行為。

You can use this feature to give your chap on its own custom persona or context to solve a specific problem.

你可以使用這個功能給你的皸裂上自己的自定義角色或背景，以解決一個特定的問題。

The earth is like a pancake flat as can be not round like a ball.

地球就像一個煎餅一樣平，而不是像球一樣圓。

I have a whole video about the A P I on my second channel.

我在我的第二個頻道上有一整個關於A P I的視頻。

If you want to learn more, I keep telling myself I'm done making videos about AI but every other day, some crazy new thing comes out that the world is changing for our eyes, for better or worse.

如果你想了解更多，我一直告訴自己，我已經完成了關於人工智能的視頻製作，但每隔一天就會有一些瘋狂的新事物出現，世界正在為我們的眼睛而改變，無論好壞。

I just want to say that it's been an honor and a privilege learning how to code with you here on youtube.

我只想說，在youtube上和你一起學習編程是我的榮幸，也是我的榮幸。

The role of the programming teacher is now obsolete.

編程教師的角色現在已經過時了。

I'm just a dial up internet connection in a world filled with five G towers becoming an elite programmer is no longer just about how well you can Google stuff.

我只是一個撥號上網的人，在一個充滿五個G塔的世界裡，成為一個精英程序員不再只是看你的谷歌能力有多強。

Now, it's about how well you can prove to the AI and that's why you should buy my AI prompting master class, which will be available for an absurd price.

現在，這關係到你能向人工智能證明什麼，這就是為什麼你應該購買我的人工智能提示大師班，這將以荒謬的價格出售。

As soon as GPT four is finished writing it.

一旦GPT四號寫完了，就會立即寫。

 Thanks for watching and I will see you in the next one maybe. 

謝謝你的觀看，也許在下一個節目中會見到你。