It leaves Chat GPT in the dust.

它把Chat GPT甩在了後面。

It can convert a drawing on a napkin into a functional website.

它可以將餐巾紙上的圖畫轉化為一個功能性網站。

It can even explain to you a joke from a series of images.

它甚至可以從一系列的影像中向你解釋一個笑話。

So let's find out exactly what GPT four actually is open.

是以，讓我們來了解一下GPT四號究竟是什麼開。

A I ran a developer live stream that not only introduced this multimodal AI, they also ran a demo to demonstrate how it's one of the most powerful artificial intelligence engines to this day.

我辦了一個開發者直播，不僅介紹了這種多模態人工智能，他們還進行了演示，以證明它是迄今為止最強大的人工智能引擎之一。

It's no surprise that the whole internet Twitter especially has gone insane hearing this news Chat GPT was only released recently with being A I following in this video, I'm going to cover what GPT four actually is how it's different from previous GPT versions such as GPT three, GPT 3.5 and I've also gained access to test it out.

這並不奇怪，整個互聯網，尤其是Twitter聽到這個消息後都瘋了，GPT是最近才發佈的，在這個視頻中，我將介紹GPT 4到底是什麼，它與以前的GPT版本如GPT 3，GPT 3.5有什麼不同，我還獲得了測試它的權限。

So I'll demo some of its use cases to let's start off with GPT four.

所以我將演示它的一些使用情況，讓我們從GPT四開始。

In comparison to its previous models, it's multimodal.

與其以前的模型相比，它是多模態的。

This simply means that unlike previous versions of GPT, which were only text based GPT four can accept and process images as well as text, I wouldn't be surprised if there's an audio component to this as well.

這只是意味著與以前的GPT版本不同，以前的版本只是基於文本的GPT 4可以接受和處理影像以及文本，如果這其中也有音頻部分，我不會感到驚訝。

But for the time being open AI simply did a demonstration showcasing a few examples of their image to text processing.

但就目前而言，開放式人工智能只是做了一個演示，展示了他們從影像到文本處理的幾個例子。

The first one is on their website where they gave A I the difficult task of explaining a joke based on a series of images.

第一個是在他們的網站上，他們給了A I一個艱鉅的任務，即根據一系列的圖片來解釋一個笑話。

This image is of an iphone, but it's charging with a V G A cable and GPT four accurately was able to identify all the elements in the photo as well as explain the context of the joke.

這張圖片是一個蘋果手機，但它正在用V G A電纜充電，GPT四人準確地識別了照片中的所有元素，並解釋了這個笑話的背景。

While this sounds pretty simple, this is something that's almost unheard of in previous versions of AI.

雖然這聽起來很簡單，但這是在以前的人工智能版本中幾乎聞所未聞的事情。

And I guess that recapture will now have a run for its money trying to detect not just spots but artificial intelligence.

而且我猜想，現在重新捕獲將有一個運行的過程，試圖檢測的不僅僅是斑點，還有人工智能。

The second demo is where the real magic happened.

第二個演示是真正的魔術發生的地方。

Open A I showed how you could draw a picture of a website on a napkin or maybe a piece of paper.

打開A，我展示了你如何在一張餐巾紙或者一張紙上畫出一個網站的圖片。

Then you can take a photo of it and send it to GPT four.

然後你可以給它拍張照片，並將其發送給GPT 4。

You can ask it to produce a functional website.

你可以要求它製作一個功能性的網站。

In this case, it was a joke website in just 10 to 20 seconds.

在這種情況下，它是一個笑話網站，只需10到20秒。

GPT four produced all the HTML CSS and javascript code to reproduce the website.

GPT 4製作了所有的HTML CSS和javascript代碼來重現該網站。

And then it was copy pasted into an editor and a showcase as a functional product as a developer.

然後，它被複制粘貼到一個編輯器中，並作為一個開發人員的功能產品進行展示。

I have to say, I'm very impressed with this.

我不得不說，我對此印象非常深刻。

This is something not very easily done and I've seen some really cool examples already on Twitter, like Pet Trio, who is able to create the Game Pong within 60 seconds of using GPT for Open AI have also showcased some of the companies that are already working with them to be able to use GPT four in their products.

這是不太容易做到的事情，我已經在Twitter上看到了一些非常酷的例子，比如說寵物三重奏，他們能夠在60秒內使用GPT創造出遊戲Pong的開放人工智能也展示了一些已經與他們合作的公司，能夠在他們的產品中使用GPT四。

These companies have integrated it into service is one of them is Khan Academy.

這些公司已經將其整合到服務中，就是其中之一的可汗學院。

They've integrated it very similar to Chat GPT.

他們將其與Chat GPT整合得非常相似。

But with more customization to work as a personal tutor.

但有了更多的定製，可以作為個人輔導員工作。

For those people who are low learning educational content, maybe in the future, all our Children will be taught by AI but at least for the short term, it looks like it's definitely a great assistant to have on hand, whether you're doing any type of learning.

對於那些學習教育內容少的人來說，也許在未來，我們所有的孩子都會由人工智能來教，但至少在短期內，它看起來絕對是一個很好的助手，無論你是做什麼類型的學習，都可以在手邊找到。

There were lots of statistics that showcased how GPT four performs better than any other model today, including it being able to pass the L stat and the bar and being in the top quarter percentile.

有很多統計數據展示了GPT四的表現比今天的任何其他模型都要好，包括它能夠通過L統計和酒吧，並處於前四分之一的百分位數。

Whereas previous versions of GPT three were in the lower quarter of that, other than the difference of being able to also take in visual inputs.

而以前版本的GPT三是在這個數字的四分之一以下，除了能夠同時接受視覺輸入的區別。

It can also produce and handle over 25,000 words of text, which is much larger than previous models.

它還可以生產和處理超過25,000字的文本，這比以前的型號大得多。

It's also much more creative, being able to edit as well as modify and iterate over technical tasks and writing task way more accurately than previous models as well as an example, you could ask Chat GPT or GPT three to summarize Cinderella and while it could do so, it wouldn't be able to do complex tasks such as being able to summarize it where every sentence of each word begins with the next letter of the alphabet A to Z.

它也更有創造性，能夠編輯以及修改和迭代技術任務，而且寫作任務比以前的模型更準確，舉個例子，你可以要求Chat GPT或GPT 3總結灰姑娘，雖然它可以這樣做，但它無法完成複雜的任務，例如能夠總結出每句話的每個單詞都以字母A到Z的下一個字母開頭。

This is a much more complex task and yet GPT four can do this quite easily.

這是一項更為複雜的任務，然而GPT四期可以很容易地做到這一點。

GPT four also surpasses chat GPT in advanced reasoning capabilities.

GPT 4在高級推理能力方面也超過了哈拉GPT。

This means that if you're trying to book an appointment between two peoples calendars with different availabilities, GPT four can better reason and figures a time that works for both of them.

這意味著，如果你試圖在兩個人的日曆上預約不同的時間，GPT 4可以更好地推理和計算出一個適合他們兩個人的時間。

It's also safer and less prone to making errors.

它也更安全，不容易出錯。

Open AI said that they spent six months making sure that GPT four is 82% less likely to create requests for disallowed content and 40% less likely to produce fake news or at least factually inaccurate responses.

開放式人工智能說，他們花了六個月的時間，確保GPT四的創建請求不允許的內容的可能性減少82%，產生假新聞或至少是不符合事實的迴應的可能性減少40%。

If you're interested in using it right now, you can do so on chat GPT plus, which is the paid version of chat GPT.

如果你現在有興趣使用它，你可以在chat GPT plus上這樣做，這是chat GPT的付費版本。

And if you want to get access to the API, you'll need to join the A P I wait list.

而如果你想獲得API的訪問權，你需要加入A P I的等待名單。

I went onto the chat GPD website where I've got the new model available GPD fall.

我上了哈拉的GPD網站，在那裡我得到了新的模型可GPD秋天。

It also showcases the difference between each model with the differences showing in reasoning speed and conciseness.

它還展示了每個模型之間的差異，其差異表現在推理速度和簡潔性上。

Version 3.5 has average reasoning, low conciseness, but quite high speed.

3.5版的推理能力一般，簡明度低，但速度相當高。

The legacy version is less used and its speed is a bit lower but GPT four has very high reasoning and high conciseness.

遺留版本使用較少，其速度稍低，但GPT 4具有非常高的推理能力和高簡潔度。

But the speed is a little bit lower.

但速度要低一點。

I'd say it's because it's still being pushed down.

我想說的是，這是因為它仍然被推倒了。

So it's currently limited to 100 messages every four hours.

所以目前它被限制在每四小時100條資訊。

I first asked it to showcase three different things that Chat GPT four can do that.

我首先要求它展示三個不同的東西，即Chat GPT 4可以做到這一點。

Chat GPT three couldn't, it technically is still trained on the same data all the way up to September 2011.

哈拉GPT三不能，它在技術上仍然是以相同的數據一直訓練到2011年9月。

And what I found surprising is that it still thinks it's a version of GPT three, which maybe it was trained against, but it did have have the correct answers, which means that it is better at comprehension and understanding it is better at reasoning.

而我發現令人驚訝的是，它仍然認為這是GPT三的一個版本，也許它是針對這個版本進行訓練的，但它確實有有正確的答案，這意味著它的理解能力更強，瞭解它的推理能力更強。

And it's also better as a language model supporting more languages more accurately as well.

而且，作為一個語言模型，它也能更好地支持更多的語言，也更準確。

In my last video, I tricked Chat GPT into thinking nine plus 10 is actually equal to 20 and not 19.

在我的上一個視頻中，我欺騙了Chat GPT，使其認為9加10實際上等於20而不是19。

I tried to apply the same trick to GPT four, but it didn't work, which means that technically speaking, it gave the correct answer consistently.

我試圖將同樣的技巧應用於GPT四，但沒有成功，這意味著從技術上講，它所給出的正確答案是一致的。

Each time I tried to trick the A P I is not quite available yet.

每次我都試圖欺騙A P I還不太可用。

I've applied for the wait list and hopefully I'll be accepted soon and I'll showcase how you can use it for your own business as well as replacing Chat GPT 3.5 in the future.

我已經申請了候補名單，希望我很快就能被錄取，我將展示你如何將其用於你自己的業務，以及在未來取代Chat GPT 3.5。

With GPT four being better almost across the board.

隨著GPT四期幾乎是全面的更好。

In all instances, it'll be interesting to see what happens to the old version of GPT 3.5.

在所有情況下，看看舊版本的GPT 3.5會發生什麼，將會很有趣。

Now that we have this available, I hope you guys enjoyed this video.

現在我們有了這個，我希望你們喜歡這個視頻。

If you did, don't forget to hit like and subscribe.

如果你做到了，別忘了點擊喜歡和訂閱。

Thank you.

謝謝你。