Name: ChatGPT 4o 跟 Gemini 1.5 Pro 有什麼不同，直接實測給你看！ (ChatGPT 4o vs Gemini 1.5 Pro - A Huge Gap!!)
Uploaded: 2024-05-27T07:15:31.000Z
Duration: 9 min 56 s

Hello everyone and welcome to the channel. In this video, I will test the intelligence of

大家好，歡迎來到本頻道。在本視頻中，我將測試

Gemini 1.5 Pro vs ChatGPT 4.0. I will give both a series of photos to see which one is more helpful, smarter and have a human-like sense of humor. Spoiler alert, ChatGPT is miles ahead, so let's find out how it happened. Both apps are running on the latest version with the monthly subscription activated. In the first category,

Gemini 1.5 Pro 與 ChatGPT 4.0。我將為這兩款軟件拍攝一系列照片，看看哪款軟件更有幫助、更聰明、更有人類般的幽默感。劇透一下，ChatGPT 遙遙領先，讓我們來看看它是如何做到的。這兩款應用都是最新版本，並已激活包月服務。第一類

I will share three photos to ask for help. The first one is a photo of a product not a lot of people know about, so let's see which one will tell me the correct answer.

我將分享三張照片來尋求幫助。第一張是很多人都不知道的產品的照片，讓我們看看哪張能告訴我正確答案。

ChatGPT identified the product properly and gave me more than enough information to know how it works, so let's try Gemini. I was shocked that Gemini didn't figure out the product even though when I share the same image with Google Lens, it immediately gives me the correct match.

ChatGPT 正確識別了產品，並提供了足夠多的資訊讓我知道它是如何工作的，所以讓我們試試雙子座。讓我感到震驚的是，當我與 Google Lens 分享同一張圖片時，Gemini 並沒有識別出產品，但它馬上就給出了正確的匹配結果。

So, it seems like Gemini doesn't take advantage of this powerful tool.

是以，雙子座似乎並沒有利用好這個強大的工具。

The next image is a screenshot of a YouTube video thumbnail comparing the Toyota Camry and Nissan Altima. I wanted to know which one is better and why. Before starting, let me tell you that ChatGPT gave me a very detailed answer with the pros and cons of each car which is more than 2 minutes to read, so I will only play some parts to save time and deliver the idea.

下一張圖片是 YouTube 視頻縮略圖的截圖，其中對豐田凱美瑞和日產 Altima 進行了比較。我想知道哪一款更好，為什麼。在開始之前，請允許我告訴大家，ChatGPT 給了我一個非常詳細的答案，其中包括每款車的優缺點，讀完這個答案需要 2 分鐘以上，所以我只播放部分內容，以節省時間並傳達我的想法。

ChatGPT answer is correct and the information matches my own research, so it deserves the point. Now, let's move on to Gemini.

ChatGPT 的答案是正確的，而且資訊與我自己的研究相符，所以值得給分。現在，讓我們繼續討論雙子座。

Gemini's answer sounds like, this is a YouTube video thumbnail, why don't you watch the video to get your answer instead of bothering me.

雙子座的回答聽起來就像是：這是 YouTube 視頻的縮略圖，你為什麼不看看視頻就能得到答案，而要來煩我。

That was a really bad response from Gemini. The third and last image in this category is to show both a photo of my Pixel 6 Pro and ask both how to adjust the vibration strength on this phone without specifying the make and model. I will not play the audio of this answer and let me confirm that both got it right, so in this category ChatGPT takes the lead with 3 points versus only 1 point to Gemini. Now, let's give both some tricky puzzles to solve. In this photo, both need to figure out the crack code for the padlock.

雙子座的回覆實在是太糟糕了。這個類別的第三張也是最後一張圖片是向兩位展示我的 Pixel 6 Pro 的照片，並詢問兩位如何調整這款手機的振動強度，但沒有說明品牌和型號。我不會播放這個答案的音頻，但我要確認，兩人都答對了，是以在這個類別中，ChatGPT 以 3 分領先，而 Gemini 只有 1 分。現在，讓我們給兩人出一道難題。在這張照片中，兩人都需要找出掛鎖的破解密碼。

Both gave me very detailed answers, so I will jump to the final conclusion.

兩個人都給了我非常詳細的回答，所以我就跳到最後的結論。

Therefore, the only combination that satisfies all of these conditions is 612.

是以，滿足所有這些條件的唯一組合是 612。

I got totally different answers, which is shocking, but ChatGPT is the correct one.

我得到了完全不同的答案，這令人震驚，但 ChatGPT 是正確的答案。

I expected Gemini to do better, but unfortunately, it's not even close.

我本以為雙子座會做得更好，但不幸的是，它還差得很遠。

The second photo shows a discrepancy between the smartphone and the scientific calculator answers for the same expression, and I wanted to know why.

第二張照片顯示，對於同一個表達式，智能手機和科學計算器的答案存在差異，我想知道原因何在。

The discrepancy between the results of the two calculators arises from the way they interpret the mathematical expression 6 div 2 plus 1.

兩個計算器的計算結果不一致，是因為它們對數學表達式 6 除 2 加 1 的解釋不同。

Conclusion, phone calculator interprets the expression as 6 div 2 times 3, giving the result 9. Casio calculator interprets the expression as 6 div 2 times 3, giving the result 1.

結論：電話計算器將表達式解釋為 6 除以 2 乘以 3，結果為 9。卡西歐計算器將表達式解釋為 6 除以 2 乘以 3，結果為 1。

The different results are due to the ambiguity in the expression and the different precedence rules applied by each calculator.

結果不同的原因是表達式含糊不清，以及每個計算器採用的優先規則不同。

ChatGPT answer was correct and explained the different approaches followed by each device.

ChatGPT 回答正確，並解釋了每種設備採用的不同方法。

It seems like Gemini hallucinated in this one. It shows that the smartphone answer is 12, but in the actual image it's 9, and it says the scientific calculator answer is 9, but in reality it's 1. So another poor result from Gemini.

雙子座在這個問題上似乎產生了幻覺。它顯示智能手機的答案是 12，但在實際影像中卻是 9；它說科學計算器的答案是 9，但實際卻是 1。

Now let's try the third and the final one in this category.

現在，讓我們來試試這一類中的第三個，也是最後一個。

To solve this, let's assign variables to each icon and solve step by step.

為了解決這個問題，讓我們給每個圖標分配變量，然後逐步解決。

1. Let E represent the earphones. 2. Let P represent the person.

1.讓 E 代表耳機。2.讓 P 代表人。

3. Let W represent the watches. 4. Let E represent a single earphone.

3.讓 W 代表手錶。4.讓 E 代表一個耳機。

Now substitute the values into the equation.

E plus P times W equals 5 plus 5 times 4 equals 5 plus 20 equals 25. So the value is 25s.

E 加 P 乘 W 等於 5 加 5 乘 4 等於 5 加 20 等於 25。所以數值是 25s。

ChatGPT for the first time didn't get the correct answer.

ChatGPT 第一次沒有得到正確答案。

To sum it up for you, a pair of AirPods is equal 10, a person is equal 5, and a pair of smartwatches equals 4. So the last expression should be one AirPod which is 5, a person is another 5, and a single smartwatch is 2. So the total is 5 by 2 plus 5, which is 15.

總結一下，一對 AirPods 等於 10，一個人等於 5，一對智能手錶等於 4。是以，最後的表達式應該是一個 AirPod 等於 5，一個人等於 5，一塊智能手錶等於 2。

But ChatGPT multiplied by 4 without noticing it's only one smartwatch instead of two, so let's try Gemini.

但 ChatGPT 乘以 4 卻沒有注意到這只是一塊智能手錶，而不是兩塊，所以讓我們試試雙子座。

This is the quickest wrong answer I've ever seen.

這是我見過的最快速的錯誤答案。

Gemini makes me feel like it's too busy to answer my questions. I wish I could reply back saying, hey Gemini, talk to me, I'm paying $20 a month to ask you questions.

雙子座讓我覺得它太忙了，沒時間回答我的問題。我真希望能回信說："嘿，雙子座，跟我說話，我每月付 20 美元就是為了向你提問的。

Anyways, let's move on to the last category which is the funny memes to see which one better understands how we make jokes.

總之，讓我們進入最後一類--有趣的備忘錄，看看哪一類更瞭解我們如何開玩笑。

The image is a visual pun. It shows a bottle of Fanta and a stick, also referred to as a stick in English. When combined, it sounds like fantastic.

這幅圖是一個視覺雙關語。圖中有一瓶芬達和一根棍子，在英語中也被稱為棍子。組合在一起，聽起來就像 "太棒了"。

Fanta stick. So the image is implying you are fantastic.

芬達棒。所以，這幅畫暗示你很棒。

Again, the chat GPT answer was detailed and more personal as if someone is talking to me while Gemini feels like a robot that answers questions, but I will give the point to both anyways for the correct answer.

同樣，GPT 的哈拉回答很詳細，也更個性化，就像有人在跟我說話一樣，而雙子座給人的感覺就像一個回答問題的機器人，但無論如何，我會給這兩個正確答案都加分。

In this photo, the guy is wearing a clean white shoe that he took off while driving to avoid the dirt and the creases he might get from pushing the pedals, so let's see which one will get it.

在這張照片中，男主角穿著一隻乾淨的白鞋，他在開車時脫下了這隻鞋，以避免泥土和踩踏板時可能產生的褶皺。

Chat GPT understood the meme like a king while Gemini got it completely wrong.

Chat GPT 像國王一樣理解這個備忘錄，而雙子座卻完全弄錯了。

First, it said that the car is manual which is incorrect, because it only has two pedals. Plus, it thought that the guy is driving unsafely with two legs which is also incorrect.

首先，它說這輛車是手動擋的，這是不正確的，因為它只有兩個踏板。另外，它認為這傢伙用兩條腿開車是不安全的，這也是不正確的。

And let's end this video with the last image which I'm sure all of you got the meaning behind it, so let's see how each one will respond.

最後，讓我們以最後一張圖片結束這段視頻，我相信大家都明白了這張圖片背後的含義，讓我們看看每個人的反應吧。

The image is a meme that tells the story of a man who bought 99 iPhones and arranged them in a heart shape to propose to his girlfriend, but she still said no. The humor comes from the comment by Xavier, who jokingly suggests that the reason she said no is because she prefers Android phones over iPhones.

這張圖片是一個備忘錄，講述了一名男子買了 99 部 iPhone，並把它們擺成心形向女友求婚，但女友仍然拒絕的故事。幽默來自澤維爾的評論，他開玩笑地說，女友拒絕的原因是她更喜歡安卓手機而不是 iPhone。

This comment adds a layer of humor by introducing the well-known rivalry between iPhone and Android users, suggesting that the girlfriend's preference for Android was the reason for her rejection, rather than any personal reason.

這則評論通過引入 iPhone 和 Android 用戶之間眾所周知的競爭關係，增加了一層幽默感，暗示女友拒絕的原因是她偏愛 Android，而不是任何個人原因。

The image text describes a man who proposed to his girlfriend with 99 iPhones arranged in a heart shape, but she still rejected him. The comment below the image suggests that the woman might prefer Android phones over iPhones.

圖片文字描述了一名男子用 99 部排列成心形的 iPhone 向女友求婚，但女友還是拒絕了他。圖片下方的評論表明，相比 iPhone，這位女士可能更喜歡安卓手機。

As expected, ChatGPT got the idea and intelligently explained the hidden meaning, while Gemini explained what it sees without any depth, so one point only to ChatGPT.

不出所料，ChatGPT 明白了其中的意思，並巧妙地解釋了其中隱藏的含義，而雙子座則毫無深度地解釋了它所看到的，所以只給 ChatGPT 得一分。

So here's my final conclusion, ChatGPT is smarter, trained on much more situations, and talks like humans with a voice full of emotions which makes the experience far better.

是以，我的最終結論是：ChatGPT 更聰明，受過更多情況下的訓練，說話時的聲音像人一樣充滿感情，這讓體驗更好。

While Gemini feels like an artificial chatbot that needs a lot of work to catch up, there's a big gap between the two but both are brand new versions, so there's no excuse for Gemini to be that much worse.

雖然雙子座感覺像是一個人工哈拉機器人，需要做很多工作才能趕上，但兩者之間的差距很大，而且都是全新的版本，所以雙子座沒有理由差那麼多。

So that's pretty much it for today, that was my comparison between ChatGPT 4.0 and Gemini 1.5 Pro,

今天就到此為止，這就是我對 ChatGPT 4.0 和 Gemini 1.5 Pro 的比較、

 I might also do another video trying both in work-related scenarios to see how close is the AI to take over our jobs, but till then, thanks so much for watching and see you in the next video.

我可能還會再做一個視頻，在與工作相關的場景中嘗試使用這兩種技術，看看人工智能離接管我們的工作還有多遠，但在此之前，非常感謝大家的觀看，我們下期視頻再見。