Name: Spotify 最愛的特色背後，竟然藏著這 AI 科技？ | 2025.12.18 (The AI-tech behind your favorite Spotify feature | December 18, 2025)
Uploaded: 2026-02-21T05:50:32.000Z
Duration: 10 min

We're at the penultimate day of the week, so let's get to your news while we still have time.

今天是這週倒數第二天，趁還有時間，我們趕快來看看今天的新聞。

The folks at one of the United States largest auto manufacturers for Ford Motor Company say they are going to slow their role on their electric vehicle plants to put more focus on their gas and hybrid vehicles.

美國數一數二大的汽車製造商 Ford Motor Company 的朋友們表示，他們將會放慢電動車工廠的腳步，把更多重心放在汽油和油電混合車款上。

Part of their new plan putting production of their flagship EV, the F150 Lightning, on hold indefinitely.

他們新計畫的一部分，是無限期暫停生產他們的旗艦電動車 F150 Lightning。

This stems in part from US President Donald Trump's policies, which are impacting the EV industry.

這部分原因來自美國總統川普的政策，這些政策正影響著電動車產業。

The move for Ford means a $19.5 billion hit against its fourth quarter earnings.

Ford 的這項舉動，將會讓他們第四季的財報損失 195 億美元。

Automakers around the world have invested big time in EVs, but

全球的汽車製造商都大力投資電動車，但是

Ford 表示他們看到了需求下降。

The federal tax credit, from which EV buyers once benefited, ended in September.

電動車買家曾經受惠的聯邦稅收抵免，在九月結束了。

The US Government also took away some of the environmental regulations that went into effect during former President Joe Biden's term.

美國政府也取消了一些在前總統拜登任期內生效的環保法規。

Did your Spotify Wrapped feel spot on this year?

你今年的 Spotify 年度回顧聽起來很準嗎？

You know where Spotify tells you which songs are your favorites?

你知道 Spotify 告訴你哪些歌是你最愛的嗎？

If so, it's because the streaming music service aggregates which or collects all the songs you played throughout the year by using machine learning algorithms behind the scenes.

如果是的話，那是因為這家串流音樂服務透過幕後使用 Machine Learning 演算法，匯總了你整年播放的所有歌曲。

It's just one example of our digital experiences becoming more personalized by technology.

這只是我們的數位體驗透過科技變得更加個人化的一個例子。

Our Claire Duffy has more hey Coy, have you ever found yourself wondering how your Spotify Wrapped seems to just get you so perfectly?

我們的 Claire Duffy 有更多報導，嘿 Coy，你曾經好奇過你的 Spotify 年度回顧為什麼好像能這麼精準地抓到你嗎？

Or why Netflix knows how to recommend the exact show you wanted to watch next?

或是 Netflix 怎麼知道要推薦你下一部想看的節目？

It's all thanks to a technology called machine learning.

這都要歸功於一項叫做 Machine Learning 的技術。

Machine learning is a process where a computer takes a bunch of data and uses an algorithm, essentially a mathematical equation, to find patterns in that data to draw conclusions or make recommendations.

Machine Learning 是一個過程，電腦會接收大量的資料，並利用演算法，也就是一個數學方程式，從這些資料中找出模式，進而得出結論或提出建議。

So think sorting through all of the music you listened to this year to determine who your favorite artists were.

所以，就像是整理你今年聽的所有音樂，來判斷出你最喜歡的歌手是誰。

You might have heard of machine learning in the context of artificial intelligence.

你可能聽過 Machine Learning 和人工智慧 (AI) 的關聯。

That's because it also powers many of the AI tools we use today.

因為它也是我們今天使用的許多 AI 工具的動力來源。

AI chatbots, for example, were trained on a bunch of human writing, and

舉例來說，AI 哈拉機器人透過大量人類的寫作內容進行訓練，

they use their knowledge of the patterns in our language to respond to our questions.

然後利用它們對我們語言模式的理解來回應我們的問題。

Here's Georgia Tech's Brian McGurko to explain.

接下來，喬治亞理工學院的 Brian McGurko 來為我們說明。

So like for for using ChatGPT or Llama or something like that, you know, these companies scraped the Internet, grabbed as much text as they could from from Reddit conversations and forums and blog posts and news articles.

所以像是使用 ChatGPT 或 Llama 這類工具時，你知道的，這些公司會爬取網路，從 Reddit 對話、論壇、部落格文章和新聞報導中抓取盡可能多的文字。

And the patterns that they found are like what words go with what Other words.

它們找到的模式就像是哪些詞會跟其他哪些詞一起出現。

And so when you give it words, it goes, hey, what's the most likely word that would go next?

所以當你給它一些詞時，它就會想，嘿，下一個最有可能出現的詞是什麼？

Now, online platforms don't just use information about us to serve us recommendations.

現在，線上平臺不只利用我們的資訊來推薦內容給我們。

They also look at what content people like us are consuming using a type of machine learning called collaborative filtering.

它們也會看看跟我們一樣的人在看什麼內容，這用到一種叫做協同過濾的 Machine Learning。

Spotify is really, really good at collecting data and describing data.

Spotify 在收集和描述資料方面做得非常非常棒。

So one of the things that they use and has been used in lots of the websites and apps that you use is this approach called collaborative filtering.

所以它們使用的一種方法，也是你使用的許多網站和 App 中用過的方法，就叫做協同過濾。

So Spotify, when you say, hey, give me some music that, that I think I'll like, like you hit the, the DJ button or something and what it's doing is it's going, okay, let me look at this guy.

所以 Spotify，當你說，嘿，給我一些我可能會喜歡的音樂，像是你按了 DJ 按鈕什麼的，它其實是在想，好的，讓我看看這個人。

好的，跟他一樣的人都聽了什麼？

And so collaborative filtering is underneath all of these sort of like, you know, magic suggest with Spotify Rapt or the auto generated playlists.

所以協同過濾其實就在這些 Spotify 的「魔幻推薦」或自動生成的播放清單背後。

It can recognize, like, hey, you know, there are not that many goth circus ragtime people out there, but

它能辨識出，嘿，你知道，喜歡哥德式馬戲團拉格泰姆音樂的人不多，但

You know, it can help you sort of identify if you're on this island of music appreciation that you can brag to your friends about.

你知道，它可以幫你找出你在音樂品味這座孤島上的獨特之處，讓你跟朋友炫耀。

It's important to understand that apps and online platforms are taking in data you share with them, whether by listening to certain songs or liking certain photos, and

重要的是要了解，App 和線上平臺會收集你分享給它們的資料，無論是透過聽某些歌曲或按讚某些照片，

That can be used for fun applications like giving you your Spotify wrapped.

這可以用在有趣的應用上，像是提供你的 Spotify 年度回顧。

Spotify says my listening age is 73, by the way, thanks to my love of classic rock.

Spotify 說我的聽歌年齡是 73 歲，順帶一提，這都要歸功於我對經典搖滾的熱愛。

But machine learning can also be used to do things like target you with ads for things a company thinks you're more likely to buy.

但 Machine Learning 也可以用來做一些事，像是針對你投放你可能比較會買的商品的廣告。

And I don't know about you, but this time of year that is the last thing I need.

我不知道你怎麼樣，但這一年中的這個時候，這是我最不需要的東西。

Still, the more we know about how tech platforms collect and use our data, the the better we can choose how to consciously engage with them.

不過，我們越瞭解科技平臺如何收集和使用我們的資料，我們就越能有意識地選擇如何與它們互動。

Volkswagen's iconic van was originally called what?

福斯汽車的經典廂型車最初叫做什麼？

The Microbus, the Type 2, the Wonder Wagon, or the Mystery Machine.

Microbus、Type 2、Wonder Wagon，還是 Mystery Machine。

如果你說是 Type 2，那你太厲害了。

The type two, later renamed the transporter, started production in 1950.

Type 2，後來改名為 Transporter，於 1950 年開始生產。

It soon became a cultural symbol of freedom, road trips and the 1960s counterculture.

它很快就成了自由、公路旅行和 1960 年代反文化運動的文化象徵。

Type 1 就是那臺獨一無二的福斯金龜車。

Now to a viral moment that became a symbol of hope.

現在我們要看一個病毒式傳播的時刻，它成了一個希望的象徵。

This vintage Volkswagen bus was spared from the LA fires even as the rest of the neighborhood was decimated.

這臺老舊的福斯廂型車在洛杉磯的火災中倖存下來，而鄰近的區域全都被摧毀了。

The bus has since undergone a full remodel and is getting some newfound love and national attention.

這臺廂型車後來經過了完整的翻修，獲得了新的關注和全國性的注目。

我們的 Nick Watt 會為你報導更多。

許多人從這張照片中看到了希望。

The little Blue Miracle, a 1977 VW Microbus among the grays and blacks of destruction and despair.

這臺 1977 年的福斯 Microbus，在滿是灰燼和絕望的毀滅景象中，被稱為「藍色小奇蹟」。

The burnt headlight out front was on the case.

前面燒毀的大燈當時就在關注這件事。

Found the bus and so we wanted to know who owned it to to get the story here.

我們找到了這臺巴士，所以想知道是誰擁有的，來把這個故事告訴大家。

Current owner Megan Weinreber Malibu surfer I'm speechless to see all the houses burnt and I can't believe that Azul the bus, that's what we named it, survived like it's.

現任車主 Megan Weinreber，一位馬里布的衝浪者，看到所有房子都燒毀了，我說不出話來，我不敢相信我們取名為 Azul 的這臺巴士竟然倖存了下來。

Well, Volkswagen folks also saw that viral pic of the little singed survivor and they asked if they could help.

嗯，Volkswagen 的員工們看到那張在網路上瘋傳的、關於這臺小小的倖存者被燒焦的照片，他們就問能不能幫忙。

他們把 Azul 接走，然後開始動工。

We found the car with the busted rear windows.

我們發現這臺車的後車窗都碎了。

We had ashes inside, we had embers inside and nothing combusted.

車裡都是灰燼，車裡有餘燼，但什麼都沒燒起來。

They worked on Azul for 2000 hours to bring a bus back to life.

他們花了 2000 小時來修復 Azul，讓這臺巴士重獲新生。

That eye catching blue paint had to be custom mixed.

那個引人注目的藍色漆，還得特別調製。

But what about the Palisades Then now nearly 7,000 structures destroyed.

但 Palisades 那邊呢？現在幾乎有 7,000 棟建築物被摧毀。

So far, only one home has been completely rebuilt.

到目前為止，只有一戶人家完全重建好。

A certificate of occupancy was issued just last week.

Azul is now back, charm intact and even better than before.

Azul 現在回來了，魅力依舊，甚至比以前更好。

字幕列表影片播放

Spotify 最愛的特色背後，竟然藏著這 AI 科技？ | 2025.12.18 (The AI-tech behind your favorite Spotify feature | December 18, 2025)

sort

bunch

awesome

grab