Apple leapfrogging the leaders in the AI race and delivering the first killer use of AI.

蘋果公司在人工智能競賽中躍居領先地位，並首次將人工智能用於殺手級應用。

The so-called laggard jumping to historic highs.

所謂的落後者跳上歷史高點。

Apple also pushing to new record highs and reclaiming the title of the world's biggest company from Microsoft for the first time this year. With an AI that's not just indexing the world, solving math problems or translating languages, but an AI that is all about you.

蘋果公司的業績也再創新高，今年首次從微軟手中奪回全球最大公司的頭銜。人工智能不只是索引世界、解決數學問題或翻譯語言，而是以你為中心。

It has to understand you and be grounded in your personal context and already in your pocket. I'm Deirdre Bosa and this week on Tech Check, Apple takes the AI lead.

它必須瞭解你，立足於你的個人背景，並且已經在你的口袋裡。我是迪爾德麗-博薩（Deirdre Bosa），本週的 Tech Check 欄目中，蘋果公司將引領人工智能的發展。

Apple may have just pulled ahead of the leaders in the high-stakes AI race.

在高風險的人工智能競賽中，蘋果公司可能剛剛領先於領先者。

It didn't just promise us another chatbot, it promised us a personal assistant. The best explanation of what is going on right now came from DA Davidson, where the analyst said AI now stands for Apple Intelligence.

它不僅承諾給我們另一個哈拉機器人，還承諾給我們一個個人助理。戴維森公司（DA Davidson）的分析師說，AI 現在代表蘋果智能（Apple Intelligence）。

AI assistants or agents, they are the holy grail of gen AI at the moment.

人工智能助理或代理，是目前人工智能領域的聖盃。

It's killer use case.

這是一個殺手級用例。

And at WWDC, Apple showed us one that we can use with our existing device or an upgrade. Siri, when is my mom's flight landing?

在 WWDC 上，蘋果公司向我們展示了一款可以在現有設備或升級設備上使用的軟件。Siri，我媽媽的班機什麼時候降落？

What's awesome is that Siri actually cross-references flight details that my mom shared with me by email with real-time flight tracking to give me her up-to-date arrival time.

最棒的是，Siri 還能將我媽媽通過電子郵件與我分享的班機詳細信息與實時班機跟蹤進行交叉對比，為我提供她的最新抵達時間。

Possible only because of Apple's singular combination of hardware, software, wide base of users, and personal data. Our phones, tablets, and laptops, they know more about us than any other device with access to our texts, emails, calendar, and all the apps and all the information within them.

之所以能夠做到這一點，完全是因為蘋果公司將硬件、軟件、廣泛的用戶群和個人數據完美地結合在一起。我們的手機、平板電腦和筆記本電腦比其他任何設備都更瞭解我們，它們可以訪問我們的簡訊、電子郵件、日曆以及所有應用程序和其中的所有資訊。

Applying gen AI to that data set would make Apple Intelligence more efficient in a way that other players just haven't. And there are already some really impressive chat tools out there that perform a vast array of tasks using world knowledge.

將基因人工智能應用於該數據集，將使蘋果智能以其他玩家所沒有的方式變得更加高效。而且，現在已經有一些非常令人印象深刻的聊天工具，可以利用世界知識執行大量任務。

But these tools know very little about you or your needs.

但這些工具對您或您的需求知之甚少。

Here's one striking example.

下面就是一個鮮明的例子。

It can understand who my daughter is, the play details she sent several days ago, the time and location for my meeting, and predicted traffic between my office and the theater. Understanding this kind of personal context is essential for delivering truly helpful intelligence.

它可以瞭解我的女兒是誰，她幾天前發送的戲劇細節，我開會的時間和地點，以及我的辦公室和劇院之間的預測交通流量。要提供真正有用的情報，瞭解這種個人背景是必不可少的。

Truly helpful artificial intelligence.

真正有用的人工智能

Making use cases before it look abstract or even gimmicky.

讓使用案例看起來抽象，甚至是噱頭。

Like OpenAS, unveiling of GPT-4.0 involved telling jokes and singing lullabies.

與 OpenAS 一樣，GPT-4.0 的發佈也包括講笑話和唱搖籃曲。

A lullaby about majestic potatoes.

一首關於馬鈴薯的搖籃曲

Now that's what I call a mashup. Google's Project Astra told an alliteration and remembered where glasses were placed.

這就是我所說的混搭。谷歌的 "阿斯特拉計劃"（Project Astra）告訴人們一個典故，並記住了眼鏡的擺放位置。

Do you remember where you saw my glasses?

你還記得在哪裡看到我的眼鏡嗎？

Yes, I do.

是的，我知道。

Your glasses were on the desk near a red apple. To be sure, remarkable showcases of how far generative AI technology has come, but somehow less complete and intuitive than Apple's AI.

你的眼鏡就放在桌子上的紅蘋果旁邊。當然，這些都是人工智能生成技術的傑出展示，但在完整性和直觀性上卻比不上蘋果的人工智能。

Microsoft doesn't have it on the personal side, neither does Google, neither does Amazon.

在個人方面，微軟沒有，谷歌沒有，亞馬遜也沒有。

They all have pieces of the puzzle, but they don't have the whole picture.

他們都有拼圖的一部分，但沒有全貌。

They don't have all the personal texts, all the personal emails, the personal health information. Apple intelligence promises the entire picture, without any of the friction.

他們沒有所有的個人簡訊、個人電子郵件和個人健康資訊。而蘋果智能系統則能提供全貌，而且不會產生任何摩擦。

They look at the technology and they say, how can we play to our strength as the maker of the hardware and beat everyone else who's focused on either hardware or software?

他們在研究技術時會說，作為硬件製造商，我們如何才能發揮自己的優勢，擊敗其他專注於硬件或軟件的公司？

And that brings us to another key advantage, Apple's user base. What we're seeing now is that huge installed base of devices of over 2 billion is becoming a huge competitive advantage because of all of the personal data that Apple has on consumers.

這就引出了另一個關鍵優勢--蘋果的用戶群。我們現在看到的是，由於蘋果擁有消費者的所有個人數據，超過 20 億臺設備的龐大裝機量正在成為一項巨大的競爭優勢。

And it's unlike any other company. A huge audience of people, some of whom will have the technology already in their pockets, able to call it up by pressing a button or saying, hey Siri, instead of having to open the chat GPT app or browser. The tech that powers Apple intelligence is already inside Apple's latest generation devices, custom silicon, the M line of semiconductors.

它與其他任何公司都不同。龐大的閱聽人群體，其中一些人的口袋裡已經裝有這項技術，只需按一下按鈕或說一聲 "嘿，Siri"，就能調用它，而不必打開哈拉 GPT 應用程序或瀏覽器。為蘋果智能提供動力的技術已經在蘋果的最新一代設備、定製硅、M 系列半導體中得到應用。

The company has been developing for over a decade.

公司已經發展了十多年。

Those chips have dramatically improved phones, iPads, and Macs by making them more powerful, efficient, and secure.

這些芯片極大地改進了手機、iPad 和 Mac，使其功能更強大、更高效、更安全。

And now they're the key to Apple AI. They're so cutting edge that AI computing can happen on the device itself instead of being sent to the cloud.

現在，它們成了蘋果人工智能的關鍵。它們是如此尖端，以至於人工智能計算可以在設備本身進行，而不是發送到雲端。

The cornerstone of the personal intelligence system is on-device processing.

個人智能系統的基石是設備處理。

We have integrated it deep into your iPhone, iPad and Mac and throughout your apps.

我們已將其深度集成到 iPhone、iPad 和 Mac 以及整個應用程序中。

So it's aware of your personal data without collecting your personal data. It's the ability to process and analyze huge amounts of data and respond quickly without tapping into servers and data centers located far away from the user. Apple made its case for AI while relegating chatbots like chat GPT and Gemini to features, not products in and of themselves.

是以，它可以在不收集個人數據的情況下了解你的個人數據。這是一種處理和分析海量數據並快速做出反應的能力，而無需接入遠離用戶的服務器和數據中心。蘋果在為人工智能辯護的同時，將哈拉 GPT 和雙子座等哈拉機器人歸為功能，而非產品本身。

There are other artificial intelligence tools available that can be useful for tasks that draw on broad world knowledge or offer specialized domain expertise.

還有其他一些人工智能工具，對於需要利用廣泛的世界知識或提供專業領域知識的任務非常有用。

We're starting out with the best of these, the pioneer and market leader chat GPT from OpenAI powered by GPT 4.0.

我們首先要介紹的是其中的佼佼者--由 OpenAI 提供、由 GPT 4.0 支持的哈拉 GPT 的先驅和市場領導者。

And incorporating other models like Google's Gemini in the future. Apple making the point that these models are commoditized for the mainstream user.

並在未來融入谷歌的 Gemini 等其他機型。蘋果公司的觀點是，對於主流用戶而言，這些機型已經商品化。

They're widely available, accessible and not that differentiated.

它們廣泛存在，容易獲得，而且差別不大。

A feature in the larger Apple platform of vertical integration, giving it a unique edge over other players. The iPhone is still the best device out there for this early age of AI.

蘋果公司垂直整合大平臺的一個特點，使其擁有超越其他公司的獨特優勢。在人工智能的早期時代，iPhone 仍然是最好的設備。

It's multimodal like AI is becoming.

它是多模態的，就像人工智能正在變得多模態一樣。

A smartphone fits in your pocket, sees through its camera, listens, and talks back and texts when it needs to. Newer form factors have fallen flat, like the humane pen.

智能手機可以放在口袋裡，通過攝像頭觀察、聆聽，必要時還能回話和發短信。而新的外形設計則顯得平淡無奇，比如人性化的鋼筆。

Anything that you would search the web for or even use chat GPT for, you can use on your pen.

您在網上搜索的任何內容，甚至是使用哈拉 GPT 的內容，都可以在您的筆上使用。

And the rabbit, R1.

還有兔子，R1。

I just created an order for a 12-inch pizza.

我剛下了一個 12 英寸披薩的訂單。

It's going to be hand-tossed with a classic marinara sauce and topped with regular cheese.

它將用手工拌上經典的意大利麵醬，再澆上普通奶酪。

Please confirm your order.

請確認您的訂單。

It's also the new worst product I think I've ever reviewed.

這也是我評測過的最糟糕的產品。

There's just so many things bad about it.

它的壞處實在太多了。

Far from ready for primetime.

遠未達到黃金時段的要求。

Weak battery life.

電池續航能力較弱。

This one can sit in front of you doing nothing and the battery just visibly is draining and dies in like four hours.

這塊電池放在你面前，什麼也不做，電池就會明顯耗盡，大概四個小時後就會沒電。

Slow.

慢。

What's the weather?

天氣如何？

Finding your temperature.

找到你的溫度

Missing features.

功能缺失。

Translate this text into English.

將這段文字翻譯成英文。

Sorry, the translation feature is not available in vision mode.

抱歉，翻譯功能在視覺模式下不可用。

Prone to overheating.

容易過熱。

But this ran so hot that I accidentally burned myself several times while either crossing my arms or leaning forward.

但這東西太燙了，我好幾次在雙手交叉或身體前傾時不小心燙傷了自己。

Often inaccurate.

經常不準確。

What is this?

這是什麼？

You are holding a taco.

你拿的是玉米卷。

And plagued by a host of other issues. I think AI gadgets might just be impossible.

而且還受到其他一系列問題的困擾。我認為人工智能小工具可能根本不可能實現。

But VC investor Vinod Khosla points out that all devices in its first iterations have shortcomings.

但風險投資人維諾德-科斯拉（Vinod Khosla）指出，所有設備在最初的迭代中都存在缺陷。

People forget.

人們忘記了。

When the iPhone was born, you couldn't even change the home screen.

iPhone 誕生時，你甚至無法更改主螢幕。

You couldn't copy and paste for three years after the iPhone was introduced.

iPhone 推出三年後，你還不能複製和粘貼。

That getting to a million devices, that is the tipping point.

設備數量達到 100 萬臺，就是臨界點。

Once you get to a million devices in use, you get so much feedback from the maker of the device that the real advantage in improving it and continuing to add features and capabilities. Khosla argues that the AI device of the future may not be a smartphone.

一旦有一百萬臺設備投入使用，你就能從設備製造商那裡獲得大量反饋，從而在改進設備、不斷增加功能和性能方面獲得真正的優勢。科斯拉認為，未來的人工智能設備可能不是智能手機。

This new kind of device for very rapid, quick interaction to get things done, to send yourself a note, a reminder, look across Google traffic and Uber arrival times and where you need to get to from your calendar.

這種新型設備可以進行非常快速的互動，快速完成工作，給自己發便條、提醒，查看谷歌交通和 Uber 到達時間，以及日曆中需要到達的地點。

All will be done by an agent, AI, just by telling it, make sure I'm on time for my next meeting.

一切都將由人工智能代理完成，只需告訴它，就能確保我準時參加下一次會議。

And that interface has to be less than 10 seconds, not open a phone, open multiple apps, and try and do these complex things.

這個界面必須在 10 秒鐘內完成，而不是打開手機，打開多個應用程序，嘗試做這些複雜的事情。

They'll be done for you. Apple Intelligence will also have to deliver.

它們將為你完成。蘋果智能公司也必須提供服務。

And it will have to be good enough for users to shell out for an upgrade.

它必須足夠好，用戶才會掏錢升級。

It'll only be available on the iPhone 15 Pro, its latest, more expensive model with better capabilities and an upgraded chip.

它只會在 iPhone 15 Pro 上使用，該機型是其最新的、更昂貴的機型，擁有更好的功能和升級的芯片。

But most iPhone owners, they carry an earlier model and they're waiting longer to upgrade. Apple and its investors are banking on those owners to upgrade when Apple unveils a new model in the fall.

但大多數 iPhone 用戶使用的是早期機型，他們等待升級的時間更長。蘋果公司及其投資者寄希望於這些用戶在蘋果秋季推出新機型時進行升級。

Some owners may want to wait and see how the new technology actually works in the wild.

有些車主可能想等等看新技術在實際應用中的效果如何。

Apple's AI demo was taped and carefully produced, but we've seen from OpenAI, Microsoft, and Google demos and launches that generative AI tools are imperfect and unpredictable in their first iterations. From AI laggard to potential AI leader, Apple is another reminder that the AI arms race moves quick and has a lot more room to run.

蘋果的人工智能演示是經過錄制和精心製作的，但我們已經從 OpenAI、微軟和谷歌的演示和發佈中看到，生成式人工智能工具在最初的迭代中是不完美和不可預測的。從人工智能的落後者到潛在的人工智能領導者，蘋果公司再次提醒我們，人工智能軍備競賽進展迅速，還有很大的發展空間。