Name: 為什麼蘋果的 Siri 在人工智能時代依然如此糟糕？ (Why Apple’s Siri Is Still So Bad In The Age Of AI)
Uploaded: 2025-06-06T02:54:44.000Z
Duration: 15 min 28 s

基礎被動語態

Back in 2011, when Apple introduced Siri, it was billed as a breakthrough feature for the iPhone.

早在 2011 年，當蘋果公司推出 Siri 時，它就被譽為 iPhone 的一項突破性功能。

程度副詞

Siri was a really big deal when it first launched back in 2011 because it was a whole new way to interact with your iPhone.

Siri 早在 2011 年推出時就備受關注，因為它是與 iPhone 互動的全新方式。

This was actually a pet project of Steve Jobs.

這實際上是喬布斯的一個寵物項目。

He saw Siri demoed as a separate app before Apple acquired the company and was just kind of blown away by what Siri could do.

在蘋果收購 Siri 之前，他看到 Siri 作為一個獨立應用進行演示，Siri 的功能讓他大開眼界。

He really saw the long term vision for this kind of paradigm where you talk to your devices instead of typing.

他真正看到了這種模式的長遠前景，在這種模式下，你可以與你的設備對話，而不是打字。

And at the time, it seemed very revolutionary.

在當時，這似乎非常具有革命性。

But 15 years after it launched, the voice assistant still mishears things, misses context.

但在推出 15 年後的今天，這款語音助手仍然會聽錯東西，錯過上下文。

And now the rollout of a new and improved Siri has been delayed.

而現在，全新改進版 Siri 的推出又被延後了。

And last year, the tech giant introduced Apple Intelligence.

去年，這家科技巨頭推出了蘋果智能系統（Apple Intelligence）。

But critics say Apple has not evolved to meet the new AI standards set by OpenAI's ChatGPT, Google's Gemini or Amazon's Alexa.

但批評者認為，蘋果還沒有發展到符合 OpenAI 的 ChatGPT、谷歌的 Gemini 或亞馬遜的 Alexa 所制定的新人工智能標準。

If you look at everything they announced in WWDC last year, a lot of that hasn't come to fruition.

如果你看看他們去年在 WWDC 上宣佈的一切，很多都還沒有實現。

Since the November 2022 release of ChatGPT, generative AI has taken off.

自 2022 年 11 月發佈 ChatGPT 以來，生成式人工智能已經開始起飛。

Tech giants like Microsoft and Google have centered their strategies around large language models, LLMs, from search to productivity.

從搜索到生產力，微軟和谷歌等科技巨頭的戰略都圍繞著大型語言模型（LLM）展開。

但也有人說，蘋果已經落後了。

You had ChatGPT, you had Anthropics Cloud and so many other of these artificial intelligence products coming out there outside of Apple's ecosystem.

在蘋果生態系統之外，還有 ChatGPT、Anthropics Cloud 和其他許多人工智能產品。

And Apple still hasn't been able to catch up.

蘋果公司的股票也受到了打擊。

Over the past six months, it's fallen more than 16 percent, while other tech companies like Microsoft and Google have pushed ahead in AI.

在過去的六個月裡，它的跌幅超過了 16%，而微軟和谷歌等其他科技公司則在人工智能領域取得了長足進步。

If Apple was as successful in their own pursuits for generative AI as we've seen some of their megacap competitors, Apple's stock price would be higher than it is today.

如果蘋果公司能像我們看到的一些大市值競爭對手一樣，在追求人工智能生成方面取得成功，那麼蘋果公司的股價就會比現在更高。

In November 2023, during an interview with CNBC, Apple's hardware leaders insisted the company wasn't falling behind.

2023 年 11 月，在接受 CNBC 採訪時，蘋果公司的硬件負責人堅稱公司並沒有落後。

Critics have been surprised that Apple appears to be falling behind when it comes to AI.

蘋果公司在人工智能方面似乎落後了，這讓評論家們感到驚訝。

Since that interview, the AI landscape has moved fast and Apple's position in it has only grown more uncertain.

自那次採訪以來，人工智能的發展日新月異，蘋果在其中的地位也越來越不確定。

And can Apple still lead in AI or is it too late to catch up?

蘋果還能在人工智能領域保持領先嗎？

While some iPhone users love the latest updates to Siri.

雖然有些 iPhone 用戶喜歡 Siri 的最新更新。

The new interface that lights up a rainbow around the edges is so much better and it reacts to your voice in real time.

新界面的邊緣會亮起彩虹，效果非常好，而且還能對你的語音做出實時反應。

But most importantly, Siri is actually smarter now.

但最重要的是，Siri 現在真的更智能了。

Others say it's not the way it should be in the age of AI.

還有人說，這不是人工智能時代應有的方式。

Dear Tim Cook, your Hey Siri does not work.

親愛的蒂姆-庫克，你的 "嘿 Siri "不能用了。

I need to know if it's only me that is having severe issues with Siri.

我想知道是否只有我才會遇到 Siri 的嚴重問題。

I, it literally can't spell a sentence like correctly.

Hey, Siri, could you send a message, please?

The brand new Siri isn't really much of an update.

全新的 Siri 其實並沒有什麼更新。

For example, is the NBA player Zach Levine playing tonight?

例如，NBA 球員扎克-萊文今晚有比賽嗎？

It should be able to tell me instead of breaking up stuff from the web.

它應該能告訴我，而不是把網上的東西拆散。

Siri is one where there was so much promise, but obviously been very disappointing, especially after that first splash.

Siri 曾經大有可為，但顯然非常令人失望，尤其是在首次亮相之後。

While Siri was groundbreaking in 2011, Apple didn't build on that early lead.

雖然 Siri 在 2011 年具有開創性意義，但蘋果並沒有在早期領先的基礎上再接再厲。

Over time, the assistant fell behind as rivals launched smarter, faster, more adaptive AI tools.

隨著時間的推移，隨著競爭對手推出更智能、更快速、適應性更強的人工智能工具，該助手逐漸落後。

There are a number of things that have held Siri back since it launched several years ago.

自幾年前推出 Siri 以來，有許多問題阻礙了它的發展。

And it didn't take long for competitors to really catch up to what Siri is doing.

沒過多久，競爭對手就真正趕上了 Siri 的腳步。

Apple has traditionally been a hardware company.

蘋果公司歷來是一家硬件公司。

They make great hardware, not so great at the software and services bit of things.

他們的硬件做得很好，但軟件和服務方面卻不盡如人意。

And so Siri kind of stayed the same for several years after it first launched.

是以，在 Siri 首次推出後的幾年裡，它一直沒有什麼變化。

Apple's strategy has been focused on building generative AI models into its devices, aiming to make daily tasks easier.

蘋果公司的戰略重點是在其設備中建立人工智能生成模型，旨在讓日常工作變得更輕鬆。

But industry watchers say Apple has been reactive and slow to match the moves of rivals like Google and Microsoft.

但業內觀察人士認為，蘋果公司在與谷歌和微軟等競爭對手的競爭中一直處於被動地位，行動遲緩。

In June of 2024, Apple told the world, hey, we're going to take this new large language model concept.

2024 年 6 月，蘋果公司告訴全世界，嘿，我們將採用這種全新的大型語言模型概念。

We're going to make Siri more like chat GPT.

This year, back just earlier this spring, Apple said, oops, we aren't able to do this.

今年春天早些時候，蘋果公司說，哎呀，我們不能這樣做。

The product is not where we thought it would be at this point.

產品目前還沒有達到我們的預期。

So this was kind of a huge miss, especially as Apple was already perceived to be behind in artificial intelligence.

是以，這是一個巨大的失誤，尤其是蘋果已經被認為在人工智能領域落後了。

那麼，是什麼拖了蘋果公司的後腿？

Experts say missed opportunities in model development and internal fragmentation.

專家表示，在模型開發和內部分裂方面錯失良機。

They're not developing the models in-house and implementing them in-house.

他們沒有在內部開發模型，也沒有在內部實施。

They're relying on third parties and then following the trend.

I think by relying on open AI or increasingly other players to be able to do that, they lose that control.

我認為，如果依靠開放式人工智能或越來越多的其他玩家來做到這一點，他們就失去了控制權。

They lose the ability to implement the guardrails, the controls necessary to be able to effectively build the correct solution and implement it most effectively and most user friendly to ensure that it actually aligns with their core principles.

他們失去了實施防護欄和必要控制的能力，從而無法有效地構建正確的解決方案，並以最有效、最方便用戶的方式加以實施，以確保解決方案真正符合他們的核心原則。

Apple restructured its AI teams, delayed key Siri updates and was slower than rivals to hire top LLM talent.

蘋果重組了人工智能團隊，延後了 Siri 的關鍵更新，在招聘頂尖法學碩士人才方面也慢於競爭對手。

In 2024, when Apple finally introduced its generative AI platform, Apple Intelligence, users were hopeful, but the rollout stumbled.

2024 年，當蘋果公司終於推出其生成式人工智能平臺 "蘋果智能 "時，用戶們滿懷希望，但推廣工作卻磕磕絆絆。

Apple Intelligence started rolling out this fall and it had several different kinds of features.

蘋果智能系統從今年秋天開始推出，它有幾種不同的功能。

Apple also did another feature that would summarize your notifications.

蘋果還做了另一項功能，可以彙總你的通知。

And Apple actually kind of got into hot water for this because the artificial intelligence system was summarizing news stories incorrectly.

實際上，蘋果公司也是以惹上了麻煩，因為人工智能系統對新聞報道的總結有誤。

So, for example, the BBC had a big problem with this, where they were pushing out news alerts to people's phones like normal.

例如，英國廣播公司（BBC）在這方面就遇到了很大的問題，他們像往常一樣向人們的手機推送新聞提醒。

But Apple Intelligence was interpreting it incorrectly and giving the wrong answer.

但蘋果智能公司的解釋有誤，給出了錯誤的答案。

Literally fake news pushed to people's screens.

簡直就是把假新聞推送到人們的螢幕上。

That was another stumble out the gate here for the Apple Intelligence system.

這也是蘋果智能系統的另一個絆腳石。

They're supposed to kind of smartly tell you what was going on with all your apps.

它們應該會聰明地告訴你所有應用程序的運行情況。

And it didn't really live up to that promise.

So they had to walk some of those features back.

是以，他們不得不將其中一些功能收回。

The major challenge has been the hallucinations around the news alerts, which have been pretty significant.

主要的挑戰是圍繞新聞警報產生的幻覺，這種幻覺相當嚴重。

The other factor is around differentiation and its inability to keep up, inability to act to the leading edge.

另一個因素是差異化及其無法跟上時代的步伐，無法走在時代的前列。

So when you when you look at Google, Amazon, Microsoft and others who are building in-house, they're developing AI capabilities.

是以，當你看到谷歌、亞馬遜、微軟和其他正在進行內部建設的公司時，他們正在開發人工智能能力。

They're able to bring products to market, test, evaluate them, add to them.

他們能夠將產品推向市場，對其進行測試、評估和改進。

Siri, in its most basic form and functionality, has probably not delivered the virtual assistant experience that your average iPhone user is seeking.

Siri 最基本的形式和功能可能無法為普通 iPhone 用戶帶來他們所追求的虛擬助手體驗。

Apple can still win within the market, but it needs those upgrades to become the virtual assistant that I think we can all dream the dream in thinking about.

蘋果仍能在市場中獲勝，但它需要升級才能成為虛擬助手，我想這是我們每個人都夢寐以求的。

Tim Cook addressed this in Apple's Q2 2025 earnings call.

蒂姆-庫克在蘋果公司 2025 年第二季度財報電話會議上談到了這一點。

With regard to the more personal Siri, as you mentioned, we just need more time to complete the work so they meet our high quality bar.

至於更個性化的 Siri，正如你提到的，我們只是需要更多時間來完成工作，使它們達到我們的高質量標準。

Apple declined CNBC's request for an interview for this story.

蘋果公司拒絕了 CNBC 的採訪要求。

蘋果公司面臨的人工智能挑戰遠不止 Siri。

They're also about infrastructure, from cloud computing to custom chips to model development.

它們還涉及基礎設施，從雲計算到定製芯片，再到模型開發。

Unlike competitors Google and Microsoft, Apple has not built the kind of full-stack AI pipelines or hyperscaled cloud systems needed to support cutting-edge generative AI.

與競爭對手谷歌和微軟不同，蘋果尚未建立支持尖端生成式人工智能所需的全棧人工智能管道或超規模雲系統。

Instead, Apple's leaned into a very different approach.

相反，蘋果公司採用了一種截然不同的方法。

Privacy first, on-device AI and outside partnerships.

隱私第一、設備上的人工智能和外部合作。

With Apple Intelligence, at least Apple's own technology, the goal is to be able to say something to the effect of, hey Siri, find the picture of my wife in a red dress.

有了蘋果智能，至少是蘋果自己的技術，我們的目標是能夠說出 "嘿，Siri，找到我妻子穿紅裙的照片 "這樣的話。

字幕列表影片播放

為什麼蘋果的 Siri 在人工智能時代依然如此糟糕？ (Why Apple’s Siri Is Still So Bad In The Age Of AI)

literally

context

opportunity

strategy