Name: Anthropic 與 OpenAI 大對決！AI 安全如何成為致勝關鍵？ (Anthropic Vs. OpenAI: How Safety Became The Advantage In AI)
Uploaded: 2026-03-21T18:23:52.000Z
Duration: 22 min 39 s

反身代名詞

The AI lab that avoided the hype has suddenly found itself at the center of it.

避開炒作的AI實驗室，突然發現自己身處風口浪尖。

Anthropic is one of the key AI names eyeing a potential IPO this year.

Anthropic 是今年有望進行 IPO 的關鍵 AI 公司之一。

And it's looking to outmaneuver big spending rivals by showing you don't need the most chips to win.

它們想透過證明不一定要有最多的晶片才能獲勝，來智取花費鉅額的競爭對手。

Anthropic story began five years ago with the defection.

Anthropic 的故事始於五年前的一次出走。

A core group of senior OpenAI researchers who helped build the systems that became ChatGPT, then left to create a lab where Safety wasn't an afterthought.

一群核心的 OpenAI 資深研究員，他們曾協助打造出 ChatGPT 的系統，之後離開，創立了一個將安全擺在首位的實驗室。

You had this really clear vision of the company that we wanted to build, a technology company, you know, developing transformative AI at the frontier, but

我們有一個非常清晰的願景，想要建立一家科技公司，你知道的，在最前沿開發變革性的 AI，但是

with this very intense sort of focus on making the system safe and reliable.

No viral consumer product, no flashy demos, just quiet progress building for enterprises.

沒有病毒式的消費產品，沒有炫目的演示，只有為企業默默進步的建設。

The Anthropic partnership is critical for us and very, very strategic.

Anthropic 的合作對我們來說至關重要，而且非常、非常具有策略性。

Anthropic has always had a fraction of what our competitors have had in terms of compute and capital, and

在運算和資金方面，Anthropic 始終只有競爭對手的一小部分，而且

yet we've pretty consistently had the sort most powerful, most performant models for the majority of the past several years.

但我們在過去幾年裡， pretty consistently 都有最強、表現最好的模型。

It is a central player in the most expensive build out in a generation.

這是這一代最昂貴建設計畫中的一個核心參與者。

A major move in America's AI arms race today with Anthropic, it launched a $50 billion national infrastructure buildout plan.

今天，Anthropic 在美國的 AI 軍備競賽中有了重大舉動，他們啟動了一項 500 億美元的全國基礎設施建設計畫。

Amazon's $11 billion AI data center in Indiana, one of the largest in the world, is now live.

亞馬遜在印第安納州耗資 110 億美元的 AI 資料中心，是全球最大的資料中心之一，現已上線。

They saw the potential of that chip and has really kind of co invested with us to jointly make that better.

他們看到了那款晶片的潛力，並與我們共同投資，一起讓它變得更好。

But at the same time, Anthropic is warning the next milestone in artificial intelligence could come faster than anyone's ready for.

但同時，Anthropic 也警告說，人工智慧的下一個里程碑可能會比任何人準備好的來得更快。

At some point we're going to get to AI systems that are better than almost all humans at almost all tasks.

總有一天，我們會達到 AI 系統在幾乎所有任務上都比幾乎所有人類做得更好的程度。

The term I've used for it in an essay I recently wrote is a country of geniuses in a data center.

我在最近寫的一篇文章中用來形容它的詞是「資料中心裡的國家級天才」。

This is the paradox of Anthropic, the lab that made Safety its brand and then had to scale fast enough to survive the moment that it warned the world about.

這就是 Anthropic 的悖論，這個將安全作為其品牌的實驗室，卻必須快速擴大規模，以度過它所警告世界的那個時刻。

In late 2020, siblings Dario and Daniela Amade, along with a core group of senior leaders, left OpenAI convinced that the next era of Frontier AI would be decided as much by restraint as by raw capability.

在 2020 年底，Dario 和 Daniela Amade 這對兄妹，以及一群核心高階主管，離開了 OpenAI，他們相信下一世代的 Frontier AI 將如同其原始能力一樣，同樣受到節制的影響。

We really just felt more like we were running towards something than running away from something.

我們真的感覺像是朝著某個目標前進，而不是逃離什麼。

The co founders, we had all sort of known each other for kind of various amounts of time and worked together over a lot longer than just our time at OpenAI.

共同創辦人們，我們都認識彼此一段時間了，而且一起共事的時間遠超過我們在 OpenAI 的時光。

We really had this kind of long history together, just as a group of people with this sort of very like minded set of values.

我們這個團隊，有著長久的合作歷史，而且價值觀非常相近。

Anthropic deserves huge credit for making the practice of AI capability and AI risk evaluation systematic and sustainable.

Anthropic 應獲得極大的肯定，因為他們讓 AI 能力和 AI 風險評估的實踐變得系統化且永續。

And inside that decision was a bet that safety wouldn't be a tax on progress, it would be part of the product.

而這個決策背後，是相信安全不會是進步的阻礙，而是產品的一部分。

There's this sort of misbelief that like safety and reliability kind of come into tension with the business side.

有一種普遍的誤解，認為安全和可靠性與商業面會產生衝突。

At the time it sounded very novel, this belief that those two things actually were correlated and that they went together, they were specific about what that meant.

當時聽起來非常新穎，這種相信這兩件事其實是相關聯的，而且是相輔相成的，他們對此有明確的定義。

The guardrails, the controls that keep a model from going off script, generating toxic content or being manipulated into bypassing its own rules, had to be engineered into the system, not bolted on after the fact.

那些防止模型脫離劇本、產生不良內容或被操縱以繞過自身規則的防護措施和控制機制，必須內建到系統中，而不是事後才加上去。

Anthropic's model family is called Claude.

Anthropic 的模型系列叫做 Claude。

It's similar to OpenAI's ChatGPT or Google's Gemini, but

它跟 OpenAI 的 ChatGPT 或 Google 的 Gemini 差不多，但是

trained differently with different data, different safety approaches and different goals.

訓練方式、使用的資料、安全策略和目標都不同。

And while OpenAI turned ChatGPT into a household name by giving it away for free, Anthropic took Claude in a different direction, straight to businesses.

OpenAI 讓 ChatGPT 免費提供，一炮而紅，而 Anthropic 則把 Claude 導向企業市場。

The kind of higher bar for safety in businesses and in enterprises was really well suited to where Anthropic excelled from sort of day one.

企業和公司對安全的更高要求，正好符合 Anthropic 從一開始就擅長的領域。

Inside OpenAI, Dario Amadei wasn't just another fast moving builder.

在 OpenAI 裡，Dario Amadei 不僅僅是個快速行動的開發者。

He was a technical leader who wrote long memos about safety when most investors were focused on speed.

他是一位技術領袖，在大多數投資者只關注速度時，他卻寫了關於安全的長篇備忘錄。

Matt Murphy, a partner at Menlo Ventures who led one of Anthropic's early funding rounds, saw something different.

Menlo Ventures 的合夥人 Matt Murphy，他主導了 Anthropic 早期的融資輪之一，看到了不一樣的地方。

And Dario, from the beginning, he's just got amazing energy and passion and kind of this boyish charm about him.

而 Dario 從一開始就展現出驚人的活力、熱情，還有那種男孩般的魅力。

Description of how to execute, what they were doing distinctively and just him as a leader all just kind of came through extremely clear.

關於如何執行的描述、他們獨特之處，以及他身為領導者的特質，都非常清晰地呈現出來。

And that's actually in our process of thinking about investing, where things really accelerated, where like this person is going to help build an amazing company.

而這其實也是我們在思考投資時，事情真正加速的過程，覺得這個人會幫助建立一家很棒的公司。

What Anthropic was proposing wasn't just a better model.

Anthropic 提出的不只是一個更好的模型。

It was a different theory of how to build them.

而是關於如何建構它們的另一種理論。

The founding crew avoided flashy announcements, no grandstanding.

創始團隊避免了浮誇的宣傳，沒有嘩眾取寵。

They pitched themselves as the steady alternative.

他們把自己定位為穩健的替代方案。

Less noise, more control and a simple premise.

減少噪音，增加控制，以及一個簡單的前提。

You need a set of rules and guardrails baked in from the start, before frontier AI gets too powerful to steer.

你需要一套從一開始就內建的規則和安全措施，在尖端 AI 強大到難以駕馭之前。

November 2022, OpenAI sounded the starting gun.

A new artificial intelligence program called ChatGPT made its debut online.

一款名為 ChatGPT 的新人工智慧程式在網路上亮相。

This project from the Open Air Research Lab can write essays and carry on convincing written conversation ChatGPT, a free chatbot anyone could talk to, went viral instantly.

這個來自 Open Air Research Lab 的專案可以寫文章，並進行令人信服的書面對話。ChatGPT，一個任何人都可以使用的免費哈拉機器人，瞬間爆紅。

Google 趕緊行動，Microsoft 全力投入。

The entire tech industry lurched into motion.

But Anthropic didn't sprint the real money.

但 Anthropic 並沒有衝刺真正的金流。

The sticky, reliable long term money wasn't in viral moments.

穩定、可靠的長期金流不在於爆紅時刻。

Anthropic just announcing a major new international expansion.

Anthropic 剛剛宣佈一項重大的新國際擴張計畫。

As the AI startup says enterprise demand surges overseas developer tools APIs sold into companies where reliability, security and compliance are mission critical.

這家 AI 新創公司表示，企業對海外開發工具 API 的需求激增，這些工具賣給了可靠性、安全性與合規性至關重要的公司。

How did you know from the beginning that enterprise was the market to go after?

你從一開始就知道企業市場是值得爭取的目標，是怎麼知道的？

I wouldn't necessarily say that we knew for sure.

我不能說我們從一開始就確定。

I think there were sort of two parts to it.

The first is Anthropic as an organization is well suited to be a B2B company.

首先，Anthropic 這個組織本身就很適合做 B2B 公司。

We really care about things like, you know, reliability and security and safety.

我們真的很在意像是可靠性、安全性跟保障這些事情。

Part of why we founded Anthropic was this belief in, in capturing the upside and sort of mitigating the downside.

我們創立 Anthropic 的部分原因，就是相信要抓住好的機會，同時也要降低風險。

And it turns out that's something businesses really value.

結果發現這正是企業非常重視的。

I think the second was much more kind of an economic theory view.

我覺得第二部分比較像是從經濟學理論的角度來看。

Even at the time, in late 2020 or early 2020, one saw this kind of potential future where Claude was just capable of many of the very kind of high intelligence tasks that humans do in the workplace.

即使在當時，也就是 2020 年底或 2020 年初，就已經可以看到 Claude 在很多高智慧的工作任務上，能做得跟人類一樣好，這就是未來的潛力。

And we thought that's a pretty big market.

我們覺得這是一個很大的市場。

Bessemer Ventures Samir Jalakia says the bet made sense for one simple reason.

Bessemer Ventures 的 Samir Jalakia 表示，這個投資有其簡單的道理。

Enterprise customers don't churn the way consumers do.

企業客戶不像一般消費者那樣容易流失。

We really loved the focus that they had and candidly, the focus of Anthropic on safety and trust, we knew was going to play really well with the enterprise buyer.

我們真的很喜歡他們專注的態度，坦白說，Anthropic 對安全和信任的重視，我們知道這對企業買家來說會非常有吸引力。

Anthropic's early traction showed up in one place over and over.

Anthropic 早期的發展軌跡一再地在一個地方顯現。

They really found a home in the developer community.

他們在開發者社群裡找到了歸屬。

So lots of enterprises are adopting Claude to support their coding activities.

所以很多企業正在採用 Claude 來支援他們的程式開發活動。

Anthropic recently hired Instagram co founder Mike Krieger as chief product officer.

Anthropic 最近聘請了 Instagram 共同創辦人 Mike Krieger 擔任產品長。

We're moving from asking questions of these AI models to doing work with these AI models.

我們正從「向 AI 模型提問」轉變為「利用 AI 模型來做事」。

So for example, we're looking at things like how do we unlock discounted cash flow analysis or due diligence on a client, or analyzing an entire market and then preparing a presentation at the end of it all.

所以舉例來說，我們正在研究如何解鎖現金流折現分析、客戶盡職調查，或是分析整個市場，然後最後準備一份簡報。

These really like core chunks of work that otherwise might have taken hours we can now do.

這些原本可能需要數小時才能完成的核心工作，現在我們都能辦到。

And now that enterprise bet is going Global.

而現在企業的佈局正走向全球。

Because we work so much in the enterprise side, I think we're a good barometer, maybe a pure barometer than the others, which kind of filter through consumers, which kind of have their habits and their use cases.

因為我們在企業端投入很多，我覺得我們是一個很好的指標，或許比其他過濾消費者的公司更純粹，因為消費者的習慣和使用情境比較多元。

It's grown 10x a year every year for the last three years.

過去三年來，我們的營收每年都成長了 10 倍。

It's going to land somewhere between 8 and 10 at the end of this year.

今年年底，大概會在8到10之間。

But the technology is driving there and the economic value is coming with it.

但技術正在驅動，經濟價值也隨之而來。

Anthropic says its business customer base has grown from under 1,000 to more than 300,000 in two years.

Anthropic 表示，他們的企業客戶群在兩年內從不到1,000家增長到超過300,000家。

And that nearly 80% of cloud activity actually comes from outside the United States.

而且近80%的雲端活動其實來自美國以外。

The customer list reads like a who's who of global enterprise.

這些客戶名單簡直就是全球企業的縮影。

字幕列表影片播放

Anthropic 與 OpenAI 大對決！AI 安全如何成為致勝關鍵？ (Anthropic Vs. OpenAI: How Safety Became The Advantage In AI)

sort

potential

critical

access