-
市場
-
產品
-
資訊
-
Moo社區
-
課堂
-
查看更多
-
功能介紹
-
費用費用透明,無最低余額限制
投資選擇、功能介紹、費用相關信息由Moomoo Financial Inc.提供
- English
- 中文繁體
- 中文简体
- 深色
- 淺色
Meta's Massive $1B Bet On Nvidia H100 GPU Cluster Will Dwarf Every Other Company's Target, But Zuckerberg Reveals Original Goal Was This, Not AI
Meta's Massive $1B Bet On Nvidia H100 GPU Cluster Will Dwarf Every Other Company's Target, But Zuckerberg Reveals Original Goal Was This, Not AI
$Meta Platforms (META.US)$ has set a goal of buying 350,000 $NVIDIA (NVDA.US)$ H100 GPUs by 2024 to power its AI ambitions, but that was not Mark Zuckerberg's original intention.
What Happened: Zuckerberg revealed that Meta's massive $1 billion bet on Nvidia's H100 GPUs was originally meant to be used for a different product, not AI.
In a recent episode on Dwarkesh Patel's podcast, Zuckerberg spilled the beans on something previously unknown – those 350,000 Nvidia H100 GPUs were to help make Instagram Reels better, not to help make Meta a leading AI company.
But how did Zuckerberg even know Meta would need $1 billion worth of GPUs? Zuckerberg explains that Instagram Reels is what made the company take the plunge.
"It was because we were working on Reels."
He explains that Meta needed more GPUs to "train the models" behind Reels recommendations.
"We made this big push to start recommending what we call unconnected content," he added, explaining that this resulted in the number of recommendable Reels rising from thousands to "hundreds of millions."
Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox.
Zuckerberg wanted to make Instagram Reels as addictive as TikTok, but Meta's infrastructure was proving to be the bottleneck.
"We have to make sure that's never an issue again, so let's order enough GPUs to do what we need to do on Reels." That's where the 350,000 Nvidia H100 GPUs come in.
Training large AI models like Llama was an unintended but fruitful coincidence for Meta, and Zuckerberg thinks it "ended up being a very good decision, in retrospect."
Why It Matters: Meta says that its latest AI model, Llama 3, has outperformed Alphabet Inc.'s Google Gemini, Anthropic's Claude 3, and Mistral AI's model.
The Llama 3 model, which comes in two sizes with 8B and 70B parameters, has shown significant improvements in text-based responses, including answering prompts and refusing fewer times.
This comes after Anthropic announced that Claude 3 outperformed rivals like Microsoft Corp.-backed OpenAI's GPT-4.
Interestingly, Meta did not mention GPT-4 in its comparison, but OpenAI is expected to release a "significantly better" GPT-5 model later this year.
$Meta Platforms (META.US)$ has set a goal of buying 350,000 $NVIDIA (NVDA.US)$ H100 GPUs by 2024 to power its AI ambitions, but that was not Mark Zuckerberg's original intention.
$Meta Platforms (META.US)$ 已經設定了購買 350,000 的目標 $英偉達 (NVDA.US)$ 到2024年,H100 GPU將爲其AI雄心壯志提供動力,但這並不是馬克·扎克伯格最初的意圖。
What Happened: Zuckerberg revealed that Meta's massive $1 billion bet on Nvidia's H100 GPUs was originally meant to be used for a different product, not AI.
發生了什麼:扎克伯格透露,Meta對英偉達H100 GPU的10億美元巨額押注最初是要用於其他產品,而不是人工智能。
In a recent episode on Dwarkesh Patel's podcast, Zuckerberg spilled the beans on something previously unknown – those 350,000 Nvidia H100 GPUs were to help make Instagram Reels better, not to help make Meta a leading AI company.
在最近一集的德瓦克什·帕特爾播客中,扎克伯格泄露了以前未知的東西——那35萬個英偉達H100顯卡是爲了幫助改善Instagram Reels,而不是幫助Meta成爲領先的人工智能公司。
But how did Zuckerberg even know Meta would need $1 billion worth of GPUs? Zuckerberg explains that Instagram Reels is what made the company take the plunge.
但是扎克伯格怎麼知道 Meta 需要價值 10 億美元的 GPU?扎克伯格解釋說,Instagram Reels是該公司冒險的原因。
"It was because we were working on Reels."
“那是因爲我們在開發 Reels。”
He explains that Meta needed more GPUs to "train the models" behind Reels recommendations.
他解釋說,Meta 需要更多 GPU 來 “訓練 Reels 建議背後的模型”。
"We made this big push to start recommending what we call unconnected content," he added, explaining that this resulted in the number of recommendable Reels rising from thousands to "hundreds of millions."
他補充說:“我們大力推動開始推薦我們所謂的非關聯內容,” 他解釋說,這導致值得推薦的卷軸數量從數千個增加到 “數億”。
Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox.
訂閱 Benzinga 技術趨勢時事通訊,將所有最新的技術發展發送到您的收件箱。
Zuckerberg wanted to make Instagram Reels as addictive as TikTok, but Meta's infrastructure was proving to be the bottleneck.
扎克伯格想讓Instagram Reels像抖音一樣令人上癮,但事實證明,Meta的基礎設施是瓶頸。
"We have to make sure that's never an issue again, so let's order enough GPUs to do what we need to do on Reels." That's where the 350,000 Nvidia H100 GPUs come in.
“我們必須確保這不再是問題,所以讓我們訂購足夠的 GPU 來完成我們需要在 Reels 上做的事情。”這就是 350,000 個 Nvidia H100 GPU 的用武之地。
Training large AI models like Llama was an unintended but fruitful coincidence for Meta, and Zuckerberg thinks it "ended up being a very good decision, in retrospect."
對於 Meta 來說,訓練像 Llama 這樣的大型 AI 模型是意想不到但富有成效的巧合,扎克伯格認爲這 “回想起來,最終是一個非常好的決定”。
Why It Matters: Meta says that its latest AI model, Llama 3, has outperformed Alphabet Inc.'s Google Gemini, Anthropic's Claude 3, and Mistral AI's model.
它爲何重要:Meta表示,其最新的人工智能模型Llama 3的表現已經超過了Alphabet Inc。”是 Google Gemini、Anthropic 的 Claude 3 和 Mistral AI 的模型。
The Llama 3 model, which comes in two sizes with 8B and 70B parameters, has shown significant improvements in text-based responses, including answering prompts and refusing fewer times.
Llama 3模型有兩種尺寸,參數爲8B和70B,在基於文本的響應方面已顯示出顯著的改進,包括回答提示和減少拒絕次數。
This comes after Anthropic announced that Claude 3 outperformed rivals like Microsoft Corp.-backed OpenAI's GPT-4.
在此之前,Anthropic宣佈Claude 3的表現優於微軟公司支持的OpenAI的 GPT-4 等競爭對手。
Interestingly, Meta did not mention GPT-4 in its comparison, but OpenAI is expected to release a "significantly better" GPT-5 model later this year.
有趣的是,Meta 在比較中沒有提到 GPT-4,但預計OpenAI將在今年晚些時候發佈一款 “好得多” 的 GPT-5 機型。
moomoo是Moomoo Technologies Inc.公司提供的金融資訊和交易應用程式。
在美國,moomoo上的投資產品和服務由Moomoo Financial Inc.提供,一家受美國證券交易委員會(SEC)監管的持牌主體。 Moomoo Financial Inc.是金融業監管局(FINRA)和證券投資者保護公司(SIPC)的成員。
在新加坡,moomoo上的投資產品和服務是通過Moomoo Financial Singapore Pte. Ltd.提供,該公司受新加坡金融管理局(MAS)監管(牌照號碼︰CMS101000) ,持有資本市場服務牌照 (CMS) ,持有財務顧問豁免(Exempt Financial Adviser)資質。本內容未經新加坡金融管理局的審查。
在澳大利亞,moomoo上的金融產品和服務是通過Futu Securities (Australia) Ltd提供,該公司是受澳大利亞證券和投資委員會(ASIC)監管的澳大利亞金融服務許可機構(AFSL No. 224663)。請閱讀並理解我們的《金融服務指南》、《條款與條件》、《隱私政策》和其他披露文件,這些文件可在我們的網站 https://www.moomoo.com/au中獲取。
在加拿大,透過moomoo應用程式提供的僅限訂單執行的券商服務由Moomoo Financial Canada Inc.提供,並受加拿大投資監管機構(CIRO)監管。
在馬來西亞,moomoo上的投資產品和服務是透過Moomoo Securities Malaysia Sdn. Bhd. 提供,該公司受馬來西亞證券監督委員會(SC)監管(牌照號碼︰eCMSL/A0397/2024) ,持有資本市場服務牌照 (CMSL) 。本內容未經馬來西亞證券監督委員會的審查。
Moomoo Technologies Inc., Moomoo Financial Inc., Moomoo Financial Singapore Pte. Ltd.,Futu Securities (Australia) Ltd, Moomoo Financial Canada Inc和Moomoo Securities Malaysia Sdn. Bhd., 是關聯公司。
風險及免責聲明
moomoo是Moomoo Technologies Inc.公司提供的金融資訊和交易應用程式。
在美國,moomoo上的投資產品和服務由Moomoo Financial Inc.提供,一家受美國證券交易委員會(SEC)監管的持牌主體。 Moomoo Financial Inc.是金融業監管局(FINRA)和證券投資者保護公司(SIPC)的成員。
在新加坡,moomoo上的投資產品和服務是通過Moomoo Financial Singapore Pte. Ltd.提供,該公司受新加坡金融管理局(MAS)監管(牌照號碼︰CMS101000) ,持有資本市場服務牌照 (CMS) ,持有財務顧問豁免(Exempt Financial Adviser)資質。本內容未經新加坡金融管理局的審查。
在澳大利亞,moomoo上的金融產品和服務是通過Futu Securities (Australia) Ltd提供,該公司是受澳大利亞證券和投資委員會(ASIC)監管的澳大利亞金融服務許可機構(AFSL No. 224663)。請閱讀並理解我們的《金融服務指南》、《條款與條件》、《隱私政策》和其他披露文件,這些文件可在我們的網站 https://www.moomoo.com/au中獲取。
在加拿大,透過moomoo應用程式提供的僅限訂單執行的券商服務由Moomoo Financial Canada Inc.提供,並受加拿大投資監管機構(CIRO)監管。
在馬來西亞,moomoo上的投資產品和服務是透過Moomoo Securities Malaysia Sdn. Bhd. 提供,該公司受馬來西亞證券監督委員會(SC)監管(牌照號碼︰eCMSL/A0397/2024) ,持有資本市場服務牌照 (CMSL) 。本內容未經馬來西亞證券監督委員會的審查。
Moomoo Technologies Inc., Moomoo Financial Inc., Moomoo Financial Singapore Pte. Ltd.,Futu Securities (Australia) Ltd, Moomoo Financial Canada Inc和Moomoo Securities Malaysia Sdn. Bhd., 是關聯公司。
- 分享到weixin
- 分享到qq
- 分享到facebook
- 分享到twitter
- 分享到微博
- 粘贴板
使用瀏覽器的分享功能,分享給你的好友吧