原文標題：
Generative AI
Intelligence services
Investors are going nuts for ChatGPT-ish artificial intelligence
生成式人工智能
智能服務(wù)
投資者熱衷于投資ChatGPT式人工智能

Even Elon Musk wants his own AI chatbot
甚至連埃隆·馬斯克都想要自己的AI聊天機器人

[Paragraph 1]
SINCE CHATGPT’S launch in November, a mini-industry has defied the broader slump in tech.
自去年11月ChatGPT推出以來，一個小型產(chǎn)業(yè)在科技普遍低迷的大環(huán)境下異軍突起。

Not a week goes by without someone unveiling a “generative” artificial intelligence (AI) based on “foundation” models—the vast and complex algorithms that give ChatGPT and other AIs like it their wits.
每周都有人推出基于“基礎(chǔ)”模型的“生成式”人工智能（AI）——這些龐大而復雜的算法為ChatGPT等AI賦予了智慧。

On February 24th Meta, Facebook’s parent company, released a model called LLaMA.
2月24日，臉書的母公司Meta發(fā)布了一個名為LLaMA的語言模型工具。

Elon Musk, boss of Tesla and Twitter, reportedly wants to create an AI that would be less “woke” than ChatGPT.
據(jù)報道，特斯拉和推特的老板埃隆·馬斯克想要創(chuàng)建一個不需要像ChatGPT那樣“喚醒才能用”的AI。

One catalogue, maintained by Ben Tossell, a British entrepreneur, has just grown to include, among others, Isaac Editor (which helps students write essays) and Ask Seneca (which answers questions based on the stoic philosopher’s writings).
英國企業(yè)家本·托塞爾維護了一個目錄，目錄剛剛新增了眾多生成式AI服務(wù)，包括Isaac Editor（幫助學生寫論文）和Ask Seneca（基于斯多葛派哲學家的著作回答問題）。

ChatGPT may be much talked about and, with over 100m users, talked to.
Yet Mr Tossell’s database hints that the real action in generative AI is in all manner of less chatty services enabled by foundation models.
雖然ChatGPT備受關(guān)注，并且有超過1億用戶與之交流。但托塞爾先生的數(shù)據(jù)庫表明，生成式AI真正應用是各種基于基礎(chǔ)模型的非問答服務(wù)。

[Paragraph 2]
Each model is trained on reams of text, images, sound files or other data. This allows them to interpret instructions in natural language and respond with text, art or music.
每個模型都經(jīng)過了大量文本、圖像、聲音文件或其他數(shù)據(jù)的訓練。這使它們能夠理解自然語言的指令，并能以文本、藝術(shù)或音樂作出響應。

Though such systems have been around for some time, it took a consumer-facing service such as ChatGPT to capture the world’s—and investors’—imagination.
盡管這樣的系統(tǒng)已經(jīng)存在一段時間，但ChatGPT等面向消費者服務(wù)的出現(xiàn)，才引起了世界投資者們的無限遐想。

As Mike Volpi of Index Ventures, a venture-capital (VC) firm, says, this happened just as his fellow tech backers, burned by the cryptocurrency crash and the empty metaverse, were on the lookout for the next big thing.
風險投資公司Index Ventures的邁克·沃爾皮稱，正當他的技術(shù)人員因加密貨幣崩盤和空洞的元宇宙而煩惱時，ChatGPT問世了，此時他們正在尋找下一個風口。

In addition, even more than web browsers and smartphones, foundation models make it easy to build new services and applications on top of them.
此外，基礎(chǔ)模型更容易在網(wǎng)頁瀏覽器和智能手機上構(gòu)建新服務(wù)和應用程序。

“You can open your laptop, get an account and start interacting with the model,” says Steve Loughlin of Accel, another VC firm.
另一家風險投資公司Accel的史蒂夫·洛夫林說：“你只要打開筆記本電腦，獲得一個帳戶，即可開始與模型互動。”

[Paragraph 3]
Money is flooding into the business.
資金正涌入這個行業(yè)。

In January it was reported that Microsoft poured $10bn in OpenAI, the startup behind ChatGPT, on top of an earlier investment of $1bn.
據(jù)報道，微軟在1月份向ChatGPT背后的初創(chuàng)公司OpenAI注資100億美元，此前曾投資10億美元。

Pete Flint of NfX, another VC firm, now counts more than 500 generative-AI startups.
根據(jù)風投公司NfX的皮特·弗林特，現(xiàn)在生成式人工智能初創(chuàng)公司已有500多家。

They have so far collectively raised more than $11bn—and that is excluding OpenAI (see chart). Mr Volpi talks of a “Cambrian explosion”.
它們迄今已經(jīng)共籌集了110多億美元——這還不包括OpenAI（見圖表）。沃爾皮稱之為“寒武紀大爆發(fā)”。

[Paragraph 4]
So which generative-AI platforms will make the big bucks? For now, this is the subject of head-scratching in tech circles.
那么，哪些生成式AI平臺將創(chuàng)造巨額利潤呢？目前科技圈尚未有定論。

“It’s just not clear if there will be a long-term, winner-take-all dynamic in generative AI,” wrote Martin Casado and colleagues at Andreessen Horowitz, one more VC firm, in a recent blog post.
VC公司Andreessen Horowitz的馬丁·卡薩多及其同事在最近的一篇博客文章中寫道：“目前在生成式AI中，還不清楚是否會出現(xiàn)長期贏家通吃的局面?！?br>
Many startups offer me-too ideas, often more feature than product.
許多初創(chuàng)公司提供了“我也有”的想法，但更多的是功能而不是產(chǎn)品。

Even the resource-intensive foundation models may end up as a low-margin commodity: though proprietary ones such as OpenAI’s GPT-3.5 are ahead, open-source alternatives aren’t far behind.
即便是資源密集型的基礎(chǔ)模型，也可能最終成為低利潤商品：雖然像OpenAI的GPT-3.5這樣的專有模型處于領(lǐng)先地位，但開源替代品也并不遙遠。

[Paragraph 5]
Generative AI is also tiptoeing into a legal minefield. The models often get things wrong. And they can go off the rails.
生成式人工智能也悄悄走進了法律雷區(qū)。這些模型經(jīng)常會出錯，也可能失控。

Sydney, the chatbot Microsoft is developing for its Bing search engine using OpenAI’s tech, has insulted a few users and professed its love to at least one (it has since been reined in).
微軟正在使用OpenAI技術(shù)為其Bing搜索引擎開發(fā)聊天機器人Sydney，該機器人曾經(jīng)侮辱了一些用戶，且至少給一位用戶表達了愛意（但現(xiàn)在已被控制）。

AI platforms may not enjoy the legal protection from liability that shields social media.
與社交平臺不同的是，人工智能平臺可能沒有法律責任豁免權(quán)。

Copyright holders of web-based content on which existing models are being trained without asking permission or paying compensation are up in arms.
在未經(jīng)請求許可或支付補償?shù)那闆r下，網(wǎng)絡(luò)內(nèi)容的版權(quán)所有者對現(xiàn)有模型的訓練非常不滿。

Getty Images, a repository of photographs, and individual artists have filed lawsuits against AI art-generators such as Stable Diffusion.
攝影作品庫Getty Images和個別藝術(shù)家已經(jīng)對諸如Stable Diffusion等AI藝術(shù)生成平臺提起了訴訟。

Stable Diffusion says, “We take these matters seriously. We are reviewing the documents and will respond accordingly.”
Stable Diffusion表示：“我們高度重視這些問題。我們正在審查文件，并將作出相應的回應?！?br>
News outlets fear text-gobbling AIs, too.

新聞媒體也擔心“洗稿”的AI技術(shù)。

[Paragraph 6]
OpenAI is already downplaying the launch later this year of GPT-4, the highly anticipated update to its foundation model.
OpenAI已經(jīng)淡化了備受期待的GPT-4基礎(chǔ)模型的發(fā)布計劃，原計劃是今年年底發(fā)布。

It won’t temper VC types’ appetite for generative AI. For more risk-averse investors, the safest bet at the moment is on the providers of the ample processing power needed to train and run foundation models.
這不會減弱風投們對生成式AI的興趣。對于風險規(guī)避意識較強的投資者而言，目前最安全的選擇是投資于有大量數(shù)據(jù)處理能力的供應商（這樣他們就可以培訓和運行基礎(chǔ)模型）。

The share price of Nvidia, which designs chips useful for AI applications, is up by 60% so far this year.
英偉達今年的股價迄今已經(jīng)上漲了60％，該公司設(shè)計的芯片在AI應用中非常領(lǐng)先。

Cloud-computing services and data-centre landlords are rubbing their hands, too.
云計算服務(wù)和數(shù)據(jù)中心業(yè)主也暗自竊喜。

Whichever AI platform comes out top, you can’t go wrong selling picks and shovels in a gold rush.
無論哪個AI平臺勝出，淘金熱中選擇賣鐵鍬賣鏟子的生意永遠不會錯。

（恭喜讀完，本篇英語詞匯量742左右）
原文出自：2023年3月4日《The Economist》Business版塊。

精讀筆記來源于：自由英語之路

本文翻譯整理: Irene本文編輯校對: Irene
僅供個人英語學習交流使用。

【補充資料】（來自于網(wǎng)絡(luò)）
"Foundation" models 是指一類經(jīng)過大規(guī)模預訓練的深度學習模型，這些模型通常具有大量的參數(shù)和復雜的網(wǎng)絡(luò)結(jié)構(gòu)，可以用于各種自然語言處理（NLP）任務(wù)，例如文本分類、文本生成、問答系統(tǒng)、機器翻譯等。這些模型通常使用無監(jiān)督的預訓練方法（如BERT、GPT等），在大規(guī)模的語料庫上進行預訓練，然后通過微調(diào)來完成特定任務(wù)。這些模型通常被稱為 "foundation" models，因為它們被認為是構(gòu)建更高級的自然語言處理模型和應用程序的基礎(chǔ)。由于這些模型經(jīng)過大量的預訓練，它們已經(jīng)具有了對自然語言的基本理解，可以顯著提高各種NLP任務(wù)的效果。

寒武紀爆發(fā)（Cambrian explosion）是指大約五億年前（寒武紀早期）生物演化歷史上迅速出現(xiàn)大量新的多細胞生物，它們的體型、生態(tài)類型和復雜性遠超過以前存在的單細胞生物和簡單多細胞生物。在大約2500萬年的時間里，各種不同的海洋生物出現(xiàn)了，包括腕足動物、軟體動物、節(jié)肢動物和脊索動物的祖先等等。這個時期的生物多樣性和數(shù)量之大，讓寒武紀爆發(fā)成為了生命演化史上的一個重要事件。

OpenAI's GPT-3.5是一種基于人工智能技術(shù)的語言模型，是GPT-3（Generative Pre-trained Transformer 3）的進一步發(fā)展。GPT-3是一個由OpenAI開發(fā)的自然語言處理模型，能夠理解和生成人類語言。與GPT-3相比，GPT-3.5包含更多的訓練數(shù)據(jù)和更大的模型，因此具有更高的準確性和更強的語言生成能力。

【重點句子】（3?個）
So which generative-AI platforms will make the big bucks? For now, this is the subject of head-scratching in tech circles.
那么，哪些生成式AI平臺將創(chuàng)造巨額利潤呢？目前科技圈尚未有定論。

Copyright holders of web-based content on which existing models are being trained without asking permission or paying compensation are up in arms.
在未經(jīng)請求許可或支付補償?shù)那闆r下，網(wǎng)絡(luò)內(nèi)容的版權(quán)所有者對現(xiàn)有模型的訓練非常不滿。

Whichever AI platform comes out top, you can’t go wrong selling picks and shovels in a gold rush.
無論哪個AI平臺勝出，淘金熱中選擇賣鐵鍬賣鏟子的生意永遠不會錯。