《ChatGPT》全文逐句翻譯：

第1&2段：

“The age of AI is upon us,” Microsoft’s chief executive Satya Nadella declared last week.

微軟的CEO Satya Nadella上周宣布：“AI的時(shí)代即將來臨”。

The giant software company is certainly backing its grand talk with hard cash and computing power.

這個(gè)巨大的軟件公司肯定會(huì)用現(xiàn)金和算力來支持它所放出的豪言壯語(指前文CEO說的話)。

Last month, Microsoft announced it was investing a further $10bn in OpenAI, the San Francisco-based research company that has electrified the digital world with the launch of its ChatGPT language-generation model.

上個(gè)月，微軟宣布在OpenAI（OpenAI是一家總部位于舊金山的研究公司）再投資100億美元，OpenAI推出的ChatGPT語言生成模型讓數(shù)字世界為之興奮。

第3段：

Within a week of opening up ChatGPT to the public, 1mn users were experimenting with the super-sophisticated chatbot to write reams of highly plausible text at astonishing speed.

在ChatGPT開放給公眾的一周內(nèi)，一百萬用戶嘗試使用這個(gè)超級(jí)先進(jìn)的聊天機(jī)器人，用它驚人的速度寫出大量高度可信的文本。

第4段：

Such “l(fā)arge language models”, as they are called, are not new, but never before have they been made so widely accessible.

這種所謂的“大型語言模型”并不新鮮，但他們以前從未如此廣泛地使用。

（注：never before have they…倒裝，語序還原為they have never before…)

Even the British finance minister Jeremy Hunt (or his speechwriter) employed ChatGPT to compose the introduction to his recent speech on innovation

甚至英國的財(cái)政部長Jeremy Hunt（或他的演講稿撰寫人）也用ChatGPT來撰寫他最近關(guān)于創(chuàng)新的演講的引入語。

第5段：

Despite the broader tech market downturn, venture capital firms have been frenziedly pouring billions of dollars into start-ups to exploit the commercial opportunities of these “generative” AI models.

盡管科技市場廣泛衰退，風(fēng)投公司瘋狂地往初創(chuàng)公司里注入數(shù)十億美元的資金，以此來利用這些“生成式”AI模型所帶來的商業(yè)機(jī)會(huì)。

venture capital（VC）：風(fēng)險(xiǎn)投資

start-up：初創(chuàng)公司

第6段：

Trained on huge amounts of content drawn from the internet, these models can also be used to generate computer code, images, audio and video.

利用從互聯(lián)網(wǎng)中獲取的大量內(nèi)容來訓(xùn)練，這些模型同樣能被用來生成代碼、圖像、音頻和視頻。

The next big thing in AI this year may well be the emergence of multi-modal models that mix several of these elements.

今年AI領(lǐng)域的下一件大事可能是多模態(tài)模型的出現(xiàn)，這種模型可以融合多個(gè)（之前提及的）這些元素（指代碼、圖像、音頻和視頻）。

“The quality and capabilities of generative models for imagery, video, and audio has shown truly stunning and extraordinary advances in 2022,” wrote Jeff Dean, one of Google’s most senior AI researchers.

來自谷歌的資深A(yù)I研究員Jeff Dean寫道：“在2022年，生成式模型在圖像、視頻和音頻方面（應(yīng)用）的質(zhì)量和能力，展現(xiàn)了真正驚人和非凡的進(jìn)步”。

第7段：

Tech evangelists argue that the widespread adoption of generative AI opens up a new era in technology, just as the emergence of personal computers, smartphones and social media gave birth to seemingly limitless business opportunities in earlier times.

技術(shù)傳播者認(rèn)為，生成式AI的廣泛采用開啟了技術(shù)新時(shí)代，就像個(gè)人電腦、智能手機(jī)和社交媒體的出現(xiàn) 在早期催生了看似無限的商機(jī)一樣。

evangelists：巡回布道者、傳播者

Generative AI may yet transform the creative industries, video games production and enterprise software, for example.

生成式AI可能最終會(huì)變革創(chuàng)造性產(chǎn)業(yè)，比如電子游戲制作和企業(yè)軟件。

yet：終歸

One of the open questions surrounding its use is whether it will strengthen the power of the tech incumbents, including Microsoft and Google, or stimulate fresh competition from insurgent start-ups, such as Cohere, Stability AI and You.com.

關(guān)于它（生成式AI）的使用的一個(gè)未有定論的問題是：它會(huì)加強(qiáng)現(xiàn)任科技公司的力量（包括微軟和谷歌），還是刺激來自新興初創(chuàng)公司的新競爭（比如Cohere、Stability AI和You.com）。

open question：未有定論的問題

incumbent：在職者、現(xiàn)任者

insurgent：起義的

第8段：

But the widespread adoption of generative AI is also triggering concerns about the theft of copyright, the spread of disinformation and deepfakes and the debasement of knowledge on the world wide web.

但生成式AI的廣泛采用同樣引發(fā)了一系列擔(dān)憂，即版權(quán)盜竊、虛假信息的擴(kuò)散、AI換臉和萬維網(wǎng)知識(shí)的貶值。

deepfake：深度偽造

=deep learning+fake，專指基于AI的人體圖像合成技術(shù)的應(yīng)用，可簡單理解為AI換臉

world wide web=WWW萬維網(wǎng)

Generative AI systems are trained to produce the most statistically probable answer, not the most accurate, and can “hallucinate” facts.

生成式AI系統(tǒng)被訓(xùn)練來產(chǎn)生統(tǒng)計(jì)學(xué)上最可能的答案，而不是最精確的答案，這會(huì)“用幻覺影響”事實(shí)。

hallucinate：vi.產(chǎn)生幻覺,vt.用幻覺影響

ChatGPT has been described as “the greatest bullshitter ever” by Arvind Narayanan, a computer science professor at Princeton University.

普林斯頓大學(xué)的一位計(jì)算機(jī)科學(xué)教授Arvind Narayanan將ChatGPT描述為“史上最胡說八道的（物品）”。

第9段：

The digital gold rush around generative AI may also be distracting us from more profound and productive uses of other machine learning models, according to one tech executive.

一位科技高管說，關(guān)于生成式AI的數(shù)字淘金熱，可能會(huì)讓我們從 其他機(jī)器學(xué)習(xí)模型更深刻和高產(chǎn)的使用中 轉(zhuǎn)移注意力。

While everyone is fixated on the industry’s “shiny new toy”, researchers have been using machine learning systems to diagnose cancers, model protein structures, optimise electricity grids and design components for heart valves, aero engines or wind turbines.

當(dāng)所有人都專注于這個(gè)行業(yè)的“閃亮新玩具”時(shí)，研究員們一直在使用機(jī)器學(xué)習(xí)系統(tǒng)來診斷癌癥，給蛋白質(zhì)結(jié)構(gòu)建模，優(yōu)化電網(wǎng)和設(shè)計(jì)心臟瓣膜、飛機(jī)引擎或風(fēng)力渦輪發(fā)電機(jī)的組件。

be fixated on sth. ：專注于…

electricity grid：電網(wǎng)

第10段：

The age of AI may well be upon us, but not quite in the way we — and our digital assistants — are currently all chatting about.

AI的時(shí)代即將來臨，但不是以我們和電子助手正在討論的這種形式。

（為便于理解，該句等同于：

The age of AI may well be upon us, but not quite in the way(that) we and our digital assistants are currently all chatting about.）

文章整體結(jié)構(gòu)如下：