OpenAI发布的GPT-4o有多厉害？

今天凌晨，OpenAI举行春季发布会，推出新模型GPT-4o，主打实时对音频、视觉和文本进行推理，语音响应时间短至232毫秒，与人类反应速度一致。

GPT-4o的“o”代表“omni”，源自拉丁语“omnis”。“omni”常被用作词根，用来表示“全部；到处，遍及；一切；所有”，如：omnipresent 无所不在的，omniscient 无所不知的。

究竟有多厉害？看看“AI家教”demo：

🤔️小作业：

1. Which of the following words can best replace "instantaneously" in the context of the second paragraph?

A) Gradually  B) Eventually

C) Promptly  D) Slowly

A) To detail the technological advancements in GPT-4o

B) To describe the competition between OpenAI and other tech companies

C) To highlight the real-time capabilities of the new AI model

D) To discuss the financial investments made by OpenAI

A) To show that OpenAI's inspiration comes from popular media

B) To compare GPT-4o's capabilities to those seen in the movie

C) To emphasize the emotional intelligence of GPT-4o

D) To explain the technological limitations of current AI models

无注释原文：

OpenAI Launches Voice Assistant Inspired by Hollywood Vision of AI

From: The Wall Street Journal

①OpenAI unveiled a less expensive version of its flagship artificial-intelligence system that includes a new voice assistant to make it easier to use, as it races other tech companies to roll out products and features to attract users.

②The new AI model, dubbed GPT-4o, can better digest images and video in addition to text, and can interact with people by voice in real time, said Mira Murati, OpenAI's chief technology officer, on Monday. People can interrupt the new voice feature while talking to it, unlike current voice assistants, and the model is capable of responding close to instantaneously, the company said.

③OpenAI executives showed in a livestreamed demonstration how the model could analyze code, translate languages between two speakers or guide users through a basic algebra problem written down on a piece of paper, all seemingly in real time.

The launch of GPT-4o reflects how OpenAI and other startups and tech giants are increasingly seeking to expand their pools of users and bring in money for their generative artificial-intelligence technology after pouring enormous sums of money into computing power and energy to develop their systems.

The OpenAI announcement comes a day before the start of Google's annual developer conference on Tuesday, where it is expected to announce its own new products. Google, an AI pioneer, has been vying with OpenAI and its partner and backer, Microsoft, for leadership in generative AI. Microsoft wasn't involved in making GPT-4o.

Chief Executive Sam Altman likened the new product to the kind of AI tools typically seen in movies. In a speech last year, he said he and other OpenAI executives found inspiration in the 2013 film “Her” about a man who falls in love with a voice assistant. He and other OpenAI employees posted references to the movie on X just after Monday's announcement.

“The original ChatGPT showed a hint of what was possible with language interfaces; this new thing feels viscerally different,” Altman wrote Monday on his personal blog.

The new model could also detect a person's emotions in their tone of voice or facial expression, OpenAI said. It also more rapidly switches between different emotional tones from a dramatic voice to a robotic tone to singing. This feature will roll out to users who pay for ChatGPT-Plus, the company's $20-a-month feature in coming weeks.

GPT-4o will also be offered to companies. Murati said the model would be twice as fast and half the cost of its current top-of-the-line offering GPT-4 Turbo. The “o” in GPT-4o stands for “omni,” the company said. People who use the free version of ChatGPT will have access to GPT-4o's image and vision features starting Monday.

OpenAI already offers a feature called “voice mode” that combines three separate models to respond to users in voice, but it could be confused by multiple speakers or background noise. It was also slow. The “chains of models” approach “doesn't cut it when you're trying to deliver at this speed,” said Mark Chen, head of frontiers research at OpenAI, in an interview.

By contrast, GPT-4o was built as a single model trained on text, vision and audio material, and can respond more quickly and accurately to cues. 

OpenAI executives declined to describe what kind of data was used to train this model. They also declined to say whether OpenAI was able to train the model with less computational power. OpenAI is also working on a whole new AI model, called GPT-5, that is expected to make a large jump from current technology.

Murati said Monday that the OpenAI team wasn't inspired so much by the movie “Her” as by human conversation. “When you stop talking, I'm about to jump in. I can kind of read your tone and respond to that. And it's really natural, rich and interactive,” she said.

- ◆ -

注：完整题目见本文开头；中文文本为华尔街日报官方译文，仅供参考

含注释全文：

OpenAI Launches Voice Assistant Inspired by Hollywood Vision of AI

From: The Wall Street Journal

OpenAI unveiled a less expensive version of its flagship artificial-intelligence system that includes a new voice assistant to make it easier to use, as it races other tech companies to roll out products and features to attract users.

OpenAI发布了其旗舰AI系统的一个更便宜的版本，其中内置了一个语音助手以方便用户使用。和其他科技公司一样，OpenAI也在加紧推出产品和功能以吸引用户。

unveil /ʌnˈveɪl/ 1）表示“ 为…揭幕；揭开…上的覆盖物；拉开…的帷幔”，英文解释为“to remove a cover or curtain from a painting, statue, etc. so that it can be seen in public for the first time”举个🌰：She unveiled a plaque to mark the official opening of the hospital. 她主持揭幕式，标志着医院正式启用。

2）表示“（首次）展示，介绍，推出；将…公之于众”，英文解释为“to show or introduce a new plan, product, etc. to the public for the first time”举个🌰：They will be unveiling their new models at the Motor Show. 他们将在汽车大展上首次推出自己的新型汽车。

🎬电影《复仇者联盟1》（The Avengers）中的台词提到：We look beyond the Earth to the greater worlds the Tesseract will unveil. 而我们则放眼地球之外 那些将被宇宙魔方开启的更广博的世界。

🎬电影《超人高校》（Sky High）中的台词提到：We can't unveil the Stronghold Three without Will. 我们不能在威尔（Will）不在场时 向大家介绍斯特朗霍德三侠（Stronghold Three）。

📍词根“-veil”表示"面纱"，前缀“un-”表示“不”。联想宫殿：在宫殿的一个房间，你看到一个被面纱覆盖的新产品即将被“揭露”。近义词：reveal（揭示） 同根词：veil（面纱），常用短语：unveil a plan（公布计划）

flagship /ˈflæɡ.ʃɪp/ 表示“（某机构的）旗舰产品，最重要理念，主建筑物”，英文解释为“the best or most important product, idea, building, etc. that an organization owns or produces”举个🌰：This machine is the flagship in our new range of computers. 这台机器是我们新计算机系列的主打产品。

artificial /ˌɑː.tɪˈfɪʃ.əl/ 表示“人造的，人工的；仿造的”，英文解释为“made by people, often as a copy of something natural”如：clothes made of artificial fibres 人造纤维面料的服装。

📍词根“art-”来自拉丁语“ars”，表示“艺术，技巧”，“-ficial”来自“facere”，表示“做”，整体意思是“通过技巧做出来的”，即“人工的”。近义词：synthetic（合成的），同根词：artifact（人工制品），常用短语：artificial light（人工光）

📍artificial intelligence /ˌɑː.tɪ.fɪʃ.əl ɪnˈtel.ɪ.dʒəns/ 表示“人工智能”，英文解释为“the study of how to produce machines that have some of the qualities that the human mind has, such as the ability to understand language, recognize pictures, solve problems, and learn”

表示“推出（新产品或服务）”，英文解释为“If a company rolls out a new product or service, or if the product or service rolls out, it is made available to the public.”举个🌰：Microsoft rolls out its new operating system. 微软推出了它的新操作系统。

📍此外，roll out还有另一个含义，表示“将…轧平”，英文解释为“to make sth flat by pushing sth over it”举个🌰：Roll out the pastry. 将油酥面团擀平。

🎬电影《老爸上战场》（Dad's Army）中刚好有这个例句：Shall I roll out the pastry?  我要把面团轧平吗？

📍rollout作名词时，表示“首次提供（产品或服务）”，英文解释为“the act of making something, especially a product or service, available for the first time”举个🌰：Since its rollout in 2011, Wechat has gained billions of members. 自从2011年首次提供使用以来，微信已经吸引了数以亿计的用户。

📍2020年政府工作报告Part 12中就提到：We will support the rollout of e-commerce and express delivery services in rural areas to expand rural consumption. 支持电商、快递进农村，拓展农村消费。

The new AI model, dubbed GPT-4o, can better digest images and video in addition to text, and can interact with people by voice in real time, said Mira Murati, OpenAI's chief technology officer, on Monday. People can interrupt the new voice feature while talking to it, unlike current voice assistants, and the model is capable of responding close to instantaneously, the company said.

OpenAI首席技术官Mira Murati周一表示，这款被称作GPT-4o的新AI模型除了文本之外还能更好地消化图像和视频，并能通过语音与用户实时互动。OpenAI表示，与现有的语音助手不同，用户在与新的语音功能对话时可以打断它，而新模型能够近乎即时地做出反应。

dub /dʌb/ 1）表示“把…称为”，英文解释为“If someone or something is dubbed a particular thing, they are given that description or name.”举个🌰：Today's session has been widely dubbed as a "LearnAndRecord"meeting. 今天的会议被大众称为“LearnAndRecord”的会议。

2）表示“为（影片或声道等）配音”，英文解释为“If a film or soundtrack in a foreign language is dubbed, a new soundtrack is added with actors giving a translation.”举个🌰：It was dubbed into Spanish for Mexican audiences. 它已为墨西哥观众配成西班牙语了。

digest /daɪˈdʒest/ 作动词，1）表示“（使）消化”，英文解释为“to change food in your stomach into substances that your body can use”举个🌰：Sit still and allow your meal to digest. 静静地坐着，让胃里的食物消化消化。

2）表示“理解，吸收（信息）”，英文解释为“to read or hear new information and take the necessary time to understand it”举个🌰：This chapter is so difficult to digest, I'll have to read it again later. 这一章不容易理解消化，稍后我还得再看一遍。

作名词，表示“文摘，摘要；文汇，简报”，英文解释为“a short written report providing the most important parts of a larger piece of writing, or one containing recent news”举个🌰：A digest of the research findings is now available. 研究成果的简报已经出来了。

interact /ˌɪn.təˈrækt/ 表示“交互，交流，交往；互相作用，互相影响”，英文解释为“to communicate with or react to”举个🌰：Her teacher says that she interacts well with the other children. 她的老师说她与别的孩子相处得很好。

capable /ˈkeɪ.pə.bəl/ 表示“有能力的；熟练的；能干的”，英文解释为“able to do things effectively and skilfully, and to achieve results”举个🌰：She's a very capable woman/worker/judge. 她是个很能干的女人/很熟练的工人／能力很强的法官。

📍-capable 表示“可…的，有…能力的（加在名次后形成形容词）”，英文解释为“added to nouns to form adjectives that mean able to use the stated thing”举个🌰：These are nuclear-capable aircraft (= they can carry nuclear weapons). 这些飞机可载核武器。

📍capable of sth/doing sth 表示“有能力做…；能做…”，英文解释为“having the ability, power, or qualities to be able to do something”举个🌰：A force ten wind is capable of blowing the roofs off houses. 10级的大风能吹走屋顶。

instantaneously /ˌɪn.stənˈteɪ.ni.əs.li/ 表示“瞬间地，即刻地”，英文解释为“in a way that happens immediately, without any delay”举个🌰：Participants in chat rooms respond instantaneously to each other. 聊天室中的参与者可以立即相互回应。

OpenAI executives showed in a livestreamed demonstration how the model could analyze code, translate languages between two speakers or guide users through a basic algebra problem written down on a piece of paper, all seemingly in real time.

OpenAI高管在直播演示中展示了新模型是如何分析代码、为两个对话者提供语言翻译，或指导用户完成写在纸上的基本代数问题的，这一切似乎都能实时做到。

livestream/live-stream 表示“互联网直播，线上直播，流直播”，英文解释为“to broadcast video and sound of an event over the internet as it happens, or to be broadcast in this way”

demonstration /ˌdem.ənˈstreɪ.ʃən/ 1）表示“演示，示范”，英文解释为“the act of showing someone how to do something, or how something works”举个🌰：This disaster is a clear demonstration of the need for tighter controls. 这场灾难清楚表明需要进一步加强控制。

2）表示“表示；表达；表现；表明；表露；展现”，英文解释为“a way of expressing a feeling or a quality”举个🌰：Huge crowds followed the funeral procession in a public demonstration of grief. 大批群众跟在葬礼队伍的后面，公开表达他们的悲痛。

algebra /ˈæl.dʒə.brə/ 表示“代数”，英文解释为“a part of mathematics in which signs and letters represent numbers”

The launch of GPT-4o reflects how OpenAI and other startups and tech giants are increasingly seeking to expand their pools of users and bring in money for their generative artificial-intelligence technology after pouring enormous sums of money into computing power and energy to develop their systems.

GPT-4o的问世表明，在投入巨资和精力开发算力和系统后，OpenAI和其他初创公司越来越希望扩大用户群，并将它们的生成式AI技术商业化。

startup / start-up /ˈstɑːt.ʌp/ 表示“刚起步的小企业，新兴小型企业；初创公司”，英文解释为“a small business that has just been started”如：start-up costs 创业成本，举个🌰：Start-ups are very vulnerable in the business world. 商界中刚起步的小企业非常脆弱。

giant /ˈdʒaɪ.ənt/表示“巨头，大公司”，英文解释为“Giant is often used to refer to any large, successful business organization or country.”如：Japanese electronics giant, Sony 日本的电子业巨头——索尼公司。

📍这个词来自希腊神话中的“巨人”（Gigantes），表示在某个领域或行业内非常强大或巨大的实体。

类似的还有：

📍mogul /ˈməʊɡəl/表示“（尤指新闻、影视界的）大人物；大亨”，英文解释为“A mogul is an important, rich, and powerful businessman, especially one in the news, film, or television industry.”如：an international media mogul一位国际传媒大亨。

📍magnate /ˈmæɡneɪt, -nɪt/ 表示“大亨，巨头”，英文解释为“a rich and powerful person in industry or business”。

📍tycoon /taɪˈkuːn/ 表示“巨头，大亨（工商界）”，英文解释为“someone who is successful in business or industry and has a lot of money and power”如：property tycoon 房地产大亨。

📍conglomerate /kənˈglɒmərɪt/ 表示“联合大公司；企业集团”，英文解释为“A conglomerate is a large business firm consisting of several different companies.”如：the world's second-largest media conglomerate 世界第二大传媒集团。

📍behemoth /ˈbiːhɪmɒθ/ 表示“巨头（指规模庞大、实力雄厚的公司或机构）”，英文解释为“a very big and powerful company or organization”。

📍titan /ˈtaɪ.tən/表示“巨人；巨头”，英文解释为“If you describe someone as a titan of a particular field, you mean that they are very important and powerful or successful in that field.”

1）表示“赚取，挣（钱）”，英文解释为“to make money”举个🌰：Their chain of restaurants brings in millions of dollars a year. 他们的连锁餐馆每年有几百万美元的收入。

2）表示“引进（新产品等）；制定，提出（新法律）”，英文解释为“to introduce something new such as a product or a law”举个🌰：New safety regulations have been brought in. 已经制定了新的安全规章。

generative /ˈdʒen.ər.ə.tɪv/ 表示“有生产能力的；能生成的”，英文解释为“able to produce or create something”如：the generative power of the mind 大脑的生成能力。

📍ChatGPT全名就是：Chat Generative Pre-trained Transformer.

📍词根“gen-”来自拉丁语“generare”，表示“产生”，后缀“-ative”表示形容词，即“有产生作用的”。想象生成式AI能够基于现有数据创造新内容。近义词：productive（生产的），同根词：generate（生成），常用短语：generative model（生成模型）

The OpenAI announcement comes a day before the start of Google's annual developer conference on Tuesday, where it is expected to announce its own new products. Google, an AI pioneer, has been vying with OpenAI and its partner and backer, Microsoft, for leadership in generative AI. Microsoft wasn't involved in making GPT-4o.

在OpenAI发布新产品一天后，谷歌（Google）也将于周二召开年度开发者大会，并预计将在会上发布自己的新产品。作为AI领域的先驱，谷歌一直在和OpenAI以及微软（Microsoft）争夺生成式AI领域的霸主地位。作为OpenAI的合作伙伴和支持者，微软没有参与GPT-4o的研发。

vie /vaɪ/ 表示“竞争；争夺”，英文解释为“to compete with other people to achieve or get something”举个🌰：The two older children tend to vie with the younger one for their mother's attention. 两个大一点的孩子正在和小的争宠，都想博得妈妈的注意。

Chief Executive Sam Altman likened the new product to the kind of AI tools typically seen in movies. In a speech last year, he said he and other OpenAI executives found inspiration in the 2013 film “Her” about a man who falls in love with a voice assistant. He and other OpenAI employees posted references to the movie on X just after Monday's announcement.

OpenAI首席执行官阿尔特曼（Sam Altman）将GPT-4o比作电影中常见的那种AI工具。他在去年的一次演讲中说，他和OpenAI的其他高管从2013年的电影《她》（Her）中获得灵感，这部电影讲述了一个男人爱上语音助手的故事。在周一发布公告后不久，他和OpenAI的其他员工在X上提到这部电影。

“The original ChatGPT showed a hint of what was possible with language interfaces; this new thing feels viscerally different,” Altman wrote Monday on his personal blog.

阿尔特曼周一在个人博客上写道：“初代ChatGPT展示了语言界面的雏形，而这款模型能给人截然不同的感觉。”

hint /hɪnt/ 作名词，1）表示“暗示，提示，示意”，英文解释为“something that you say or do that shows what you think or want, usually in a way that is not direct”举个🌰：He's dropped (= given) several hints to the boss that he'll quit if he doesn't get a promotion. 他已经向老板暗示了好几次，如果不获得升职他就会辞职。

2）表示“建议”，英文解释为“a piece of advice that helps you to do something”举个🌰：Could you give us a hint about how to do this exercise, please? 你能不能给我们一些关于如何做这个练习的建议？

3）表示“很少量”，英文解释为“a very small amount of something”举个🌰：There's just a hint of brandy in the sauce. 调味汁中只有很少的白兰地。

作动词，表示“暗示，示意”，英文解释为“to suggest something in an indirect way, but so that someone can guess your meaning”举个🌰：What are you hinting at? 你在暗示什么？

interface /ˈɪn.tə.feɪs/ 表示“接口；界面”，英文解释为“a connection between two pieces of electronic equipment, or between a person and a computer”举个🌰：My computer has a network interface, which allows me to get to other computers. 我的计算机有网络接口，可以与其他计算机连在一起。

viscerally /ˈvɪs.ər.əl.i/ 表示“发自内心地”，英文解释为“in a way that is based on deep feeling and emotional reactions rather than on reason or thought”举个🌰：The performance was both viscerally exciting and intellectually satisfying. 这场演出既让人发自内心地感到兴奋，又让人在知识层面上得到满足。

The new model could also detect a person's emotions in their tone of voice or facial expression, OpenAI said. It also more rapidly switches between different emotional tones from a dramatic voice to a robotic tone to singing. This feature will roll out to users who pay for ChatGPT-Plus, the company's $20-a-month feature in coming weeks.

OpenAI说，新模型还能从语音语调或面部表情中检测人类的情绪。它还能在不同的情感语调之间更快速地切换，从戏剧性的声音到机器人语调，再到歌声。这项功能未来几周将向每月支付20美元的ChatGPT-Plus用户推出。

tone /təʊn/ 1）表示“（常指对听话人的）语气，腔调，口吻”，英文解释为“a quality in the voice that expresses the speaker's feelings or thoughts, often towards the person being spoken to”举个🌰：I tried to use a sympathetic tone of voice. 我试着用一种同情的语气说话。

2）表示“风格；基调，格调；气氛”，英文解释为“the general mood or main qualities of something”举个🌰：I didn't like the jokey tone of the article - I thought it was inappropriate. 我不喜欢这篇文章的搞笑风格——我觉得这不恰当。

3）表示“色调；明暗；影调”，英文解释为“a form or degree of a colour”，如：warm tones of brown and yellow 棕色和黄色这种暖色调。

GPT-4o will also be offered to companies. Murati said the model would be twice as fast and half the cost of its current top-of-the-line offering GPT-4 Turbo. The “o” in GPT-4o stands for “omni,” the company said. People who use the free version of ChatGPT will have access to GPT-4o's image and vision features starting Monday.

GPT-4o还将面向企业推出。Murati说，该模型的速度将是目前顶级产品GPT-4 Turbo的两倍，而成本只有后者的一半。该公司表示，GPT-4o中的“o”代表“一切（omni）”。从下周一开始，使用免费版ChatGPT的用户将可以使用GPT-4o的图像和视觉功能。

top-of-the-line = top-of-the-range /ˌtɒp.əv.ðəˈreɪndʒ/ 表示“（产品）顶级的，最昂贵的”，英文解释为“used for describing something that is the most expensive in a group of similar products”如：a top-of-the-range Mercedes 一辆顶级奔驰车。

omni- /ɒm.nɪ-/ 表示“到处，遍及；一切”，英文解释为“everywhere or everything”如：omnipresent 无所不在的，omniscient 无所不知的。

OpenAI already offers a feature called “voice mode” that combines three separate models to respond to users in voice, but it could be confused by multiple speakers or background noise. It was also slow. The “chains of models” approach “doesn't cut it when you're trying to deliver at this speed,” said Mark Chen, head of frontiers research at OpenAI, in an interview.

OpenAI已经提供了一种名为“语音模式”的功能，是将三个独立模型结合在一起并以语音的方式回应用户，但当多个人一起说话或是背景比较嘈杂的时候，这个功能可能发生混淆，而且速度也不够快。OpenAI前沿研究主管Mark Chen在接受采访时说，这种速度是无法适应“模型链”的。

chain /tʃeɪn/ 1）表示“链条，链子；锁链；项链”，英文解释为“(a length of) rings usually made of metal that are connected together and used for fastening, pulling, supporting, or limiting freedom, or as jewellery”举个🌰：The gates were locked with a padlock and a heavy steel chain. 大门被一把挂锁和一根沉重的钢链锁着。

2）表示“一连串，一系列（的事物）”，英文解释为“a set of connected or related things”举个🌰：She has built up a chain of 180 bookshops across the country. 她在全国各地开了180家连锁书店。

By contrast, GPT-4o was built as a single model trained on text, vision and audio material, and can respond more quickly and accurately to cues. 

相比之下，GPT-4o是根据文本、视觉和音频材料训练而成的单一模型，可以更快更准确地响应提示。

OpenAI executives declined to describe what kind of data was used to train this model. They also declined to say whether OpenAI was able to train the model with less computational power. OpenAI is also working on a whole new AI model, called GPT-5, that is expected to make a large jump from current technology.

OpenAI高管不愿透露训练GPT-4o模型时使用了什么样的数据，也未透露OpenAI是否能利用较少的算力训练该模型。OpenAI还在开发一种名为GPT-5的全新AI模型，有望在现有技术的基础上实现大幅跃升。

Murati said Monday that the OpenAI team wasn't inspired so much by the movie “Her” as by human conversation. “When you stop talking, I'm about to jump in. I can kind of read your tone and respond to that. And it's really natural, rich and interactive,” she said.

Murati周一说，比起电影《她》，OpenAI团队从人类对话中获得的启发更大。“你不说话了，我就接一句。我能读懂你的语气并做出回应。这非常自然，丰富，而且充满互动。”

jump /dʒʌmp/ in表示“打断，插嘴”，英文解释为“to interrupt when someone else is speaking”举个🌰：I wish you'd stop jumping in and finishing my sentences for me all the time. 希望你别老打断我，接我的话茬。

- 词汇盘点 -

- 词汇助记 By ChatGPT -

LearnAndRecord

2015年2月8日

2024年5月14日

第3384天

每天持续行动学外语