Just two years ago, Google looked like it was in trouble. In a desperate move to play catch-up with the OpenAI chatbot that had upended the tech industry, the search giant debuted an unpolished version of its artificial intelligence on Google.com. The A.I. spat out shoddy information, including advice for people to eat rocks and put glue on their pizza.

仅仅在两年前，谷歌似乎还深陷困境。为了拼命追赶颠覆了整个科技行业的OpenAI聊天机器人，这家搜索巨头在Google.com上推出了一个尚未打磨完善的人工智能版本。该系统输出了许多极其糟糕的信息，包括建议人们吃石头，以及在披萨上抹胶水。

Google’s reign over the internet seemed at risk.

谷歌对互联网的统治地位似乎摇摇欲坠。

But today, consensus is forming in Silicon Valley not only that Google has recovered and caught up but that it could actually win the A.I. race, a testament to how so much can change in so little time.

但在今天，硅谷正形成一种共识：谷歌不仅已经恢复元气并实现了赶超，而且实际上有可能赢得这场人工智能竞赛。这证明即使在短短时间内也可以发生巨大的变化。

Google announced on Tuesday at Google I/O, its conference for software developers, that in only one year the number of people regularly using its chatbot, Gemini, had more than doubled to 900 million, on a par with OpenAI’s self-reported number of active users for ChatGPT and nearly 30 times the estimated web traffic of Anthropic’s Claude chatbot, which is more focused on business customers.

谷歌在周二的软件开发者大会Google I/O上宣布，在短短一年内，定期使用其聊天机器人Gemini的人数增加了一倍多，达到9亿人。这一数字与OpenAI自报的ChatGPT活跃用户数持平，且几乎是Anthropic旗下更侧重企业客户的Claude聊天机器人预估网络流量的30倍。

And unlike Anthropic and OpenAI, which are still losing money from operating expensive A.I. data centers, Google is rapidly developing ways to use A.I. to increase profit with online advertising, its bread and butter. In its last quarter, Google reported, its advertising revenue rose 16 percent to $77 billion, fueled by A.I. technology that has helped marketers collect deeper information about users’ interests.

与仍在因运营昂贵的人工智能数据中心而亏损的Anthropic和OpenAI不同，谷歌正在迅速开发新方法，利用人工智能来增加其核心业务网络广告利润。谷歌报告称，在上一个季度，由于人工智能技术帮助营销人员收集了关于用户兴趣的更深层次信息，其广告收入增长了16%，达到770亿美元。

Pretty soon, Gemini will become a staple on a competitor’s platform: Apple’s iPhone. Google and Apple said in January that Gemini would become the foundational A.I. technology for a future version of Siri, Apple’s voice assistant. Gemini is already included on Android devices, so the partnership with Apple will, in effect, bake Gemini into virtually all of the world’s phones.

Gemini很快将成为竞品平台苹果iPhone上的核心配置。谷歌和苹果在1月份表示，Gemini将成为苹果语音助手Siri未来版本的底层人工智能技术。由于Gemini已经内置于安卓设备中，因此与苹果的合作实际上将把Gemini植入全球几乎所有的手机中。

At I/O this week, Google shared its vision for how it would use A.I. to increase sales for online retailers, including through a new type of shopping cart that automatically surfaces promotions to shoppers.

在本周的I/O大会上，谷歌分享了利用人工智能增加在线零售商销量的愿景，其中包括一种新型购物车，可自动向购物者推送促销信息。

The company also said it was continuing to add its A.I. into tools that people use every day — for instance, by letting a person using Google Docs ask Gemini to draft a speech that includes personal anecdotes and jokes.

该公司还表示，它将继续把人工智能融入人们日常使用的工具中——例如，允许谷歌文档用户要求Gemini起草一份包含个人轶事和笑话的演讲稿。

“If I had to put a wager on the biggest winner of A.I., I would say it’s Google,” said Gary Rivlin, a former New York Times reporter who wrote a book about the tech industry’s race to cash in on A.I.

“如果让我押注谁是人工智能最大的赢家，我会说是谷歌，”《纽约时报》前记者加里·里夫林表示，他曾写过一本关于科技行业竞相从人工智能中获利的书。

“If you asked me that same question a year and a half ago, the answer wouldn’t have been Google,” he added.

“如果你在一年半前问我同样的问题，答案绝不会是谷歌，”他还说。

Google’s success with Gemini is a case study in how even a juggernaut can seemingly transform itself overnight to remain dominant. But past incumbents faced with similar existential threats did not have the same story.

谷歌在Gemini上取得的成功是一个经典案例研究，展示了即使是行业巨头，似乎也能在一夜之间完成自我转型，从而保持主导地位。然而，过去那些面临类似生存威胁的老牌巨头却没能得到这样的成功

About 15 years ago, Microsoft, whose leadership in personal computing was challenged by the rise of Apple’s iPhone and Google’s Android ecosystems, failed to stay relevant in the mobile phone era even after releasing its Windows Phone devices and acquiring the handset maker Nokia for $7.2 billion. Microsoft remains extremely profitable thanks to business customers, and it has a large video game business, but it no longer leads the pack among consumers.

大约15年前，微软在个人电脑领域的领导地位受到了苹果iPhone和谷歌安卓生态系统崛起的挑战。即使推出了Windows Phone设备，并以72亿美元收购了手机制造商诺基亚，微软也没能在移动电话时代保持竞争力。得益于企业客户以及庞大的电子游戏业务，微软依然盈利丰厚，但在消费者市场中已不再领跑。

(The Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to A.I. systems. The two companies have denied the suit’s claims.)

（《纽约时报》已对OpenAI和微软提起诉讼，指控其侵犯了与人工智能系统相关的新闻内容版权。这两家公司均否认了诉讼中的指控。）

Instead of isolating Gemini, Google reinvented itself by blending A.I. into all of its most important products. To name a few: There is a team working on Gemini for Gmail, another working on Gemini inside Google Maps and yet another focused on the Gemini chatbot app.

谷歌并没有将Gemini孤立起来，而是通过将人工智能融入其所有最重要的产品，从而实现自我重塑。例如：有一个团队正致力于将Gemini应用于Gmail，一个团队致力于将Gemini整合到谷歌地图中，还有一个团队专注于Gemini聊天机器人应用本身的开发。

Google is also in a position to use A.I. to capitalize on the travel industry. Gemini is the only chatbot with native access to Google Flights and Google Hotels for looking up airfare and lodging, which makes it a superior travel agent.

谷歌也具备利用人工智能在旅游业中获利的条件。Gemini是唯一能够原生接入谷歌航班和谷歌酒店以查询机票和住宿的聊天机器人，这使它成为一个更优秀的旅行顾问。

To put it another way: What Google has that other A.I. companies lack is cultural cachet. And with the Gemini model integrated into so many popular services that people rely on every day for work and play, Google has made it much more likely that people will interact with its A.I., said Carolina Milanesi, the president of Creative Strategies, a consumer technology research firm.

换句话说：谷歌拥有其他人工智能公司所缺乏的东西——文化威望。消费科技研究公司创意战略的总裁卡罗琳娜·米拉内西表示，随着Gemini模型被整合到人们日常工作和娱乐所依赖的众多热门服务中，谷歌极大地增加了人们与其人工智能进行互动的可能性。

“It’s not about just how good the model is, but where it is and how easy it is for people to discover Gemini and then use it and get value,” Ms. Milanesi said.

“这不仅仅在于模型有多好，而是在于它所在的位置，以及人们能否轻松发现Gemini、使用它、并且从中获得价值，”米拉内西说。

Google clearly leaned into its advantages to improve its technology. It made A.I.-generated answers a permanent fixture on Google.com — still the most visited website in the world — without giving users the ability to opt out of seeing them.

谷歌显然利用了自身优势来改进其技术。它将人工智能生成的答案作为固定内容呈现于至今仍是全球访问量最大网站Google.com，且没有给用户提供关闭该功能的选项。

Despite making glaring mistakes about the nutritional value of rocks, among other errors, in its early days, Google quickly honed the model. But there is still some work to do. A recent analysis from The Times concluded that Google’s A.I.-generated responses were correct 90 percent of the time. Google disputed the study and argued that its answers were more accurate.

尽管在早期犯下了关于石头的营养价值等显而易见的错误，谷歌还是迅速优化了该模型。不过仍有一些工作要做。《纽约时报》最近的一项分析得出结论，谷歌人工智能生成的回答准确率为90%。谷歌对这项研究提出了异议，并坚称其回答更加准确。

“They just have this reach that few, if any, companies on Earth have,” Mr. Rivlin said. “Looking for the next Google? It’s probably Google.”

“他们拥有这种地球上极少数公司才拥有的触达能力，”里夫林说。“寻找下一个谷歌？那可能还是谷歌。”