Author丨Kong Haili.
Editor丨Zhu Yimin.
Source丨Xinhua News Agency.
The battle for large models is in full swing, and Google, which has been robbed of the first-mover advantage by OpenAI, is catching up.
On the night of February 8, Beijing time, Google made a series of big moves:Bard, which originally provided AI interactive functions, was officially renamed Gemini, and Gemini model functions will be fully introduced into Google products, including Workspace (such as Docs, Sheets, and Slides) and Google Cloud.
In order to make it easier for users to access Gemini, Google has specially added the Gemini App for Android models, and the iOS system can use Gemini directly from the Google App.
Gemini Ultra, known as Google's strongest AI model, is also officially launched, and users can pass $19 per month$99 Google One AI Premium plan to experience Gemini Ultra.
This design is quite similar to OpenAI's ChatGPT Plus subscription service, which offers the GPT-4 version for a monthly service fee of $20The meaning of-for-tat is quite obvious.
After more than a year of accumulation, Google intends to regain lost ground in the large-scale model competition with the new version of Gemini, and even regain the leading rhythm.
Gemini gets a makeover
In November 2022, ChatGPT was born, innovating the way of human-computer interaction, and the battle of large models has been out of control since then, with OpenAI and Microsoft out of the limelight.
As the original pioneer of artificial intelligence, in order to benchmark ChatGPT, Google has successively launched the chatbot Bard in March 2023 and the Gemini large model in December, which still carries functions with Bard. At that time, according to the benchmark released by Google, Gemini scored more than GPT-4 in 30 out of 32 performance test criteria.
However, users have mixed reviews of Gemini, and its demo** is widely questioned as a fine-cut compilation album, which is suspected of hype marketing.
Subsequently, Oriol Vinyals, VP of Research at Google DeepMind, admitted that all of the human prompts and AI outputs in ** are real, but shortened for brevity.
Now, with the name of Bard changed to Gemini, Google has extended the large model horizontally.
According to Google CEO Sundar Pichai, the Gemini model will soon be widely available in Google products, including Workspace and Google Cloud.
It is reported that more than 1 million people are already using features such as Duet AI's "Help Me Write" in Google Workspace to improve productivity and creativity through Duet AI. Next, Duet AI functions will be integrated into Gemini.
For Google Cloud users, Duet AI will also change to Gemini in the coming weeks. Gemini will play a role in helping companies increase productivity, increase developer coding speed, help protect organizations from cyberattacks, and more.
Developers are the testing ground for every major technological advancement, and they will also play an important role in the Gemini ecosystem. Sundar Pichai said hundreds of thousands of developers and businesses are already building with the Gemini model, and Google is already training the next iteration of the Gemini model.
In order to make it easier for users to use the Gemini model on their mobile phones, Google has created a new app for Android.
Similar to Siri's role for iOS, Google's app for Android is also positioned as a mobile phone assistant, just like waking up the phone with "Hey, Siri", "Hey, Google" will also become a new way to call the Android system.
Android users can call Gemini by pressing the power button, swiping from the corner of the screen, or shouting "Hey, Google". This will result in a whole new overlay experience, where users can easily use large model features on the screen by waking up the gemini, such as generating a title for a freshly shot or asking a question about an article being read.
According to Sissie Hsiao, vice president and general manager of Google Assistant & Bard, many of the features originally provided by Google Assistant will be available through the Gemini app, including setting timers, dialing**, and controlling smart home devices.
iOS phones that already have Siri Assistant will be able to use Gemini directly from the Google App in the coming weeks.
Starting on the evening of February 8, Beijing time, Gemini will launch the English version on Android and iOS phones in the United States, and from next week, more countries will be able to access Gemini in English, Japanese, and Korean.
A new Gemini Ultra experience
The legend of Gemini Ultra has been foreshadowed, and it has been hailed as the most powerful and advanced version of Google's model, carrying Google's expectations and users' curiosity.
ultra 1.0 is the first model to outperform human experts in MMLU (Massive Multitasking Language Understanding), which uses a combination of 57 subjects, including mathematics, physics, history, law, medicine, and ethics, to test the knowledge and problem-solving skills of large models.
Nowadays, ultra 10 products have finally landed. In order to unify the pace with the Gemini universe, Ultra 1Version 0 is named Gemini Advanced.
Sissie Hsiao said that in the third-party assessment,Users feel that Gemini Advanced is the most popular chatbot right now.
According to Google's official introduction, Gemini Advanced not only supports longer and more detailed conversations, but also can better understand contextual prompts, and can be competent for more advanced programming schemes, more complex logical reasoning, more in-depth data analysis and other functions.
Gemini Advanced is now available in English in more than 150 countries and territories, and Google will expand it to more languages over time.
But just like OpenAI's paid service for GPT-4, Google has also set up payment standards for Gemini Advanced.
Google's new Google One AI Premium plan will include Gemini Advanced for $19 per month$99 with a free trial for the first two months.
That is, the original charge was 9$9 per month for the Google One Premium plan, in addition to the existing service, Gemini Advanced will come as an additional plus for $19$9 per month is included in the package.
In addition, AI Premium subscribers will soon be able to use Gemini in Gmail, Docs, Slides, Sheets, and more (formerly Duet AI).
Google also emphasized that they will continue to be bold and responsible, taking AI technology to the world. And, to reduce issues such as unsafe content or bias, Google is focusing on building security into its AI products. Prior to the launch of Gemini Advanced, they conducted extensive trust and safety checks to improve the model's capabilities based on feedback.
In the deployment of large AI models, in the past year, Google has indeed missed the first-mover opportunity, and now it has launched the Gemini Ultra model, which is more than Gemini 10 may have obvious upgrades and improvements, and more real answers should refer to the follow-up feedback of Google One AI Premium members.
If there is not much difference in the abilities of the two sides, 19The $9 month Gemini Advanced will have many more ecological advantages than the $20 month GPT-4.
Google, which is intensively training large models, is still unknown how far it is from its goal of "providing the world's most advanced, safest and most responsible AI", but Google is backed by the largest search database and has the advantage of a large number of users, these factors may add some buffs to its large model competition.
sfc
Editor: Liu Xueying, intern: Zhao Fengling.
21 Jun recommended reading
Google announces: remove cookies
Nearly 5 billion yuan in 7 years! What happened to the "first share of AI chips"?
The Great Reversal! OpenAI announced: Ultraman will return as CEO