Shulou(Shulou.com)12/24 Report--

Thank CTOnews.com netizens soft media Xinyou 2203184 for the clue delivery! CTOnews.com, December 3 (Xinhua)-- according to The Information local time on Saturday, two people familiar with the matter revealed that Google has quietly postponed the public launch of its AI model Gemini to January next year. Before that, Gemini's "debut" and a series of related events were scheduled to be held in California, New York and Washington next week.

According to one of the people familiar with the matter, the postponement was made by Google CEO Sandal Pichai after Google found that Gemini could not reliably handle some non-English query instructions. According to the report, Gemini was supposed to be Google's "most important" product launch this year. Google is eager to catch up with OpenAI, consuming a lot of computing resources and merging large teams.

Currently, the Gemini team wants its primary model to reach the same level as GPT-4, or even surpass the latter. People familiar with the matter claim that Google has met these standards in "some ways" but is still improving it. Because the team wants Gemini to run normally in multiple languages around the world.

CTOnews.com quoted the same media report last month, when there was already news of Gemini's "dystocia": its progress had been delayed to the first quarter of next year. The recent report by The Information revealed more details about the ticket skipping of Gemini.

The system combines the technology behind AlphaGo with the large language model, and the goal is to make the system have new capabilities, such as planning or solving problems, which are stronger than OpenAI's GPT-4 capabilities.

When Gemini is completed, it may play an important role in Google's competition against generative AI technologies such as ChatGPT. Google pioneered a number of technologies that led to breakthroughs in AI, but chose to carefully develop and deploy products based on these technologies.

Gemini AI combines text generation capabilities similar to GPT-4, and supports "create images based on text description" features, so the AI tool can achieve "chat dialogue", "analyze chart data", "create images" and other functions, and even use natural language commands to control software.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.