#Mobieee #Tech News - 0003
Beginning with Gemini Ultra, Google will make available to users its most sophisticated models and features with the launch of Bard Advanced early in the following year.
 |
| Google Bard.AI |
With the release of Gemini, the most recent and sophisticated AI model from the company, Google's artificial intelligence (AI) chatbot Bard is receiving the most significant upgrade to date.
Gemini will come in three sizes: Nano (for on-device tasks), Pro (for scaling across a wide range of tasks), and Ultra (for extremely complex tasks).
Bard's vice-president for Google Assistant, Sissie Hsiao, announced that Gemini will be released to Bard in two stages: Beginning on December 6, Bard will be powered by a specially tuned version of Gemini Pro, increasing its comprehension and summarization, reasoning, coding, and planning capabilities.
Currently, Gemini Pro will support text-based prompts in Bard; multimodal support will be added in the upcoming months. More than 170 countries and territories will be able to access it in English at first, and in the near future, additional languages and regions, including Europe, will be supported.
Google is set to launch Bard Advanced early in the upcoming year, providing users with access to the company's most cutting-edge models and features, beginning with Gemini Ultra.
See: Google introduces Gemini, the biggest AI model it has to compete with OpenAI
"Gemini Ultra was designed to process and respond to a variety of input formats, including text, images, audio, video, and code, rapidly. The most widely used programming languages in the world can be understood, explained, and high-quality code can be produced by Gemini Ultra thanks to its multimodal reasoning capabilities, according to Hsiao.
Originally released by Google in February 2023 as a response to OpenAI's ChatGPT, Bard was previously powered by the PaLM 2 large language model from the company.
Prior to its public release, Hsiao claimed that Gemini Pro beat OpenAI's GPT-3.5 in six out of eight benchmarks, including MMLU (massive multitask language understanding) and GSM8K (Grade School Math 8K), which assesses grade school math ability.
While GSM8K consists of 8,500 excellent multi-step reasoning grade school math problems, MMLU tests world knowledge and problem-solving skills across 57 subjects, including math, physics, history, law, medicine, and ethics.
GPT-4 Turbo, an enhanced version of OpenAI's flagship GPT-4 model, was just released. Nevertheless, GPT-3.5 continues to power ChatGPT, the company's free chatbot. Users need to pay a monthly subscription fee in order to access the most recent OpenAI models.
It will be interesting to see if Google decides to follow OpenAI's lead and charges a subscription fee for Bard Advanced, providing users with a state-of-the-art AI experience.
Hsiao stated that they had nothing specific to share about this topic in a press briefing. "At this time, our priority is providing Gemini users with the greatest possible product experience. We will investigate potential forms of monetization as the experience progresses," she stated.
According to Hsiao, they are finishing up safety and trust audits and will soon begin a "trusted tester" program for Bard Advanced, with a wider rollout planned for early next year.
Google has been pushing to integrate generative AI capabilities throughout their product suite, and in recent months, they have added several new features to Bard.
 |
| Google : YouTube, Gmail, Docs, and Maps |
Currently, Bard is deeply integrated with a number of the company's consumer apps and services, such as YouTube, Gmail, Docs, and Maps. It can now comprehend YouTube videos as well as respond to specific queries about the subject matter of some of these videos.
The company also enhanced Bard's visual search features, enabling users to upload photos via Google Lens and receive images from Google Search in their replies. Additionally, the chatbot can adapt the tone and manner of responses to over 40 supported languages and nations.
Google introduced Assistant with Bard in October 2023. Bard is a virtual assistant that combines Assistant's personalized assistance with its generative and reasoning powers.
"We are thrilled to be able to realize our vision for Google Assistant thanks to Bard's generative abilities, sophisticated reasoning, and language understanding. When the announcement was made, Hsiao stated, "It's a step towards our vision to deliver the world's most helpful personal assistant."