Google's Bard Upgraded with Gemini Multimodal AI
Today, Google debuted Gemini, its newest "largest and most capable" artificial intelligence model. Since it was designed to be multimodal, it can generalize and comprehend several forms of data simultaneously, including text, pictures, audio, video, and code. This improves its ability to understand subtleties and improves its ability to respond to inquiries about complex subjects. As a result, it excels at clarifying thinking in challenging disciplines like physics and arithmetic.
There are three sizes available: Ultra, Pro, and Nano. Pro is the "best model for scaling across a wide range of tasks," Nano is the "most efficient model for on-device tasks," and Ultra is "the largest and most capable model for highly complex tasks."
Gemini Ultra can read, comprehend, and produce excellent Python, Java, C++, and Go code. It is multilingual and capable of reasoning over complicated data. It performs exceptionally well in several coding benchmarks, such as HumanEval and Natural2Code, Google's proprietary held-out dataset that takes advantage of author-generated sources rather than web-based data.
It's unclear what "fine-tuned version" of Gemini Pro Google's Bard AI is employing. According to Google, this should provide it with "more advanced reasoning, planning, understanding, and more". Will it surpass ChatGPT as a result? Although it's entirely up to you to guess, Google undoubtedly wants you to believe that.
Indeed, it claims that in six of the eight benchmarks it tested, Gemini Pro scored better than GPT-3.5. Additionally, Google discovered that "Bard is now the most preferred free chatbot compared to leading alternatives" after conducting some blind assessments with outside raters. To be clear, Google hasn't identified the "alternatives" it is referring to, therefore neither do we.
To date, you may test Bard with Gemini Pro now for text-based prompts in English in 170 regions, but not Europe (which is coming shortly). Gemini Ultra will be introduced to "a new Bard Advanced experience" early in the next year.
But there's still more! Unexpectedly, the first smartphone "engineered for Gemini Nano" is the Pixel 8 Pro—again, whatever that means. You read correctly: only the Pixel 8 Pro, not the Pixel 8 as well. It appears that Gemini Nano is already being used by the Pro's Summarize in Recorder and Smart Reply in Gboard capabilities.
To speed up the experience, the corporation is also "starting to experiment with Gemini in Search". Additionally, Gemini will power features in additional Google products and services, "like Ads, Chrome, and Duet AI," "in the coming months."