Gemini: Google’s Artificial Intelligence Tool

Googles GEMINI Just SHOCKED The ENTIRE INDUSTRY! (GPT-4 Beaten) Full Breakdown + Technical Report [Video]

Video uploaded by TheAIGRID on December 6, 2023.

Gemini, the Most Powerful AI Tool on the Market?

Google launched its new artificial intelligence tool. Named Gemini, it is more powerful than any other on the market, including the popular ChatGPT, created by OpenAI, now partnered with Microsoft.

“It marks the beginning of a new AI era at Google: the Gemini era,” said Sundar Pichai, CEO of the tech giant. He added, “This is a huge leap forward in an artificial intelligence model that will impact virtually all Google products.”

Among its features, its multimodality stands out: according to Google, Gemini is capable of fluently reasoning through texts, images, video, audio, and codes, using what is known as the MMLU model, the acronym for Massively Multitask Language Understanding.

In terms of its versions, Gemini will be available in three formats: Nano (the simplest, in the form of a mobile phone application), Pro, and Ultra, the most powerful AI tool for highly complex tasks.

The new AI can receive visual or auditory information (musical notes, images, words) and from there generate its own content, which in turn can be in text, audio, or image form. In the tool’s presentation, Google demonstrated that Gemini is capable of detecting errors in a complex mathematical problem, giving the correct answer, and explaining the steps to reach it. And, for fun, it can play with the user in popular challenges such as Rock, Paper, Scissors, or the ball under the cup puzzle.

Regarding its code knowledge, Gemini works with the most common languages, such as Java, Python, C++, and Go, according to its creators, and although it makes some mistakes, Google stated that it can resolve them 90% of the time when communicated.

When Gemini will be launched and integrated into Google

The executives and developers who worked on Gemini explained that the full potential of the tool will be applied to Bard, their current AI chatbot system that already operates in several languages, including Spanish.

For now, the new AI is being deployed and is already operational: Bard now works with Gemini Pro, and users of Pixel 8 Pro will get some new functions thanks to Gemini Nano. For now, Gemini is only available in English, although it is evident that it will soon be available in other languages.

Finally, Pichai stated that Gemini will be fully integrated into Google’s search engine, its advertising products, the Chrome browser, and many more tools.

Google has introduced Gemini Pro to its AI Studio and Vertex AI platforms, with plans to integrate it into Duet AI in the coming weeks

AI Studio, a web-based coding interface, simplifies app development using mostly natural-language commands, ideal for individual programmers and small teams.

Meanwhile, Duet AI, Google’s coding assistant enhanced with generative AI, will also incorporate Gemini in the near future. Announced as generally available, Duet AI is set to benefit from Gemini’s capabilities.

Thomas Kurian, CEO of Google Cloud, highlighted Gemini’s integration into Google’s vast AI infrastructure, powered by the custom Tensor Processing Unit (TPU). This infrastructure, previously internal, is now accessible to customers.

AI Studio’s preview of Gemini Pro offers an impressive 60 requests per minute for free, significantly more than other similar tools. A demonstration showcased how AI Studio can create a real estate app using natural language and photo descriptions.

Vertex AI, unlike AI Studio, focuses on large-scale corporate data, integrating with various data sources, including Google Cloud partners, for enterprise-level applications.

Google Cloud is cautious about Gemini Pro’s broader rollout, focusing on its current availability in AI Studio and Vertex AI. This announcement coincides with Google Cloud’s free Applied AI summit for developers.

Users can explore Vertex AI with a free trial, enjoying the same quota of 60 requests per minute as AI Studio.

Google Cloud plans to introduce formal pricing for both AI Studio and Vertex AI at a fraction of a cent per 1,000 characters once Gemini Pro is widely available early next year.

The company has also reduced Google Cloud pricing significantly for inputs and outputs, according to Kurian.

Gemini has three versions: Pro, Ultra, and Nano. While Ultra is in private preview, Nano is slated for mobile device release and is already operational on the Pixel 8 smartphone.

Google Cloud also unveiled Imagen 2, an enhanced text-to-image neural network, available in Vertex AI’s Model Garden, which now includes third-party neural networks like Mistral, ImageBind, and DITO.

Links

• Gemini – Google DeepMind
• Sell your old Google Pixel smartphone online – iGotOffer