Google on Wednesday introduced its most sophisticated artificial intelligence model, Gemini 2.0, as the world’s tech giants sprint to achieve the top position in this rapidly developing technology.
CEO Sundar Pichai proclaimed that Gemini 2.0 is a model that will mark what the company names “a new agentic era” in AI development by means of AI models that are capable of interpreting and making decisions about the world.
“Gemini 2.0 is all about making data a lot more valuable,” Pichai explained in the introduction, stressing the model’s advanced skills to comprehend context, think a few extra steps ahead, and take steps on behalf of users with supervision. “These advances come closer to our dream of a universal assistant,” he included.
The release caused shares of Google to soar to over four percent on Wall Street, the day following a 3.5 percent gain already after the release of a breakthrough quantum chip.
Tech giants are working on faster and more potent AI models even if they are really expensive and have no direct impact on the everyday economy. Agents are the newest digital aides in Silicon Valley, they are supposed to be aware of the surroundings, make decisions, and execute the tasks to bring out the desired goals.
The tech giants lay out the vision of agents being the next step in the AI revolution initiated by 2022’s release of
ChatGPT, which has become popular around the world. Gemini 2.0 is initially being rolled out to developers and trusted testers, with plans for broader integration across Google’s products, particularly in Search and the Gemini platform.
No Nvidia
The era is powered with the aid of Google’s sixth-era TPU (Tensor Processing Unit) hardware, dubbed Trillium, which the company has now made generally available to customers.
Google emphasised that Trillium processors have been used completely for each training and jogging Gemini 2.0.
Most AI training has been monopolized by means of chip juggernaut Nvidia, which has been catapulted via the AI explosion to turn out to be one of the world’s maximum treasured organizations.
Google said that thousands and thousands of developers are already constructing packages with Gemini technology, which has been integrated into seven Google merchandise, each serving more than two billion customers.
Gemini 2.0’s powers are predicted to are available in early 2025 to Google’s seek application, nevertheless the company’s fundamental cash-maker.
The first release from the two.0 circle of relatives of models may be Flash, offering quicker overall performance at the same time as managing multiple varieties of enter (text, pictures, video, audio) and output (together with generated pix and speech).
Gemini customers global can already tap into a chat-only model of Flash, the enterprise said, with testers given get right of entry to to a multimodal version which could interpret pictures and environment.
Google also said it became experimenting with a product that may use software apps, websites and different online tools, much like a human consumer. OpenAI and Anthropic have unveiled similar features.
The agency additionally teased a new edition of Project Astra, a telephone virtual assistant like Apple’s Siri that responds to pics in addition to verbal instructions.
