Google unveils Gemini 2.0, tech giant's 'most capable model yet ...
Google has recently unveiled its most advanced artificial intelligence model, Gemini 2.0, which represents a significant step forward in the tech giant's pursuit of a "universal assistant". According to Google CEO Sundar Pichai, Gemini 2.0 is the 'most capable model yet' and signals the beginning of a new era for the company, known as the "new agentic era".
Advancements in Multimodality
Pichai highlighted the new features of Gemini 2.0, including advancements in multimodality such as native image and audio output, as well as native tool use. These enhancements will allow Google to develop new AI agents that bring the company closer to its vision of a universal assistant.
Google initially introduced its first AI model, Gemini 1.0, back in December 2023. The release of Gemini 2.0 has further propelled Google's advancements in multimodality and context understanding, enabling the model to process various forms of multimedia information effectively.
Market Impact
The unveiling of Gemini 2.0 had a significant impact on Google's stock performance, with shares soaring by more than four per cent on Wall Street. This surge followed a 3.5 per cent gain in the company's stock after the introduction of its 'mind-boggling' quantum chip, Willow.
AI Agent Innovation
Gemini 2.0 is part of Silicon Valley's latest trend of AI 'agents' - digital assistants designed to perceive their environment, make decisions, and execute actions to achieve specific objectives. The model is expected to enhance user experiences and streamline interactions.
Technical Details
Google revealed that Gemini 2.0 is developed on custom hardware, including Trillium, the sixth-generation TPU (Tensor Processing Unit) from Google. This hardware is now accessible to all customers and powered the training and inference processes of Gemini 2.0.
Future Developments
Looking ahead, Google plans to make Gemini 2.0 Flash available to all users in January, along with additional model sizes. The company also intends to expand the integration of Gemini 2.0 into more of its products in early 2025.
Additionally, Google hinted at a new iteration of Project Astra, a smartphone digital assistant comparable to Apple's Siri, with enhanced capabilities to converse in multiple languages. This development aligns with Google's commitment to enhancing user experiences through innovative AI technologies.
