• by Laurie Sullivan  @lauriesullivan, December 11, 2024

In a race to become the dominant player in artificial intelligence (AI), Google on Wednesday released what it calls its most "capable" AI assistant as the technology becomes the answer to organizing the world's information to keep users on its search engine and advertisers on its ad platforms.

"Information is at the core of human progress," Google Chief Executive Officer Sundar Pichai, along with Google DeepMind CEO and CTO, wrote in a blog post published Wednesday. "It’s why we’ve focused for more than 26 years on our mission to organize the world’s information and make it accessible and useful."

This is also why Google continues to push into AI to organize that information across every input to make it accessible via any output, the group wrote. 

Think of how AI can support all the world's information in search and advertising.

No product has been transformed by AI more than search. The company's AI Overviews now reaches 1 billion people, enabling them to ask entirely new types of questions.

Soon AI Overviews will have the advanced reasoning capabilities of Gemini 2.0, allowing users to tackle more complex topics and multi-step questions, including advanced math equations, multimodal queries and coding.

Limited testing began this week. It will begin rolling out more broadly early next year.

A lot is at stake -- not only for Google, but other companies like OpenAI, Microsoft and many other tech companies and startups that are using AI to change consumer and user behavior. 

Pichai also introduced Google's next era of models -- Gemini 2.0 -- built for the agentic era.

These models are a type of AI system that act autonomously on behalf of its users, making decisions and taking actions to achieve goals. 

The model includes advancements in multimodality such as native image and audio output, and native tool use. Pichai explained how this technology will enable Google to build new AI agents that bring the company closer to its vision of a universal assistant.

Google has been working quickly to get the technology into its products, starting with Gemini and Search. Its Gemini 2.0 Flash experimental model became available today to all Gemini users.

The group also wrote about launching a new feature called Deep Research that uses advanced reasoning and long context capabilities to act as a research assistant, exploring complex topics and compiling reports on the user's behalf. It is now available in Gemini Advanced.

Gemini 2.0’s advances were built on decade-long investments in its full-stack approach to AI.

It is built on custom hardware like Trillium, its sixth-generation Tensor Processing Unit (TPU), a custom application specific integrated circuit (ASIC) that speeds machine learning workloads.

TPUs powered 100% of Gemini 2.0 training and inference, and today Trillium is generally available to customers so they also can build with it.