At the beginning of 2025, the rapid development of the AI industry has become a focus that cannot be ignored. Major companies have increased their investment in hardware and professional applications, and these investments are now gradually turning into technological achievements that are within the reach of consumers.
Taking Google as an example, its dual-track parallel strategy is particularly eye-catching. On the one hand, Gemini 1 has been widely deployed in terminal devices and gradually popularized to ordinary users. on the other hand, the development of Gemini 2 is also in full swing, and the project team continues to show new AI agent technology. At the same time, Google’s Notebook LM has also attracted widespread attention. It has quickly grown from a simple note-taking tool to an AI application with more functions.
Different Aspects of AI Agent Capabilities
On the Google blog, Google CEO Sundar Pichai and others elaborated on the construction concept of this AI technology, among which “native tool use” became the core keyword. This means that AI entities will be able to use various tools as skillfully as humans to perform more complex tasks. Google announced that Gemini 2.0 is now available to developers and trusted testers, and plans to quickly integrate it into products, especially Gemini and search. The Gemini 2.0 Flash experimental model is now available to all Gemini users, and a new feature called Deep Research has also been launched in Gemini Advanced, which uses advanced reasoning and long-context capabilities to serve as a research assistant to explore complex topics and write reports.
The rise of AI agents indicates that they will begin to take on more tasks that were originally done by humans. In actual demonstrations, Notebook LM showed its strong potential in the sales field. AI agents can automatically contact the last touchpoint in the sales funnel, obtain information, and create an opportunity roadmap. They can also browse proprietary content management systems and make decisions based on the latest published content. This endless possibility makes us think about the new role of AI in business processes.
In addition to performing specific tasks, AI agents also show amazing capabilities in planning. In the demonstration, the presenter used Notebook LM for long-term planning, such as developing a five-year project plan or conducting a SWOT analysis. In the past, these tasks may require people to use sticky notes to record and organize processes, but now, AI agents can easily do it. This transformation not only improves work efficiency, but also allows us to see the huge potential of AI technology.
In the field of search, AI has also brought revolutionary changes. Sundar Pichai said that AI Overview has covered 1 billion people, enabling them to ask new types of questions and quickly became one of the most popular search features. In order to further enhance the user experience, Google is introducing the advanced reasoning capabilities of Gemini 2.0 to AI Overview to handle more complex topics and multi-step questions. Currently, the feature has begun limited testing and is scheduled to be launched more widely early next year. In the future, AI Overview will also introduce more countries and languages to provide global users with a more convenient and efficient search experience.
The application of AI agents in Gmail is also worth looking forward to. Faced with a large amount of inbox content, we often find it difficult to keep up with the pace. If AI agents can automatically browse your Gmail account and filter out the most relevant and actionable content for you, then this will undoubtedly greatly improve our work efficiency and quality of life.
Notebook LM Benefits Research And Writing Work
Google has given Notebook LM the ability to source content from Google Slides and URLs, in addition to existing support for Google Docs, PDFs, and text files. This allows users to create annotations or ask questions about content in a document (whether it’s images or text) or explore online material. Google has also added inline citations to help users view supporting paragraphs in sources, fact-check AI-generated responses, and read the original text for more context.
Steven Johnson, editorial director at Google Labs, said Notebook LM was developed in conjunction with authors, students, and educators, and the company is seeing early adopters integrating its source infrastructure into their research and writing workflows. The company said Notebook LM has also been used to create hyperlocal newsletters, summarize interview transcripts, develop grant proposals, and even manage descriptions of fantasy worlds. The model’s native multimodal capabilities allow users to ask questions about visuals in sources and even use Notebook LM to provide citations to related images as supporting evidence in their responses.
However, Notebook LM is unique in that it builds on the user’s own documents and sources. This grounded architecture reduces the risk of AI models “hallucinating” or generating biased or incorrect information, a common problem with other AI chatbots. By limiting the AI to only user-provided data, Notebook LM creates a personalized AI research assistant that is an expert on the information most relevant to the individual user. This makes it a powerful tool for authors, students, educators, and professionals in a variety of fields. As these technologies advance at a rapid pace, we need to confront this latest trend and think about a new relationship between humans and AI.
