Trying out Gemini in Chrome leaves me wanting more | The Verge

Gemini in Chrome: Initial Impressions

This article reviews the early access integration of Google's Gemini AI assistant within the Chrome browser. The author explores its capabilities, limitations, and potential future developments.

Key Features and Use Cases

Gemini's integration allows users to interact with the AI directly within Chrome, leveraging its ability to 'see' the screen content. The author highlights several use cases:

• Summarizing web articles and identifying key information on webpages.
• Extracting information from YouTube videos, including identifying tools and summarizing content.
• Retrieving recipes from YouTube videos and identifying products from online listings.
• Utilizing a 'Live' feature for voice-based queries and responses.

Limitations and Inconsistencies

The author points out several limitations:

• Gemini's ability to 'see' is limited to the currently active tab and visible content.
• Summaries can be lengthy and not always concise.
• Access to real-time information, such as product listings and live locations, is inconsistent.
• Responses can be repetitive and sometimes inaccurate.

Future Potential

Despite these shortcomings, the author sees significant potential for future improvements, particularly with Google's 'agentic' AI development initiatives, like Project Mariner. The author speculates on Gemini's potential to perform tasks like ordering food, managing bookmarks, or saving YouTube videos.