I Tried Gemini Live’s Mind-Blowing Avatar: Here’s What I Discovered!

It can be a little strange when an AI, sounding almost too friendly, tells me to clean up my workspace. There’s something amusing about a robot giving me homemaking tips. My sister would probably agree. Google’s Gemini AI chatbot recently got an upgrade that turns it into quite the helpful companion.

This update, known as Project Astra, is the result of years of development. It brings a new feature called Gemini Live, which lets users interact with the AI through their camera and screen sharing. While it’s not the catchiest name, it boasts impressive capabilities.

Currently, Gemini Live is available for Pixel 9 and Galaxy S25 users, or anyone with an Android phone and a subscription. That subscription runs about $20 a month. I tested it on those devices and also on my OnePlus 13 without any hassle. You can simply summon Gemini with a quick button press or swipe, no complicated setup necessary.

When I pointed the camera at a painting, Gemini instantly recognized it as a Madhubani style artwork. It even provided a fascinating history of the style. I appreciated that it keeps things concise, encouraging you to ask follow-up questions rather than overwhelming you with information.

I decided to test Gemini Live further by showing it my book on machine learning. It recognized the title and offered a detailed summary, even pausing to ask if I wanted more information about specific chapters. This interaction felt remarkably seamless, illustrating the AI’s capability to adapt to user behavior.

However, there are some limitations. Gemini doesn’t use Google Lens yet and can’t pull real-time data from the web. For example, I asked about plant species and restaurant listings. While it did well, I noticed it didn’t provide the most current information.

Gemini also proved to be a great assistant when tackling complex subjects. It summarized intricate topics from books, including one of Bhisham Sahni’s novels. It even included details beyond what was on the page. In a particularly surprising moment, it accurately summarized a chapter of Urdu poetry, though it struggled with pronunciation and accent.

The AI shines in unexpected areas. For instance, I put my sister’s fashion sketch in front of it, and it offered thoughtful feedback and advice on digitizing the design. It’s clear that Gemini is not just about answering questions; it’s about solving problems creatively.

Moreover, the contextual awareness impressed me. For example, it identified Duracell batteries and suggested where to buy them locally, specifically through services like Blinkit and Swiggy Instamart, which cater to urban India.

Nonetheless, it has its flaws. During a test regarding a fitness band, it recognized the device but mistakenly identified it as a different brand when I asked for a follow-up. It also misrepresented some facts confidently; for instance, it claimed Digital Trends hadn’t reviewed a certain device, even though we had published the piece a week earlier.

The AI often hesitates with sensitive topics, such as medical advice or dietary recommendations, prompting users to consult experts instead. This cautious approach is clear during interactions that involve life decisions.

While Gemini Live shows a lot of promise as an AI assistant, its memory issues and occasional inaccuracies, especially with languages other than English, need addressing. The voice can awkwardly shift in tone and speed while reading, which can be jarring.

In summary, Gemini Live is a hefty leap in AI technology—probably one of the most effective uses of generative AI to date. With some improvements, it could become the ultimate tech companion. It’s already capable of much, but it still has some hurdles to overcome before it reaches its full potential.

The journey of AI continues, showing us a glimpse into a future where technology can enhance our everyday lives—if it can just iron out a few kinks along the way.



Source link

Emerging Tech,ai,Artificial intelligence,Gemini,Gemini Live,Google Gemini