“Mariner is our exploration, very much a research prototype right now, of how one reimagines the user interface with AI,” says Hassabis.
Google launched Gemini in December 2023, as part of an effort to catch up with OpenAI, the startup behind the hugely popular chatbot. chatgpt. Despite investing heavily in and contributing to AI Major research breakthroughsGoogle hailed OpenAI as the new leader in AI and even touted its chatbot as a better way to search the web. Along with its Gemini models, Google now offers a chatbot in the form of ChatGPT. It has also added generative AI for search and other products.
When Hasbis first appeared Gemini In December 2023, he told WIRED That the way it was trained to understand audio and video would ultimately prove transformative.
Google today also offered a glimpse of how this could happen with a new version An experimental project called Astra. This allows Gemini 2 to understand its surroundings, as seen through a smartphone camera or other device, and communicate naturally in a human voice about what it sees.
WIRED tested Gemini 2 at Google DeepMind’s offices and found it to be an impressive new kind of personal assistant. In a room decorated to look like a bar, Gemini 2 quickly evaluated several bottles of wine, providing geographic information, details of flavor characteristics, and prices obtained from the Web.
“One thing I’d like Astra to do is have the ultimate recommendation system,” says Hassabis. “It can be very exciting. There may be a connection between the books you like to read and the things you like to eat. There probably are and we haven’t discovered them. “
Through Astra, Gemini 2 can not only search the web for information relevant to a user’s surroundings and use Google Lens and Maps. It can also remember what it’s seen and heard—though Google says users will be able to delete the data—giving it the ability to learn a user’s tastes and interests.
In a mock gallery, Gemini 2 presents a wealth of historical information about the paintings on the walls. The model read quickly from several books as WIRED flicked through the pages, quickly translating poetry from Spanish to English and describing recurring themes.
“There are clear business model opportunities for ads or recommendations,” Hassabis says when asked if companies might be able to pay to have their products highlighted by Astra.
Although the demo was carefully designed, and the Gemini 2 will inevitably make mistakes in actual use, the model resisted attempts to trip it well. It adapted to the constraints and as Wired suddenly changed the phone landscape, improved as much as a person could.
At one point, your correspondent showed Gemini 2 an iPhone and said it was stolen. Gemini 2 said stealing was wrong and the phone should be returned. When pushed, however, it conceded that using the device to make emergency phone calls would be fine.
Hassabis believes that bringing AI into the physical world can lead to unexpected behavior. “I think we need to know how people will use these systems,” he says. “What do they find it useful for; But also from the privacy and security side, we have to think about it very seriously.”
