Google has presented Project Marineran agent artificial intelligence which has been designed to understand what appears on the browser screen and be able to perform actions as if it were the user themselves. Formerly known as Project Jarvisthis based on Gemini 2.0, Google’s latest language model and is presented as a revolution to take AI to new levels.
What is Project Mariner and how does it work?
As we say, the Mountain View company considers it a revolution because it has the ability to interact with a large number of web pages thanks to a extension that is available in the google chrome browseryes, for the moment, it is experimental. Mariner not only understands textbut can analyze and work with pixels, images and code; also forms, completing them autonomously, in addition to automatically navigating between the sections of a website.
According to Googlethe agent is very useful in automation of repetitive tasks as it could be to collect data, it is also quite efficient in time savingoptimizing processes that could classically require a lot of time. Finally, emphasize intelligent interactionwhich means that, if a request is somewhat ambiguous, Mariner will request explanations and adapt to what the user needs.
In a video, the Google team has shown how Mariner can collect different contact emails from companies that were noted in a spreadsheet. Specifically, they ask you to take the name of the companies and search their websites to obtain their contact email. When the order is sent, it begins to carry it out and, while browsing the pages looking for the information, reports progress in a sidebar that serves as a visual report.
We remember that this is still an experimental version, so I am expected to make mistakes. and is only available for some trusted beta developers of the company, but these types of innovations demonstrate the great advance that artificial intelligence presents every day.
Sign up for our newsletter and receive the latest technology news in your email.