Google announced on Wednesday the arrival of Gemini 2.0, its most advanced generative artificial intelligence (AI) model to date, designed to compete with other tech giants in a fast-growing sector.
According to Google boss Sundar Pichai, this new version should usher in ‘a new era of generative AI capable of directly facilitating users' everyday lives.
For the time being, the new version of Gemini is only available to certain people, particularly developers, ahead of wider release in early 2025, according to Google. The company then plans to integrate it into its range of different products, such as its search engine, and in several languages.
“Gemini 2.0 is about making information much more useful,” Pichai said in a blog post announcing Gemini 2, highlighting the model’s enhanced ability to understand context, think multiple steps ahead, and take supervised actions on behalf of users.
Google, along with ChatGPT creator Open AI, Meta - Facebook's parent company - and Amazon, are in a race to develop new generative AI models at breakneck speed, despite their massive costs and doubts about their real usefulness to society at this stage.
What Google is seeking to do is the new trend in Silicon Valley: to turn AI tools into the new digital butlers, all-knowing secretaries accessible at all hours and capable of carrying out numerous tasks on the user's behalf.
Its promoters claim that this use represents the next major step in the arrival of AI for the general public, following the sensational arrival of ChatGPT in 2022.
According to Google, millions of developers are already using previous versions of Gemini.
Gemini 2.0 is trained and operated using an internally developed chip called Trillium. Most of the development of generative AI models has been carried out using hardware from the US giant Nvidia, which specialises in graphics chips, or GPUs.
