“This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company,” Alphabet CEO Sundar Pichai wrote in a blog post today on the introduction of Gemini, its most advanced AI model to date.
This technology is adept at processing diverse forms of information including video, audio, and text, offering more sophisticated reasoning and nuanced understanding than Google’s previous iterations. The emergence of Gemini marks Google’s response to the competitive landscape shaped by AI advancements from rivals like OpenAI’s ChatGPT, and is seen as a strategic move to maintain technological leadership.
In a major development, Google has integrated a part of the Gemini model into its AI assistant, Bard, with plans to unveil the most advanced version of Gemini through Bard in the upcoming year. Alphabet is diversifying Gemini into three variants, each tailored to different processing power needs. The range includes a high-powered version for data centers and a more compact version suitable for mobile devices.
Complementing Gemini’s capabilities, Alphabet has also introduced a new generation of its custom-built AI chips, the Cloud TPU v5p. These chips are engineered to efficiently train large AI models and are assembled in pods comprising 8,960 chips. Offering nearly triple the training speed of previous models, these new processors are now available to developers in a preview format, marking another step forward in Alphabet’s ambitious AI journey.
“Today, we’re a step closer to this vision as we introduce Gemini, the most capable and general model we’ve ever built.
Gemini is the result of large-scale collaborative efforts by teams across Google, including our colleagues at Google Research. It was built from the ground up to be multimodal, which means it can generalize and seamlessly understand, operate across and combine different types of information including text, code, audio, image and video.”
— Demis Hassabis, CEO and Co-Founder of Google DeepMind, on behalf of the Gemini team
Featured image: Credit: DeepMind
