Gemini

Gemini is the result of the merger of two Google divisions, Google Brain and DeepMind, which came together in early 2023 to form Google DeepMind. The project was a strategic response to the emergence of ChatGPT. Gemini was initially launched in March 2023 as Bard, but in February 2024, its name was changed to Gemini to signal a new era of AI at Google.
Gemini is led by a major technology company.
Key figures include:
Demis Hassabis: CEO of Google DeepMind. He serves as technical lead.
Sundar Pichai: CEO of Google and Alphabet, who has described the project as the company’s “largest scientific and engineering effort.”
Sergey Brin: Co-founder of Google who helped the development team improve Gemini’s models.
Views
Google designed Gemini with a few main goals:
Multimodality: Gemini was trained from the ground up to understand text, images, video, audio, and programming code.
Integration: The goal is for Gemini to be present across all Google tools and act like a personal assistant.
Responsible AI: Google is emphasizing ethical principles and avoiding bias in Gemini.
The Gemini architecture comes in different versions:
Gemini Nano: for offline processing on phones.
Gemini Pro: a versatile model.
Gemini Ultra: for solving complex scientific problems.
Gemini Flash: fast and efficient.
Gemini uses Transformer technology, which allows it to hold large amounts of information in its memory.