Newsletter
In May 2023, at the Google I/O developer conference, CEO Sundar Pichai unveiled Gemini, an upcoming Google GPT-4 rival AI system. The Google DeepMind division, comprising the Brain Team and DeepMind, is developing this large language model (LLM). According to sources, Gemini could launch pretty soon.
Gemini, set to be multimodal, can handle text, images, and other data types for more natural conversations.CEO Sundar Pichai also hinted at potential capabilities like memory and planning for tasks requiring reasoning.
ChatGPT is a well-known name in generative AI. Google is developing its version, Gemini, to compete. The Information reported that Google is giving some developers early access to Gemini via its API. This lets them integrate Gemini into various applications. While these developers are using a limited version of Gemini, Google is still training the full version, which is expected to compete with GPT-4 and its release date is not far from now.
Google Gemini could launch to the public soon
In an interview with Wired a few days later, Pichai gave the clearest indication of Gemini’s role in Google’s roadmap. He stated that conversational AI systems like Bard are transitional, not the final destination, leading to advanced chatbots. Pichai emphasized that Gemini and future versions will evolve into universal personal assistants deeply integrated into people’s daily lives, including travel, work, and entertainment. He reiterated that Gemini would blend text and images, making current chatbots seem trivial in comparison within a few years.
Google offers generative AI tools through Vertex AI for its Cloud customers. Gemini is likely to be part of this offering. Business users can also access Duet AI, which is integrated into Workspace apps, for an additional $30 per month. However, Duet AI relies on Google’s existing AI models and primarily handles text-based queries.
It’s important to note that Gemini has the potential to generate different kinds of data, including text, images, and code. This makes it different from Google’s previous models like PaLM 2, which are limited to text-based input and output. While we don’t have all the details yet, Google’s efforts in this field show promise, and we can expect more sophisticated generative AI tools from the company in the future.