Google Bard's Gemini: Updated Capabilities

As Artificial Intelligence (AI) continues to grow, Google released an update today about a product it's been working on called Gemini.

I've listed some general information and summary below:

Introduced by Sundar Pichai and Demis Hassabis

Key points:

• Gemini is the most capable and general AI model Google has ever built.
• It is multimodal, meaning it can understand and operate across text, code, audio, image, and video.
• It is also flexible, running on everything from data centers to mobile devices.
• Gemini surpasses state-of-the-art performance on a range of benchmarks.
• It has sophisticated reasoning capabilities, enabling it to uncover insights from complex data.
• Gemini can understand, explain, and generate high-quality code in popular programming languages.
• An advanced version, AlphaCode 2, excels at competitive programming problems.
• Gemini is built with responsibility and safety at its core, with comprehensive safety evaluations and dedicated safety classifiers.
• It is rolling out across Google products and services, including Bard, Pixel, Search, and more.
• Developers and enterprise customers can access Gemini Pro starting December 13th.
• Gemini Ultra will be available for early experimentation and feedback soon.
• This marks a significant milestone in AI development and the start of a new era for Google.

Additional details:

• Gemini 1.0 is optimized for three different sizes: Ultra, Pro, and Nano.
• Gemini Ultra performs better than human experts on the MMLU benchmark.
• It also surpasses state-of-the-art performance on the MMMU benchmark.
• Gemini Pro is being used in Bard, Pixel, and other Google products.
• Android developers can build with Gemini Nano via AICore in Android 14.
• Bard Advanced will give users access to Gemini Ultra and other cutting-edge models.

Overall, Gemini is a powerful new AI model that has the potential to revolutionize many fields. It is built with responsibility and safety in mind, and it is being made available to developers and enterprise customers so that they can start building innovative applications.

I think it's another step and improvement upon that which is built. Still no where near of the "fear of AI" that it's going to take over the world. However a way to see some of the capabilities how we might use this as a tool going forward.