Yesterday Sundar Pichai tweeted and introduced Gemini 1.0 and mentioned it as Google’s most capable and general AI model yet.
Introducing Gemini 1.0, our most capable and general AI model yet. Built natively to be multimodal, it’s the first step in our Gemini-era of models. Gemini is optimized in three sizes - Ultra, Pro, and Nano
Gemini Ultra’s performance exceeds current state-of-the-art results on… pic.twitter.com/pzIw6iCPPN
— Sundar Pichai (@sundarpichai) December 6, 2023
Demis Hassabiss, CEO and Co-Founder of Google DeepMind also tweeted and introduced Gemini 1.0.
We’re excited to announce 𝗚𝗲𝗺𝗶𝗻𝗶: @Google’s largest and most capable AI model.
Built to be natively multimodal, it can understand and operate across text, code, audio, image and video - and achieves state-of-the-art performance across many tasks. 🧵 https://t.co/mwHZTDTBuG pic.twitter.com/zfLlCGuzmV
— Google DeepMind (@GoogleDeepMind) December 6, 2023
What is Gemini from Google?
Gemini is built from the ground up for multimodality — reasoning seamlessly across text, images, video, audio, and code. Google Gemini is multimodal and first one in the Gemini-era of models. It is the one of the most popular methods to test the knowledge and problem solving abilities of AI models. Gemini is optimized in three sizes - Ultra, Pro, and Nano. Pichai added, Ultra’s performance exceeds current state-of-the-art results on 30 of the 32 widely-used academic benchmarks. With a score of 90.0%, Gemini Ultra is the first model to outperform human experts on MMLU.
Gemini 1.0 is optimized in three sizes
- Gemini Ultra— Google’s largest and most capable model for highly complex tasks.
- Gemini Pro— Google’s best model for scaling across a wide range of tasks.
- Gemini Nano— Google’s most efficient model for on-device tasks.
What is MMLU?
MMLU (Massive Multitask Language Understanding) is a new benchmark designed to measure knowledge acquired during pretraining by evaluating models exclusively in zero-shot and few-shot settings. This makes the benchmark more challenging and more similar to how we evaluate humans. The benchmark covers 57 subjects across STEM, the humanities, the social sciences, and more. It ranges in difficulty from an elementary level to an advanced professional level, and it tests both world knowledge and problem solving ability. Subjects range from traditional areas, such as mathematics and history, to more specialized areas like law and ethics. The granularity and breadth of the subjects makes the benchmark ideal for identifying a model’s blind spots.
What can Gemini do?
Google Gemini can help in the following problem solving scenarios:
- It Excels at competitive programming
- It Unlocks insights in scientific literature
- It can Process and understand raw audio signal end-to-end
- It can explain and help understand in math and physics
- It can reason about the user intent to generate bespoke experiences
On the Google technology blog Demis added Its remarkable ability to extract insights from hundreds of thousands of documents through reading, filtering and understanding information will help deliver new breakthroughs at digital speeds in many fields from science to finance.
Demis also added that Gemini was combined with robust filters, this layered approach is designed to make Gemini safer and more inclusive for everyone. Additionally, we’re continuing to address known challenges for models such as factuality, grounding, attribution and corroboration. It will surely will enhance creativity, extend knowledge, advance science and transform the way billions of people live and work around the world.
With these kind of features it can be helpful to coders, businesses, students, teachers and parents.
Google also announced that they have incorporated Gemini Pro in Bard for new ways to collaborate with AI. Gemini Ultra will come to Bard early next year in a new experience called Bard Advanced.
What is Bard?
Bard is Google's experimental, conversational, AI chat service. It is meant to function similarly to ChatGPT, with the biggest difference being that Google's service will pull its information from the web.
Bard has a share conversation function and a double check function that helps users fact-check generated results. Bard can also access information from a number of Google apps and services, including YouTube, Maps, Hotels, Flights, Gmail, Docs and Drive, letting users apply Bard to their personal content.
Let's roll back to late November 30, 2022, when ChatGPT was released. Less than a week after launching, ChatGPT had more than one million users. According to analysis by Swiss bank UBS, ChatGPT became the fastest-growing 'app' of all time. Other tech companies, including Google, saw this success and wanted a piece of the action.
In the same week that Google unveiled Bard in February, 2023, Microsoft unveiled a new AI-improved Bing, which runs on a next-generation OpenAI LLM customized specifically for search.
ChatGPT was released on November 30, 2022. In just under a week post-launch, ChatGPT amassed over one million users, escalating to the status of the fastest-growing 'app' in history, according to an analysis conducted by the Swiss bank UBS. The success of ChatGPT caught the attention of various tech giants, Google among them, all eager to partake in this burgeoning phenomenon.
In a parallel move during the same week in February 2023, Google unveiled Bard. Simultaneously, Microsoft introduced an upgraded version of Bing powered by AI advancements. This new Bing iteration utilized a cutting-edge OpenAI LLM specifically tailored for enhanced search functionality.
Google Gemini Availability
- Google Gemini is now available on Pixel 8 Pro and Bard
- It will be accessible to developers and enterprise users from December 13.
- Gemini Ultra is still under evaluation and will be released for wide usage next year.
The AI game has just begun but it is for sure that it is here to stay.
This video below highlights some interactions with Gemini.
Learn more and try the “multimodal prompting” model on the link below:
https://developers.googleblog.com/2023/12/how-its-made-gemini-multimodal-prompting.html
