Google transforming AI landscape with Gemini
AS the year ends, Google wants to close it on a high note. This comes as, on Wednesday last week, Google launched Gemini, a game-changing artificial intelligence (AI) model capable of working with various data types, which might become the most disruptive algorithm in the company’s history.
Gemini is an advanced AI model poised to compete with GPT models from OpenAI.
As the debate surrounding the leading force in AI continues, Google’s unveiling of Gemini serves as a bold declaration of its position.
Starting on Wednesday this week, the game-changing AI model, Gemini, will be available to developers worldwide through Google Cloud’s application programming interface (API). This marks the beginning of a global rollout that will see Gemini integrated into various Google products and services in the coming months.
Pixel 8 users will be among the first to experience Gemini’s power, with the model powering suggested messaging replies directly on their keyboards.
Google also says Gemini will enhance Google Search, ads and Chrome, revolutionising how users interact with these essential tools.
Gemini is set apart by its multimodal capability, a departure from traditional AI models confined to specific types of inputs, be they text or images.
Gemini embraces a variety of media — amalgamating text, images, audio, video and even the programming code.
This innovation marks a substantial leap in Google’s scientific and engineering endeavours, as highlighted by CEO Sundar Pichai. Google’s proprietary AI chatbot, Bard, has already undergone an enhancement with a version of the Gemini model.
The company plans to extend Gemini’s integration to products used daily, including the Google search engine and the Chrome web browser, impacting billions of users globally.
This strategic move follows an industry-wide surge that was prompted by the unexpected success of OpenAI’s ChatGPT a year ago, prompting a swift response from major tech players. Gemini’s release serves as a technological leap and an effort to fortify Google’s standing in the public cloud services market, with promises of accelerated AI training
I you remember how OpenAI launched ChatGPT as the year 2022 came to an end. It appears search engine giant Google is following the same script.
through powerful cloud-based processors.
Gemini comes in three sizes: Nano, optimised for mobile devices; Pro, a versatile model for diverse tasks and users; and Ultra, the most sophisticated version currently undergoing safety testing.
Google emphasises advancements in cloud computing, showcasing new processors capable of training large AI models nearly three times faster than previous versions.
Gemini outperformed competing AI models across multiple benchmarks in rigorous testing, evaluating reading comprehension, mathematical prowess and multistep reasoning skills. Eli Collins — vice president (product) at Google DeepMind — acknowledged the frontier-pushing capabilities of Gemini while addressing concerns about AI models generating misleading results.
To mitigate risks, Google announced a gradual release of Gemini Ultra to select customers, developers, partners and safety experts for early experimentation and feedback.
While Gemini represents a significant leap forward in AI capabilities, Google remains committed to addressing challenges and responsibly introducing these powerful models.
The year 2024 promises to be pivotal for the AI revolution, fuelled by recent advancements like Google’s groundbreaking Gemini model.
It is important to note that Gemini is currently in its early stages of development. A limited version is already powering suggested messaging replies on Pixel 8 smartphones.