Google (GOOGL, GOOG) has announced Gemini, its answer to OpenAI’s GPT-4, which will launch with three variants: Gemini Nano for Android devices, Gemini Pro for Google AI services (including Bard), and Gemini Ultra for data centers and enterprise applications. Each variant is tailored to specific uses, highlighting Google’s commitment to diverse AI applications.One of Gemini’s standout features is its multimodal capabilities, enabling it to process and interact with text, images, video, and audio. Google hopes this integrated approach will offer a more comprehensive range of functionalities compared to OpenAI’s separate models for different media types. Going forward, I think multimodal models are table stakes; Google’s main advantage will be Gemini’s integration into Google’s ecosystem, including search, advertising, and the Chrome browser.Google also anticipates that Gemini will have a notable impact on programming, with its new AlphaCode 2 system showing significant advancements over its predecessor.Efficiency is a key aspect of Gemini’s design. Developed on Google’s Tensor Processing Units, it is both faster and more cost-effective than previous models. The concurrent launch of the TPU v5p system underscores Google’s emphasis on sustainable and scalable AI technology.Google has pledged a balance of innovation and caution, which is particularly welcome as the industry edges closer to the concept of artificial general intelligence (AGI).Early demos of Gemini look amazing. The company says that Gemini beats GPT-4 in 30 out of 32 benchmarks… but that’s just the way engineers flex. This contest will be fought on your devices, and you will judge which platform(s) are best for you.This may feel like a fight between the most powerful Jedi masters, but “there is another…”More By This Author:Microsoft Upgrades Copilot To GPT-4 Turbo And DALL-E 3
Meta And IBM Unite For AI Innovation
AI And The End Of Humanity
