Google is set to launch its most extensive and capable artificial intelligence model, Gemini, amid increasing pressure to clarify its AI monetization strategy. Gemini comprises three sizes: Gemini Ultra, the largest and most capable category; Gemini Pro, versatile across various tasks; and Gemini Nano, designed for specific tasks and mobile devices.
Initially, Google plans to license Gemini to customers through Google Cloud for use in their applications. Starting December 13, developers and enterprise customers can access Gemini Pro via the Gemini API in Google AI Studio or Google Cloud Vertex AI. Android developers will also incorporate Gemini Nano. Additionally, Gemini will power Google products such as the Bard chatbot and Search Generative Experience (SGE).
Gemini holds potential for advanced customer service engagement, chatbots, product recommendations, and trend identification for advertising. It could also be utilized in content creation for marketing campaigns, blog content, and productivity apps, summarizing meetings or generating code.
Gemini Ultra stands out as the first model to surpass human experts in massive multitask language understanding (MMLU), spanning subjects like math, physics, history, law, medicine, and ethics. Sundar Pichai, Google’s CEO, emphasized Gemini’s multimodal capabilities, seamlessly handling text, code, audio, image, and video.
Bard, Google’s chatbot, will use Gemini Pro for advanced reasoning, planning, and understanding. An upcoming release, “Bard Advanced,” will leverage Gemini Ultra, representing a substantial update to the existing chatbot.
While Gemini outperforms GPT-3.5, details regarding its comparison with GPT-4 remain undisclosed. Google’s general manager for Bard, Sissie Hsiao, emphasized creating a positive user experience and deferred specifics on monetization.
Gemini’s launch faced delays, attributed to comprehensive testing of its advanced models. Despite being the largest model, Gemini Ultra is reportedly more cost-effective to serve, reflecting enhanced efficiency in training.
Google will release a technical white paper for Gemini, with more details on the model’s capabilities. Simultaneously, the company introduced its next-generation tensor processing unit, TPU v5p, offering improved performance for training AI models.
Investors have questioned Google’s AI profitability strategy, and Gemini’s launch aligns with the company’s efforts to explore AI monetization avenues. Google’s experimental project, Search Generative Experience (SGE), reflects a conversational approach, yet its public launch timeline remains uncertain. Incorporating Gemini into SGE is expected within the next year.
Sundar Pichai expressed excitement about Gemini’s potential, marking it as a significant science and engineering effort that will unlock opportunities globally.
