Google has unleashed its Gemini Pro API, ushering in a new era for developers and organizations. The API, now available on Google Studio and through Google Cloud’s Vertex AI platform, promises state-of-the-art capabilities that are set to redefine the landscape of AI tools, models, and infrastructure.
Thomas Kurian, the CEO of Google Cloud, expressed the company's commitment to rapidly integrating Gemini's cutting-edge features across all services. In an interview with indianexpress.com, Kurian stated, "We are moving quickly to bring the state-of-the-art capabilities of Gemini to all our services. And you should assume that we’re gonna bring it to every single surface that Google has."
The Gemini Pro API is not the only star in Google's constellation of AI advancements. The tech giant has introduced several other models on the Vertex AI platform, including an upgraded Imagen 2 text-to-image diffusion tool and a family of foundation models tailored for the healthcare industry called MedLM. Duet AI, both for developers and in security operations, has also been announced, marking a comprehensive expansion of Google's AI offerings.
Google's approach to refining and enhancing its AI models involves a disciplined process, as Kurian explained. The company initiates a preview phase, enabling collaboration with a broad community of developers and consumers. This iterative process allows Google to incorporate valuable feedback, ensuring the models evolve to meet diverse user needs.
According to Kurian, user feedback is a critical component of fine-tuning their models. The reinforcement learning from human feedback is automated, providing insights into user experiences and preferences. This collaborative approach extends to specific industries, recognizing the unique language nuances and requirements of domains such as legal and medical professions.
Josh Woodward, VP of Google Labs, emphasized the company's active engagement in hackathons to stay closely connected to user feedback. This user-centric approach, including observations during onboarding processes, informs the design and development of Google's AI tools.
The introduction of Gemini Pro has not only opened new possibilities for developers but also showcases Google's commitment to ongoing improvement. While the initial version is accessible via the Gemini API, Google plans to fine-tune the model in the coming weeks and months. The Gemini Pro boasts features such as embeddings, function calling, semantic retrieval, custom knowledge grounding, and chat functionality. Supporting 38 languages across over 180 countries and territories, Gemini Pro is positioned as a versatile tool for diverse applications.
Developers currently have free access to both Gemini Pro and Pro Vision through Google Studio, making it an ideal choice for app development needs. Moreover, users of Vertex AI can explore the same models with similar limits at no cost until general availability in early 2024.
In addition to Gemini Pro, Google has integrated Imagen 2 and MedLM into Vertex AI, offering advanced text-to-image diffusion technology and healthcare-focused foundation models. The company has also introduced DuetAI, bringing AI-powered code and chat assistance to developers and integrating it into Security Operations, making it the first major cloud provider to offer generative AI in a unified SecOps platform.
With competitive pricing and expanded indemnification, Google is addressing user concerns related to copyright, solidifying its position as a leader in the AI landscape. The unveiling of Gemini Pro and the associated AI tools marks a significant leap forward, showcasing Google's dedication to innovation and collaboration in the ever-evolving field of artificial intelligence.
