Gemini is designed to perform complex tasks such as summarizing or generating text, creating contextual images, analyzing flow charts, and even assisting software engineers in writing code. Google CEO Sundar Pichai stated in May that Gemini was still in training, but already showing impressive multimodal capabilities. Once thoroughly tested for safety, Gemini will be available in various sizes and capabilities.
Key takeaways:
- Google has provided early access to its latest AI model, Gemini, to a select group of companies, indicating that it is nearing readiness for incorporation into consumer services and business sales.
- The Gemini model is expected to compete with OpenAI's GPT-4 and will be used to power Google's AI chatbot Bard, which currently runs on Palm 2.
- Gemini is also expected to be integrated into Google's other products, including enterprise apps like Google Docs and Slides.
- The model is designed to perform complex tasks such as summarizing or generating text, creating contextual images, analyzing flow charts, controlling software with voice, and assisting software engineers in writing code.