In April, Alphabet CEO Sundar Pichai took an unusual step: bringing together two large artificial intelligence teams—with different cultures and code—to catch up and beat OpenAI and other competitors. Now that effort is being put to the test with hundreds of people scrambling to run a group of large machine learning models — one of the highest-stakes products the company is making this fall. The information: The models, known as Gemini, are expected to give Google the ability to create products that its competitors can’t, according to a person involved in Gemini development. OpenAI’s GPT-4 multilingual model can understand and produce conversational text. Gemini will go beyond that, combining the text capabilities of LLMs like GPT-4 with the ability to create AI images based on a text description, similar to the AI image generators Midjourney and Stable Diffusion, this person said. Gemini’s imaging capabilities have not been reported so far.
Googlers have also discussed using Gemini to offer functions such as analyzing charts or creating graphics with text descriptions and controlling software through text or voice commands. Google is betting on Gemini to power services ranging from its Bard chatbot, which competes with OpenAI’s ChatGPT, to enterprise apps like Google Docs and Slides. Google also wants to charge app developers for access to Gemini through its Google Cloud server rental unit. Google Cloud currently sells access to more primitive AI models created by Google through a product called Vertex AI. These new features could help Google catch up with Microsoft, which is racing with new AI features for its Office 365 apps and also selling access to OpenAI models to its app customers.
Comments are closed.