Meta recently introduced a new AI image generation feature called Imagine, which allows users to create images by typing a prompt. This feature is powered by Meta’s AI model, Llama 3, which produces high-quality images with text. Users can access Imagine on WhatsApp and the Meta AI website in beta.
To use Imagine, users type the prompt “Imagine” followed by a description of the image they want to generate. The image changes as they type more letters, and they can animate it, iterate on it, or turn it into a GIF. Meta’s AI assistant has also been updated to provide specific recommendations, like restaurant options with sunset views or vegan dishes.
Although Meta’s CEO, Mark Zuckerberg, called the company’s AI advancements “wild,” Meta still has work to do to catch up with competitors like OpenAI and Anthropic. The company is training larger models of Llama 3, with plans to release them in the future. This release is expected to be on par with OpenAI’s GPT-4 and Anthropic’s Claude 3 Opus.