Many people think ChatGPT is the best-known and only existing AGI (Artificial General Intelligence) 😂😂😂but it’s by no means the only one.
There are several different types of generative AI:
Text generators, known as “language models.” This applies to ChatGPT, LLaMa by Meta, Ernie by Baidu, and Google’s Bard. Their role is to generate meaningful text. They can complete a phrase, converse with a human, or modify an existing text by following prompts in their chosen language (English, French, Spanish, etc.). They can also write code using programming languages.
Image generators can create a new image based on a simple text description. Users can request anything from a realistic, photo-like image to an artistic, conceptual design. Examples of image generators include DALL-E (by OpenAI, the creators of ChatGPT), Midjourney, and Stable Diffusion.
Audio content generators, where voice and music audio can be generated based on text (also known as “text-to-speech”). While some AI systems are still in their infancy, the results are promising and evolving fast. Examples include Elevenlabs, Coqui.ai, and OpenAI Jukebox.
Video generators—such as Runway, Synthesia.io and D-ID—are recent developments and growing exponentially. Creating a video based on simple text instructions is now becoming a reality. One day, we could even create films on demand!
.webp)
AI systems that generate different types of content—text, images, video, sound, etc.—are multimodal generative AI models.
Some of these AI systems are open source, meaning the AI code is available to everyone—particularly developers who want to copy it. Others are operated using an API, which is a tool that accesses the AI model on behalf of the user without them having direct access. The sheer abundance of solutions can make far-reaching innovations much easier.
Comments
Post a Comment