ElevenLabs released the beta version of its AI voice synthesis platform and announced a $2 million pre-seed round. It drew attention for text-to-speech that reproduces natural, human-like intonation.
Midjourney released "niji," an image generation model co-developed with anime studio Spellbrush. It debuted as the company's first model specialized in anime-style illustration.
Just one week after ChatGPT's public release, Perplexity AI launched an "answer engine" that responds to questions directly, with citations. Its approach of merging search and chatbot drew attention as something new.
OpenAI released the conversational AI "ChatGPT," based on GPT-3.5, as a free research preview. It surpassed 1 million users in just 5 days, kicking off the generative AI boom.
Stability AI released Stable Diffusion 2.0. While a new text encoder and stronger safety filters improved quality, the changes sparked debate over prompt incompatibility and content restrictions.
Midjourney released its V4 model, built on a completely new AI architecture. Depicting people and objects became far more accurate, and it became the foundation for all versions that followed.
OpenAI released a DALL-E 2 API, allowing developers to embed image generation directly into their own apps and services. A 1024x1024 image was priced at $0.02.
OpenAI removed DALL-E 2's invite-only waitlist, letting anyone start using it immediately after signing up. By this point, more than 1.5 million people were generating over 2 million images a day.
Adept released its first model, "ACT-1 (Action Transformer)," which follows natural-language instructions to click and type within a browser. The company positioned it as a first step toward a foundation model capable of operating any software.