Stability AI loses key researchers and adds new API services
Key Points
- Stability AI loses key employees, including Robin Rombach and other leading researchers involved in the development of the Stable Diffusion text-to-image model.
- Despite this, Stability AI is introducing a comprehensive suite of new API image services to the developer platform based on the latest image models, including SD3.
- The new API services provide developers and creators with tools to create, enhance, and edit images with generative AI, and are divided into four categories, including new original media generation, 4K scaling, intuitive image editing, and soon, image post-processing control tools.
The AI image generator Stable Diffusion turned Stability AI into a unicorn. Now a large part of the core team is leaving the company. At the same time, Stability is launching new API services.
British AI unicorn Stability AI is dealing with an exodus of key employees. According to Forbes, Robin Rombach and other key researchers who were instrumental in developing the text-to-image model Stable Diffusion have left the company. CEO Emad Mostaque reportedly announced the departures at an all-hands meeting last week
Rombach led the research team that conducted the foundational research on Stable Diffusion at the Ludwig Maximilian University of Munich and the University of Heidelberg. Along with Andreas Blattmann and Dominik Lorenz, he is one of the five authors of the seminal research paper.
Stability hired the researchers after the paper was published. Most recently, they helped develop Stable Diffusion 3, which for the first time combines the diffusion structure of earlier versions with the transformers of ChatGPT.
New API services for image generation and editing
Despite this, Stability AI has introduced a comprehensive suite of new API image services on the Stability AI developer platform. Utilizing the latest image models, including SD3, these services provide developers and creators with extensive tools for generating, enhancing, expanding, and editing images using generative AI.
The goal of the services is to simplify the development of user-centric solutions and facilitate the creation of high-quality images without complex prompt engineering, according to the company.
The services fall into four categories: creating new original media from text descriptions, upscaling images to 4K, intuitive image editing with natural language instructions, and upcoming control tools to ensure that the results match the user's intent.
The latter is particularly useful for developers and creatives who require consistent and predictable results. The new services are available via a REST API.
AI News Without the Hype – Curated by Humans
As a THE DECODER subscriber, you get ad-free reading, our weekly AI newsletter, the exclusive "AI Radar" Frontier Report 6× per year, access to comments, and our complete archive.
Subscribe now