Two events in 2022 define how the public and the industry will think about AI for years.
Stable Diffusion is released in August 2022. Unlike DALL-E and Midjourney, which operate through controlled APIs, Stable Diffusion is open-source and runs on consumer GPUs. The ability to generate photorealistic images from text, locally and without cost, is immediately understood as a structural change to creative software. The broader point — that foundation model weights can be distributed and run at the edge — will have lasting implications for how AI is deployed in enterprise systems.
ChatGPT launches on 30 November 2022 and reaches one million users in five days. No previous software product has grown at this rate. The interface — a conversational text box, familiar to anyone who has ever used a chat application — removes every barrier to trying the technology. Users discover, interactively and in real time, what large language models can and cannot do. The product demonstrates that model capability and interface design are both necessary: GPT-3 had existed for two years, but a well-designed interface changes who can use it and how.
For enterprise software teams, the close of 2022 brings a clear question: what does this capability mean for the products we build, and how quickly do we need to act?