In the case of Stable Diffusion vs OpenAI, a model was released to the public for anyone on earth to experiment with, creating a diverse ecosystem of ideas. There are thousands of use cases now that are being tested live on the open market -- a number that any single company would find incredibly difficult to compete with. Most ideas won't gain any footing, but some will. By the time those ideas gain popularity, OpenAI will only be left to copy them.

This is specifically with regards to Stable Diffusion vs DALL-E. I think OpenAI has plenty of breathing room when it comes to innovation in the broader spectrum of new AI applications. I think there might be some interesting discussions here regarding when it's a good idea to provide a cheap and available version of your offering vs when to provide a quality offering. Both definitely have their place in the market and can result in superior products, but right now to me it's not clear where that distinction should be drawn.