Was recently catching up on the latest OpenAI publications and came across a relatively new one titled, "Image GPT". It's a pretty intriguing and highlights the use of GPT ("Generative Pretrained Transformer") to generate images, hence, the name.
Honestly, it's all kind of bananas in a neat, curious and shocking kind of way of what is becoming possible. It'll be interesting to see how this research evolves and what applications it can be used for.
More info on OpenAI's website. Plus, there's some open source code available on GitHub too.