pixelated images ChatGPT Image prompts

2 months ago

A non-binary person in their 20s with wavy shoulder-length hair and subtle retro makeup stands confidently on an urban 2000s-style street during golden hour. They’re dressed in Y2K fashion: baggy jeans, a cropped graphic tee, and tinted sunglasses, exuding effortless cool. The scene is shot with a Nikon Coolpix 995, built-in lens, auto aperture, at a low and slightly tilted angle, giving it that classic CCD sensor aesthetic — soft tones, pixel bleed, subtle motion blur, and a visible date stamp in the corner. The image is framed in 4:3, evoking the authentic early-2000s digital photo look.

3 months ago

A mesmerizing digital art piece in the style of glitch art, showcasing a cybernetic marvel amidst a chaotic scene. Shattered expectations are visualized through distorted, pixelated textures and fractured geometries, creating a sense of digital disarray. The glitch matrix aesthetic is dominant, with vibrant neon colors bleeding and flickering across the fragmented forms. This artwork blends cyberpunk elements with the aesthetic of digital glitches, creating an immersive, chaotic landscape where reality and technology collide. The composition is a dynamic interplay of shattered images and shimmering glitches, capturing the essence of a digital marvel in a state of constant flux. The colors are intense, and the atmosphere is charged with a sense of electric energy, reflecting the chaos and uncertainty of the digital realm

4 months ago

A wide image taken with a phone of a glass whiteboard, in a room overlooking the Bay Bridge. The field of view shows a woman writing, sporting a tshirt wiith a large OpenAI logo. The handwriting looks natural and a bit messy, and we see the photographer's reflection. The text reads: (left) "Transfer between Modalities: Suppose we directly model p(text, pixels, sound) [equation] with one big autoregressive transformer. Pros: * image generation augmented with vast world knowledge * next-level text rendering * native in-context learning * unified post-training stack Cons: * varying bit-rate across modalities * compute not adaptive" (Right) "Fixes: * model compressed representations * compose autoregressive prior with a powerful decoder" On the bottom right of the board, she draws a diagram: "tokens -> [transformer] -> [diffusion] -> pixels"