Former OpenAI researcher unveils flipbook prototype using AI to generate screen pixels

ME News reports that on April 23 (UTC+8), according to monitoring by Beating, former OpenAI researcher Zain Shah and his team have released Flipbook, an experimental prototype that generates screen pixels directly using AI models, replacing traditional web technologies such as HTML and CSS. Every “page” viewed by users is an AI-generated image; clicking any area of the image generates a new image to proceed further. The entire interface contains no HTML code, no fixed links, no predefined buttons, and even text consists of pixels within the image. The video mode is based on LTX Studio, an open-source DiT (Diffusion Transformer) video generation model developed by the Israeli company Lightricks, optimized to stream in real time at 1080p 24fps via WebSocket to the user’s screen, with the backend powered by Modal Labs’ serverless GPUs. Shah stated that Flipbook currently has limited functionality, as the team has designed it around visual interpretation, but it demonstrates a broader direction: as models become more accurate and stateful, future extensions could include structured UIs, even programming environments. Shah previously worked at OpenAI on AI and robotics research, then served as a Creative Technology Expert at Samsung, and is also an alum of Y Combinator’s S13 cohort. Other team members include Eddie Jiao, former engineer at Humane and Slack, and Drew O’Carr, former Apple engineer. (Source: BlockBeats)