Recent advancements in artificial intelligence have enabled the creation of real-time generated games without traditional game engines. A neural network generates future frames based on mouse and keyboard inputs, allowing for a game like Counter-Strike to be visualized entirely through AI diffusion models. This technology, demonstrated on standard home computers and 4060 laptops, utilizes GPUs and frameworks such as Torch with CUDA. The potential applications extend beyond gaming, with implications for robotics and real-world interactions. Open-source safety layers are anticipated to ensure the safe deployment of these world models. Additionally, future games may incorporate neural rendering and NPCs running llama1b, and the technology's impact is expected to be profound, as experienced by users playing in the latent space.