Models & Labs
Worldmodeldata brokers video game data for AI training
WIRED AIhigh confidence
Why it matters
- →Addresses the critical shortage of high-quality physical interaction data for world models.
- →Offers a scalable, cost-effective alternative to manual sensor data collection.
- →Highlights the ongoing debate between using simulated vs. real-world physics for AI training.
Worldmodeldata, a UK-based startup advised by Yann LeCun, has licensed nearly one million hours of video game data to train world models. The company acts as a broker, aggregating controller inputs and visual data from various studios to create training datasets for embodied AI. While Nvidia and other labs prefer custom physics engines for fine motor control, Worldmodeldata argues that the volume and variety of gameplay data are essential for general navigation and handling edge cases in autonomous systems. Industry experts remain divided on whether game physics is sufficiently accurate for real-world deployment.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
More from WIRED AI
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
